There is now a very comprehensive chicken linkage map containing well over 800 loci. However, in order to use this map to identify the genes responsible for traits it is necessary to progress from linkage information to physical clones of the relevant region. The identification of chicken yeast artificial chromosome (YAC) clones and the gene of interest is the first step towards physical cloning of a candidate gene but at present the efficiency of this process would be greatly improved if the density of genetic markers in the targeted region could be increased. The initial approach to this problem was to place as many known genes as possible on the map, by using expressed sequence tags (ESTs) as markers. This placed seven novel chicken genes on the linkage map. However, this method was inefficient, so a method of saturating the genome with markers was applied. Random amplified polymorphic DNA (RAPD) analysis was investigated as a means of increasing marker density over the whole genome, but was also inefficient at producing large numbers of markers. Then the potential of representational difference analysis (RDA) was investigated to target markers to specific regions of the chicken genome. As an initial test of the method the genomes of two inbred lines were compared. This provided a relatively undemanding test of the technique and generated large numbers of polymorphic clones. As a more stringent test of the method, and to provide marker loci in a region of particular interest, a second comparison was made targeting chicken chromosome 16. Chromosome 16 contains the major histocompatibility complex (MHC), nucleolar organiser region (NOR) and Rfp-Y complex. This succeeded in targeting 10 out of 18 RDA clones to chromosome 16. It appeared possible to target RDA towards the direct identification of a trait gene or genes and as a model for this, RDA was applied to the identification of genes affecting resistance to Marek's disease. This generated four RDA loci which were located on chromosome 1 three of which showed association with both mortality and Marek's disease virus quantitative PCR product values. These RDA loci appear to identify a novel region of the chicken genome containing a trait gene conferring resistance to Marek's disease. Thus RDA appears to provide a viable solution to the generation of targeted markers, which will aid YAC identification in specific areas.

I would like to thank Professor John Bourne and the Institute for Animal Health, Compton for allowing me to undertake the research for this degree.

I would also like to thank my supervisor Dr Nat Bumstead for all his help and support, especially with the proof reading of this thesis and his belief that PCR is not controlled by the tooth fairies! My thanks also go to Dr Ayo Toye, for generating the chicken YAC library, letting me use it and giving positive encouragement in the final stages of writing up. I would also like to thank all those within the Avian Genetics lab. especially Julie Sillibourne and Nigel Salmon without whose technical support everything would have been a great deal harder.

I would also like to acknowledge the help of my external supervisor Dr Terry Burke and all those who have previously provided markers for the Compton chicken linkage map, without which a great deal of this work would have been impossible.

I am also grateful to all other staff at the Institute who have made my working time more pleasant: particularly the librarians Chris Gibbons and Diane Collins, the DID Jim White-Cooper, Bernard Clark and formerly Dave Hawkins, the Computing Dept. Roger Harrison, Fiona Wycherley, the canteen staff and Alison Thomson and everyone in Media and Wash-up. I would specifically like to thank Christine Jones for her understanding of my need for isotope and Dr Pete Kaiser for his reference checking, proof reading and bullying.

I would like to say a special thank you to Janene, Thomas and Simon Bumstead for tolerating my somewhat unexpected intrusions into their lives and Marie Quick for encouraging me to start this PhD. Another very special thank you must go to my parents Drs Chris and Bill Wain without whom I would never have got this far.

Finally I am also extremely appreciative of my husband Graham for putting up with my computer problems, long hours away from home and continuous Dial-a-pizzas for dinner.

This thesis is wholly the result of my own work. No part of it has been submitted to any other board for another qualification. The views expressed are those of the author and not of the University.

The HTML contents of this thesis vary from the original publication, due to size contraints.  For any further information please email Hester.

Title page
Abstract
Acknowledgements
Declaration
Contents
List of Tables (removed)
List of Figures (removed)
Abbreviations
Genes and Loci
Chapter One
Introduction
1.1 The Chicken
1.2 Genomic structure and organisation of the chicken
1.2.1 Repetitive Sequences
1.3 Genomic Mapping
1.3.1 Genetic Mapping
1.3.1.1 Restriction Fragment Length Polymorphism
1.3.1.2 Single-Stranded Conformational Polymorphism
1.3.1.3 VNTR-Hypervariable Minisatellites
1.3.1.4 Microsatellites
1.3.1.5 Random Amplified Polymorphic DNA markers
1.3.1.6 Amplified Fragment Length Polymorphism
1.3.1.7 Targeted Mapping
1.3.1.8 Expressed Sequence Tags
1.3.1.9 Subtractive hybridisation
1.3.1.10 RFLP subtraction
1.3.1.11 Differential Display-Reverse Transcription (DD-RT)
1.3.1.12 Genomic Mismatch Scanning
1.3.1.13 Representational Difference Analysis
1.3.2 Physical Mapping
1.3.2.1 Pulse Field Gel Electrophoresis
1.3.2.2 Radiation Hybrids
1.3.2.3 Flow Karyotyping
1.3.2.4 Chromosome microdissection
1.3.2.5 Yeast Artificial Chromosomes (YACs)
1.3.2.6 Sequence Tagged Sites (STS)
1.3.2.7 Fluorescent In-Situ Hybridisation (FISH)
1.4 Traits of interest in the chicken
1.4.1 Production Traits
1.4.2 Disease Traits
1.4.2.1 The major histocompatibility complex (MHC)
1.4.2.2 Salmonellosis
1.4.2.3 Infectious bursal disease virus
1.4.2.4 Marek's disease virus
1.4.2.5 Newcastle disease
1.4.2.6 Infectious bronchitis virus
1.4.2.7 Avian leukosis and sarcoma virus
1.4.2.8 Coccidiosis
1.4.2.9 Fowl cholera
1.5 Objectives
1.6 Approaches
1.6.1 Expressed Sequence Tags (EST)
1.6.2 Random Amplified Polymorphic DNA markers (RAPD)
1.6.3 Representational Difference Analysis
Chapter Two
Materials and Methods
2.1 Materials
2.2 Bacterial strains and plasmids
2.2.1 cDNA Library IAHchB1
2.2.2 Transformation of IAHchB1 by electroporation of
Escherichia coli MC1061/P3
2.2.3 Cloning RDA products into pGEM-T vector
2.3 Chicken Strains and Crosses
2.4 DNA Preparation
2.4.1 Genomic DNA Extraction
2.4.2 Plasmid DNA Preparation-ABI Protocol
2.4.3 Plasmid DNA Preparation-Hybaid maxiprep Protocol
2.5 Restriction Endonuclease Digestion
2.6 Electrophoresis
2.6.1 Agarose Gel Electrophoresis
2.6.2 ABI Genescan electrophoresis
2.6.2.1 Creation of a Matrix File
2.6.2.2 RAPD-PCR Electrophoresis
2.6.3 Sequence Gel Electrophoresis
2.7 PCR Amplification
2.7.1 Plasmid insert PCR Amplification
2.7.2 Colony PCR of RDA products
2.8 Gel purification of PCR products
2.9 Automated Sequencing
2.9.1 Sequence Analysis
2.10 Southern Blotting
2.10.1 Probe preparation and hybridisation
2.10.2 Nick translation system (Gibco BRL)
2.10.3 Prime-It RmT random primer labelling kit (Stratagene)
2.11 Mapping Techniques
2.11.1 Restriction Fragment Length Polymorphism Analysis
2.11.1.1 Mapping Segregation Analyses
2.11.2 RAPD Analysis
2.11.2.1 RAPD PCR Reaction
2.11.2.2 Genescan Analysis
2.12 Representational Difference Analysis
2.12.1 RDA comparison of Line N and line 15I
2.12.1.1 Screening and characterisation of RDA clones
2.12.2 RDA targeted to chromosome 16
2.12.2.1 BamHI Representation
2.12.2.2 TaqI Representation
2.12.2.3 NheI Representation
2.12.3 RDA targeted to Marek's Disease Resistance Genes
Chapter Three
Mapping Expressed Sequence Tags
3.1 Introduction
3.2 Results
3.2.1 Mapping ten novel chicken genes
3.2.2 RFLP Mapping
3.3 Discussion
3.3.1 Analysis of clones from the bursal cDNA library
3.3.2 RFLP Analysis and Mapping of clones
Chapter Four
Random Amplified Polymorphic DNA Analysis
4.1 Introduction
4.1.1 Choice of RAPD Primer
4.2 Results
4.2.1 Creation of the Matrix File
4.2.2 RAPD Analyses using fluorescent primers
4.3 Discussion
Chapter Five
Representational Difference Analysis comparison of line N
and line 15I
5.1 Introduction
5.2 Results
5.2.1 Comparison of Line N and line 15I
5.2.2 Colony PCR of RDA products
5.2.3 Screening and characterisation of RDA clones
5.2.4 Mapping RDA Clones
5.2.5 YAC Hybridisation
5.2.6 Nucleotide Sequences
5.3 Discussion
Chapter Six
Targeted mapping of chromosome 16 by Representational
Difference Analysis
6.1 Introduction
6.2 Results
6.2.1 BamHI Representation
6.2.2 TaqI Representation
6.2.3 NheI Representation
6.2.4 Screening and characterisation of RDA clones
6.2.5 Mapping RDA Clones
6.2.6 Nucleotide Sequence
6.2.7 YAC Hybridisation
6.3 Discussion
Chapter Seven
Resistance to Marek's disease targeted by Representational
Difference Analysis
7.1 Introduction
7.2 Experimental Design
7.3 Results
7.3.1 RDA using BamHI representation
7.3.2 Screening and Characterisation of RDA clones
7.3.3 Mapping RDA Clones
7.3.4 Analysis of association of RDA clones to MD resistance
7.3.5 Segregation of hV32 in the F₂ population
7.4 Discussion
Chapter Eight
General Discussion
8.1 The Chicken
8.2 Genome Mapping
8.3 Saturation Mapping
8.4 Random amplified polymorphic DNA (RAPD) analysis
8.5 Microsatellite Markers
8.6 Amplified fragment length polymorphisms (AFLP)
8.6 EST mapping
8.7 Subtractive Hybridisation
8.8 Differential display-reverse transcription (DD-RT)
8.9 RFLP subtraction
8.10 The RDA Approach
8.11 The Future of Mapping
8.12 Future Work
References
Appendices (removed)

°C degree(s) Celsius
A adenine
aa amino acid
ABI Applied Biosystems (Perkin Elmer)
AceDB A C.elegans database
AFLP Amplified fragment length polymorphism
AIDS Acquired Immunodeficiency syndrome
ALSV Avian leukosis and sarcoma viruses
ALSV-A Subgroup A Avian leukosis and sarcoma viruses
Alu-PCR PCR of inter Alu fragments
ALV Avian leukosis virus
Amp Ampicillin
APC Antigen presenting cell
App. Appendix
APS Ammonium Persulphate
ARS Autonomous Replication Sequence
BCE Before common era
bp base pair
BSA Bovine Serum Albumen
C cytosine
C band chromatin band
cDNA complementary Deoxyribonucleic acid
CH₃COONH₄ Ammonium acetate
CHEF Clamped homogeneous electrophoresis field
cM centi-Morgan(s)
cm centimetre
CMRP Compton mapping reference population
contig contiguous
COS7 Monkey kidney cell line
CpG Cytosine and guanine dinucleotide
CR1 Chicken Repeat 1
Cys Cysteine
dCTP 2'-deoxycytidine triphosphate
DD-RT Differential Display-Reverse Transcriptase
DMF Dimethyl formamide
DNA Deoxyribose nucleic acid
dNTP deoxynucleoside triphosphate
DOP-PCR Degenerate oligo. primed PCR
dsb Double stranded break
dT deoxy-Thymine
DTT dithiothreitol
dUTP 2'-deoxyuridine triphosphate
EDTA Ethylenediaminetetra acetic acid
EE EDTA and EPPS buffer
EPPS N-(2-hydroxyethy)piperazine-N'-(3-propanesulfonic acid)
EMBL European molecular biology laboratory
ES Embryonic stem cell
EST Expressed sequence tag
F₁ First cross
F₂ Intercross between two F₁
FAM 5-carboxyfluorescein (ABI fluorescent label)
Fig Figure
FISH Fluorescent In Situ Hybridisation
g Acceleration due to gravity
G guanine
GCG Genetics computer group
GCN4 DNA binding protein of ds DNA
GDB Genome database
GDRDA Genetically Directed Representational Difference Analysis
GL Suggestive linkage
GMS Genomic mismatch scanning
GTE Glucose, tris, EDTA buffer
HAT Hypoxanthine/aminopterin/thymidine
HepG2 Liver cancer cell line
HEX 6-carboxy-2',4',7'4,7-Hexachlorofluorescein
(ABI fluorescent label)
His Histidine
HPRT Hypoxanthine phosphoribosyltransferase
HSL Highly significant linkage
HVT Herpes virus of turkeys
IAH Institute for Animal Health
IBDV Infectious bursal disease virus
IBV Infectious bronchitis virus
IFGT Irradiation and fusion gene transfer
IPTG Isopropyl-[beta]-D-thiogalactopyranoside
JBAM 2^nd RDA oligo for BamHI derived fragments
JTAQ 2^nd RDA oligo for TaqI derived fragments
kb kilobase
kV kilo volts
l litre
[lambda] Wavelength
L-agar Luria-Agar
LOD Logarithms of odds ratio
LR Long Ranger
LTR Long terminal repeat
M Molar
MAS Marker assisted selection
Mb Megabase
Mbq Mega bequerel
µCi micro Curie
MD Marek's disease
MDV Marek's disease virus
mer Oligomer
mF milli Farad
µg microgram
mg milligram
MgCl₂ Magnesium chloride
MGD Mouse genome database
MHC Major Histocompatiblity Complex
MIC Micro-chromosome
min minute(s)
µl microlitre
ml millilitre
µm micrometer
MPC Biomagnetic separator
MQ Milli-Q pure water
mRNA messenger Ribonucleic Acid
N Normal
NaCl Sodium chloride
NaClO4 Sodium perchlorate
NaOH Sodium hydroxide
NBAM 3^rd RDA oligo for BamHI derived fragments
ng nano gram
nmol nano mole
NOR Nucleolar organiser region
NTAQ 3^rd RDA oligo for TaqI derived fragments)
OD Optical Density
oligo Oligomer
p short arm of chromosome
P Probability
PB Qiagen's PB buffer
PBS Phosphate Buffered Saline
PCR Polymerase Chain Reaction
PCR-ELA PCR-Enzyme Linked Assay
PCR-SSP PCR-Sequence Specific Primer
PEG Polyethylene Glycol
pers. com. Personal communication
PFGE Pulsed field gel electrophoresis
pfu plaque forming units
pg pico grams
PI Post infection
pmol pico moles
Poly A Poly Adenylate
PVP Polyvinylpyrrolidone
q long arm of chromosome
QTL Quantitative Trait Locus
rad Unit of Radiation dose
RAPD Random Amplified Polymorphic DNA
RBAM 1^st RDA oligo for BamHI derived fragments
RDA Representational Difference Analysis
rDNA Ribosomal DNA
RE Restriction Endonuclease
RFLP Restriction Fragment Length Polymorphism
RH Radiation hybrid
RNA Ribonucleic Acid
RNaseA Ribonuclease A
RNHE 1^st RDA oligo for NheI derived fragments
rpm revolutions per minute
RPRL Regional poultry research laboratory
RT Room Temperature
RTAQ 1^st RDA oligo for TaqI derived fragments
SB Special buffer
SDS Sodium Dodecyl Sulphate
sec Second(s)
Seq Ed Sequence editor
SL Significant linkage
ss single stranded
SSC Saline sodium citrate
SSCP Single Stranded Conformational Polymorphism
STS Sequence tagged sites
T Thymine
T_a Annealing Temperature
TAE Tris-acetate EDTA
TAP2 Transporter associated with antigen processing 2
Taq Thermus aquaticus DNA polymerase
TB Terrific broth
TBE Tris-borate EDTA
TE Tris EDTA
T_e Extension Temperature
TEMED N,N,N',N'-tetramethylethylenediamine
Tet Tetracyclin
T_m Melting Temperature
Tris 2-amino-2-(hydromethyl) propane-1, 3 diol
tRNA transfer RNA
U Unit
UK United Kingdom
UTR Untranslated region
UV Ultraviolet
V Volt
VNTR Variable number tandem repeat
vs. versus
[Omega] Ohm
WG-RH Whole genome radiation hybrid
X-Gal 5-Bromo-4-chloro-indoyl-[beta]-D galactoside
YAC Yeast artificial chromosome
YS Yoshida sarcoma

Loci/prefixes
ADL Microsatellite markers from East Lansing
COM Marker generated at IAH Compton
MCW Microsatellite markers from Wageninen
28S 28S ribosomal protein
60S 60S ribosomal protein
B-F Chicken MHC Class I locus
B-L Chicken MHC Class II locus
B-G Chicken MHC Class IV locus
BAT8 (homologue of G9a) function unknown
bcl-x bcl2 family, apoptosis gene
BG32.1 Chicken MHC Class IV locus marker 32.1
Bu-1 antigen expressed on B cells (marker=CHB6)
C4 MHC class III gene
CMYC28 Chicken myelocytomatosis viral oncogene 28
CNBP Cellular nucleic acid binding protein
COL4a1 Collagen type 4, alpha 1 chain
Db Eumuelanin restrictor gene
ERV Endogenous retrovirus
ev21 endogenous virus locus
FN1 fibronectin 1 gene
G6PD Glucose 6 phosphate dehydrogenase gene
G9a (homologue of BAT8)
GDI GDP dissociation inhibitor
GDID4 GDP dissociation inhibitor for the rho (GDI) protein
GDP Guanosine diphosphate
hg high growth gene
HLA Human leukocyte antigen loci A-D
HMG1 High mobility group 1
HPRT Hypoxanthine phosphoribosultransferase
Insr Insulin receptor gene
ITPR2 Inositol tri-phosphate receptor 2
Ity Salmonella typhimurium resistance gene
jcpk Juvenile congenital polycystic kidney disease
Lec Lectin type C gene
Ly-4 receptor gene on lymphocytes
MHCII[beta] Major Histocompatiblity Complex II [beta]
Ml Eumuelanin extension gene
NF2 Neurofibromatosis 2 gene
NOR Nucleolar organiser region
Nramp Natural resistance associated macrophage protein
nude Nude locus (mouse chromosome 1)
Pg Pattern gene
pol polymerase gene
PSF PTB associated splicing factor (marker SFPQ)
PTB Polypyrimidine tract binding
pudgy Pudgy locus
RAG-1 Recombination activation gene 1
RAG-2 Recombination activation gene 2
Rfp-Y Restriction fragment pattern-Y
SFPQ PTB associated splicing factor (PSF)
SRE Sterol regulatory element
TAP2 Transporter associated with antigen processing 2
TET Tetracycline resistance gene
TFRC transferrin receptor gene
Th-1 antigen expressed on T cells
tottering tottering gene
TUBA Tubulin alpha
tv-a receptor locus for ALSV-A
unl unlinked
VIL villin gene
Xic X inactivation centre
Xist X-inactive specific transcripts

1.1 The Chicken
The domestic chicken (Gallus gallus domesticus) originated from south-western Asia as a descendant of the Red Jungle Fowl and was first introduced into China in about 1400 before the common era (BCE). Chickens are also depicted in Babylonian carvings of about 600 BCE and are mentioned by ancient Greek writers, particularly Aristophanes in 400 BCE (reviewed by Stockbridge, 1995). Since that time, chickens were kept in small flocks for home consumption until the 20^th Century when poultry farming became commercialised. A modern poultry farm may contain from several hundred thousand to over a million chickens, either layers for egg production or broilers for the meat industry. The chicken industry is now a multi-million pound business with a gross output value of £1339 million in the UK in 1991 (Law and Payne, 1996), which depends on high-production egg layers, very rapidly growing broilers and disease-free stock. These three factors have been the targets of poultry breeders who have tried to select the best genetic traits in their stock, possibly eradicating some of the main poultry pathogens.
However, disease is still a major cause of loss in production and still proves of great economic concern to poultry farmers, and from the point of view of the birds' welfare is unethical. Control of disease has been improved by good husbandry and the use of vaccines, but, it would be far more efficient if the birds could be bred to be disease resistant. Therefore it would be useful to know the identity of the gene or genes affecting the resistance traits for each disease, as with this detailed information it will be possible to directly select for the required trait. To this end linkage maps of the chicken have been developed, with the intention of generating enough marker loci to identify linkage with the resistance trait, and ultimately by positional cloning to isolate the gene concerned. Once resistance genes have been isolated and characterised it will be possible to use them to screen commercial breeding stock and selectively breed for resistance. Knowledge of the gene may also suggest improved pharmaceutical or immunological therapies.
Chickens are animals of agricultural importance as well as a valuable model organism. Their use as a model organism is due to the many inbred lines of chickens, which are well characterised for disease and production traits. The parent stocks produce large numbers of progeny which eases genetic experimentation. All commercially bred chickens have accurately detailed pedigrees and are of low cost per animal. Because it is possible to produce and assess large progeny families, the chicken is well suited to genome analysis, particularly as the red blood cells are nucleated, facilitating large scale good quality DNA extraction. Chickens have a small genome, about half the size of mammals' but three times as big as that of the pufferfish (Elgar et al., 1996). The chicken genome is therefore more easily accessible to cloning than those of mammalian species, and yet retains sufficient sequence homology to mammals to provide comparisons which could elucidate function. It is not yet clear whether this similarity also applies to gene order between these species.
Chickens can harbour a number of different types of pathogen; viruses, bacteria and parasites, all of which cause major diseases in poultry. A wide variety of resistance mechanisms operate against these pathogens. The identification and characterisation of these resistance genes in chickens could help elucidate basic mechanisms for resistance in other organisms as well as chickens.
1.2 Genomic structure and organisation of the chicken
The genome of the domestic chicken has a haploid number of 39 chromosomes; the ten largest are referred to as macro-chromosomes, and the other 29 are termed micro-chromosomes (MICs) (Yamashina, 1944). In chickens chromosomes have been numbered in size order, the biggest first. The large number of MICs is typical of avian species (Abbott and Yee, 1975). In comparison to man, the first six chromosomes are of similar size, the largest being 8 µm. However the MICs are much smaller (the smallest being about 7 Mb) than the smallest human chromosome, which contains about 50 Mb of DNA (Bloom and Bacon, 1985). There is a size difference of 23 times between the largest and the smallest chromosomes in the chicken. Chickens, like other avian species, differ from mammals in that the female is the heterogametic sex (ZW) and the male is the homogametic sex (ZZ), the Z and W chromosomes displaying heteromorphism. The chicken chromosomes are mostly euchromatic with the exceptions of a large terminal C-band (chromatin-band) on the Z chromosome and an almost totally heterochromatic W chromosome, with small C-bands on most of the MICs (reviewed by Fecheimer (1990)). The chicken genome is relatively small, about 1.2 x 10⁹ bases (Olofsson and Bernardi, 1983), less than half that of the mouse and human genomes. This makes the chicken's genomic structure and organisation particularly interesting, as evolution appears to have pruned the genome to a minimal size. Alternatively it is possible that mammalian genomes have expanded in the 300 million years since splitting from the avian lineage. However whatever the mechanism of change between the two lineages, the reason for this difference is still unknown. One theory is that small genome size was favoured by directional selection in birds in order to cope with the metabolic demands of flight (Hughes and Hughes, 1995). However, one aspect of the chicken's unique genome is the relative paucity of repetitive sequences.
1.2.1 Repetitive Sequences
The chicken genome like that of other animals contains repetitive sequences. In the genomes of many animals, for example the amphibian Xenopus, there is a short interspersion pattern of 0.3 kb repetitive sequences with 2 kb single copy sequences (Davidson et al., 1973). However, the genomes of avian species show organisation similar to that of the long period interspersion pattern of the Drosophila genome (Crain et al., 1976). Estimates vary as to the degree of repetition in the chicken genome. Eden and Hendrick (1978) concluded that 87% was single copy and 13% repetitive sequences, 42% of which contained ten to twenty copies and 58% contained about 1500 copies. Another study by Epplen et al. (1978) using reassociation kinetics obtained similar results showing that the chicken genome contained 15% highly repetitive sequences, 10% intermediate repetitive sequences and 70% single copy sequences, of which 28% are 2.3 kb single copy DNA pieces interspersed with 1.5 kb long repetitive fragments. Olofsson and Bernardi (1983) fractionated the genome using density gradient centrifugation and similarly concluded that the genome was composed of 84% unique sequences and 13% repetitive sequences. Arthur and Straus (1983), however, concluded that the genome contained 34% of unique sequences interspersed with repeated sequences, 38% of long stretches of unique sequences and 19% of foldback elements. Sequences complementary to mRNA were found to be randomly distributed with respect to the interspersion patterns, implying that the distribution of the repetitive sequences does not parallel that of the structural genes.
The only characterised chicken repeat sequences identified so far are the chicken repeat 1 (CR1) elements, which are middle repetitive sequences discovered by Stumph et al. (1981). CR1 elements are an ancient group showing six sub-families in chickens, with four sharing a common progenitor (Vandergon and Reitman, 1994). Over 95 CR1 elements have been identified, of an average length of 300 bp, the largest being 2.3 kb (Burch et al., 1993). They have truncated 5' ends and a consensus 3' end containing two or more repeats of an eight nucleotide sequence. CR1 elements are non-long terminal repeat (LTR) retrotransposons (Burch et al., 1993), with a pol-like open reading frame encoding reverse transcriptase which is responsible for dispersal of the element throughout the genome. There are an estimated 100,000 copies of CR1 throughout the chicken genome (Vandergon and Reitman, 1994). CR1 elements are also represented in other avian genomes. Nine orders have so far been studied all of which contain CR1 elements (Chen et al., 1991). All CR1 elements have two regions of high homology (Stumph et al., 1984), which extend to other avian species. These regions appear to be conserved, the first coding for a silencer domain and the second for a nuclear protein binding domain located at the 3' end of the element, probably involved in transposition (Chen et al., 1991).
Reptiles have also been shown to possess CR1 elements (Vandergon and Reitman, 1994), showing that CR1 elements existed before the divergence of birds and reptiles. CR1 elements are not randomly distributed but occur mostly in G-C rich regions of the genome (Olofsson and Bernardi, 1983). These regions are also the most gene-rich and at least 16% of the [beta]-globin gene cluster contains CR1 elements (Reitman et al., 1993). CR1 elements are associated with a number of genes, particularly near transition regions of chromatin structure, for example the ovalbumin gene (Stumph et al., 1983).
In mammals repeat elements are highly useful as, like the Alu repeats of man, they can be used to characterise the genome by hybridisation or PCR. PCR amplification of inter-Alu fragments (Alu-PCR) has been used for high resolution physical mapping of radiation hybrids, cosmids and yeast artificial chromosomes (Monaco et al., 1991; Aburatani et al., 1996). However, CR1 repeats in chickens are less common and also much less conserved than the repeats in mammals and it seems unlikely that they can be used for these purposes. Similarly although chickens possess simple microsatellite repeats, there are far fewer of these than in humans and mice. For example in chickens estimates of the numbers of (AC)n repeats range as low as 7,000, almost 10 fold less than in mammals and they appear not to be uniformly distributed throughout the genome (Toye, 1993).
1.3 Genomic Mapping
The first published genetic map was of the Drosophila X chromosome (Sturtevant, 1913). Since then, genome mapping technology has greatly improved and genetic maps for over one hundred species have been generated.
Two types of genomic maps have been established for many animals. Linkage maps are determined from the frequency of meiotic exchange among linked polymorphic markers in sexual crosses. Physical maps relate traits to actual chromosomal location. Ultimately, the aim is to merge these two maps into one giving both physical and linkage data.
1.3.1 Genetic Mapping
Genetic or linkage maps are necessary for the location of genes responsible for monogenic and polygenic traits. Linkage analysis can be applied to backcrosses of the same breed, or interspecies crosses can be used to generate more polymorphisms. The murine genome has been analysed using interspecific crosses (Avner et al., 1988). The advantages of interspecific crosses are that they maximise the number of possible genetic polymorphisms, and that pedigree analysis can be aligned with linkage analysis in order to place loci on a particular chromosome (Copeland and Jenkins, 1991).
In chickens, genome mapping was first achieved by the analysis of phenotype and limited maps established (Hutt, 1949). The first genetically studied crosses were generated by Bateson and Saunders (1902) but it was not until 1936 that Hutt generated the first chicken linkage map. Chickens of differing phenotypes were crossed and the linkage between different genes established. For example, the eumelanin extension gene (Ml) lies between the loci of the eumelanin restrictor (Db) and the pattern gene (Pg) (Carefoot, 1987). This system is inherently limited as these markers require expression as phenotypes and are only present in rare lines of chickens. These morphological polymorphisms are also infrequently observed, as only one or two occur in any one line. This makes segregation analysis difficult as no single population is capable of giving scores for each polymorphism, thus leading to ambiguity in the results.
Bumstead and Palyga (1992) published the first map of the chicken genome derived from Restriction Fragment Length Polymorphism (RFLP) analysis.
68 White leghorn progeny were generated from a backcross between an F₁ (Line 15I x line N) female and her male line 15I parent. An initial map of 100 markers was developed from segregation analyses, covering a minimum of 585 cM of the chicken genome and identifying 18 linkage groups.
More recently, Crittenden et al. (1993) crossed a Red Jungle Fowl and a White Leghorn in order to maximise the chance of finding DNA polymorphisms in the backcross. It was determined that RFLP were observed between the parental lines with in the order of 91% of those enzymes used, compared to 50% in the inbred White Leghorn lines used by Bumstead and Palyga (1992). This gave a polymorphic population providing a good resource for collaborative mapping. A variety of marker types were used, including linkage mapping of feather colour and blood group loci, RFLP analysis and RAPD (Random Amplified Polymorphic DNA) analysis. Further analysis of this intercross has provided a linkage map (Levin et al., 1994b). Recent comparisons of the Compton and East Lansing maps has shown that there is in fact greater recombination in the Compton population, containing 2700 cM in comparison to 2100 cM in the East Lansing population (Mariani et al., In Press). Within the Compton Linkage map there are over 100 genes and 300 other loci, and now fewer than 2% of markers do not show linkage (of LOD>3.0).
There are two main strategies for genetic mapping; saturation or random mapping and targeted mapping. Saturation or random mapping entails the generation of large numbers of markers in an attempt to cover the whole genome densely, thus ensuring that any trait of interest must lie near to at least one marker. Examples of techniques which achieve this are: microsatellite analysis, RAPD STS, AFLP, RFLP and SSCP. Targeted mapping is the isolation of markers in a specific region, or for a specific trait. This is achieved by a variety of methodologies including EST, which can be targeted by the isolation of cDNAs from specific tissues, as well as more refined methods of subtractive hybridisation and representational difference analysis.
Markers generated by these methods can be used for comparative mapping, but are defined as two classes based on their polymorphic nature. Comparative mapping is the positioning of identical markers on maps of different species or strains. These markers are termed anchor or index loci and are mostly used to define new linkage maps. Type I anchor loci are coding sequences which are relatively conserved across different species. However, this sometimes means that they contain no identifiable polymorphisms. So, although they are particularly useful in interspecific mapping for the identification of genes they cannot always be used. Type II anchor loci are defined as high resolution polymorphisms. These are typified by microsatellite and minisatellite polymorphisms which have traditionally not been used as loci for interspecific mapping, due to their non-coding nature. However, O'Brien and Graves (1991) have shown that these type II anchor loci can also play an important role in the identification of syntenic regions between maps of different species.
1.3.1.1 Restriction Fragment Length Polymorphism
RFLPs are generated by the digestion of DNA with restriction endonucleases. RFLP was used by Donis-Keller et al. (1987) in the construction of the first genetic linkage map of the human genome. RFLP are often visualised by hybridisation of radioactive cDNA probes to the digested genomic DNA fragments. RFLP analysis of the chicken B-F and B-L genes has been used to confirm serological B-typing (Juul-Madsen et al., 1993).
There are disadvantages with RFLP analysis, as often a number of probe and enzyme combinations have to be tested in order to generate significant numbers of RFLP. This method generally uses radioactive probes, and requires a large amount of target DNA. In fact, RFLPs detect about 1 in 10000 polymorphic nucleotides in the human genome (Soller and Beckmann, 1986). The problem with searching for direct associations between RFLP and traits of economic value is the low likelihood of finding one; at best 1:200, but probably 1:20000 (Soller and Beckmann, 1986). RFLP markers are also difficult to transfer between different linkage maps, as what is polymorphic in one population may well not be in another. This, however, is true for any randomly selected type of marker, but RFLPs do generate a reasonable number of markers which can be easily placed onto a map. This technique has been extensively employed by Bumstead and Palyga (1992) to place new markers on the chicken linkage map.
1.3.1.2 Single-Stranded Conformational Polymorphism
SSCP involves the digestion of genomic DNA by restriction endonucleases, denaturation with NaOH and EDTA and electrophoresis on a non-denaturing polyacrylamide gel. Polymorphisms are detected as a mobility shift of the single stranded DNA, due to conformational changes in alternative secondary structures (Nakabayashi and Nishigaki, 1996). Even single-base substitutions can be detected as shifts in electrophoretic mobility (Orita et al., 1989a). This technique has been successfully used to detect three point mutations in highly conserved residues found in mildly-affected cystic fibrosis patients (Dean et al., 1990).
SSCP-PCR is a development of this technique, where the sequences to be examined are amplified by PCR using labelled primers for the clones of interest; these could include fluorescently-labelled primers. The samples are then denatured and analysed by gel electrophoresis (Orita et al., 1989b). This technique is usually used for genetic screening, as it requires sequence data for the design of the primers. However, if RFLP analysis shows no polymorphism with a particular clone, this method can be used once the genomic sequence has been determined permitting design of primers to the most variable region possible, usually introns or 3' UTR. But it is costly in primer synthesis and often many primer pairs have to be tested before a polymorphism is detected.
1.3.1.3 VNTR-Hypervariable Minisatellites
Minisatellites or VNTR (Variable number tandem repeats) are regions of DNA containing simple tandem repeats of 10-15 bp segments. They are detected by hybridisation of a probe to a core sequence within the repeat. This generates a multi-band pattern or fingerprint which can vary between individuals (Jeffreys et al., 1985). These band patterns are analysed by radioactive labelling of the products and electrophoresis on a polyacrylamide gel. This technique is mainly used in forensic and paternity investigations but can also be used for mapping. Fingerprinting usually detects a large number of loci from a single probing but these loci are mostly unlinked and scattered throughout the chicken genome; however, hybridisation with cloned single locus minisatellites shows that some could be clustered (Bruford et al., 1994). However, this necessitates knowledge of the core sequence and the identification of polymorphisms between the parents of the cross before it can be used for linkage mapping. VNTR analysis is also time consuming and involves the use of large amounts of DNA and radioactive label.
1.3.1.4 Microsatellites
Microsatellite markers consist of di-, tri-, or tetra-nucleotides repeated up to 40 times. Oligonucleotides are synthesised to sequences flanking these repeats and can be used in PCR, to amplify loci from genomic DNA (Holmes et al., 1993). Microsatellites are now being used as markers for linkage analysis as they can be highly polymorphic, polyallelic, very abundant and usually behave as single locus markers (Weber, 1990). Disadvantages include the large amount of labour required and initial expense of the oligonucleotide primers necessary to identify enough microsatellite markers (Rafalski and Tingey, 1993). Microsatellites are useful type II anchor loci as they are often polymorphic between more than one linkage cross. However, within chickens less than 50% of microsatellites are polymorphic between two inbred lines (Khatib et al., 1993). This is particularly noticeable in the Compton mapping reference population where microsatellite polymorphisms occur in less than 40% of primers (Bumstead pers. com.). However, many microsatellites have been used in the chicken linkage maps and recently a third map has been generated, based entirely on microsatellites (Crooijmans et al., 1996). With increased use of fluorescence technology, using the ABI Genescan system, many fluorescent primers can be used in each PCR reaction and the products separated by electrophoresis. While re-loading gels three or four times greatly increases the productivity, cloning and screening the initial microsatellite is still a time-consuming process.
1.3.1.5 Random Amplified Polymorphic DNA markers
RAPD markers can be generated using short arbitrary primers to amplify genomic DNA, giving a genotype-specific pattern of bands. This is discussed later in section 1.6.2.
1.3.1.6 Amplified Fragment Length Polymorphism
AFLP is based on selective amplification of digested genomic DNA by a series of extended primers. The genomic DNA is digested with two enzymes, a frequent and a rare cutter, e.g. MseI and EcoRI. Adaptors, with a core sequence and the appropriate overhangs for the enzymes used are ligated to the ends of the fragments. Pre-amplification of these fragments by PCR is achieved using primers with a core sequence, an enzyme specific sequence and a random-extension nucleotide. This generates a pool of fragments which are re-amplified by two primers of core sequence, enzyme specific sequence and this time further random-extension nucleotides. The primer with the rare cutting enzyme sequence is end-labelled before the reaction so that those fragments only will be detected by electrophoresis and autoradiography (Vos et al., 1995). Fragments generated by MseI digestion will be small enough to amplify by PCR, while those with one MseI and one EcoRI end will be much rarer but will also easily amplify. During the second amplification, further selection of fragments occurs using the extension nucleotides and finally only those fragments generated by the EcoRI-labelled primer will be resolved visually by autoradiography following polyacrylamide gel electrophoresis. The use of two enzymes also enables the AFLP reaction to be fine-tuned for the production of the maximum number of bands. This technique was originally developed by plant geneticists but can easily be transferred to the analysis of the chicken genome. The development of many polymorphic bands for each reaction should enable large numbers of marker loci to be mapped for a relatively small amount of work and input DNA. However, all these markers will be anonymous, and to clone one requires the extraction of DNA from a gel matrix. This method can be visualised by silver staining rather than by radioactive labelling, but the complexity would not then be further reduced, and might need a further PCR step, which will ultimately be more expensive.
1.3.1.7 Targeted Mapping
Recently, a number of strategies have been directed towards the isolation of candidate genes. In comparison with the saturation mapping approach, these techniques would not necessarily place other informative loci on the map, but would significantly reduce the amount of work needed to isolate a candidate gene.
1.3.1.8 Expressed Sequence tags
EST are partial sequences of cDNA clones. These will be discussed later in section 1.6.1.
1.3.1.9 Subtractive hybridisation
Subtractive hybridisation has been used to find candidate tumour suppressor genes. This entails the positive selection of cDNA expressed only in non-tumour cells of interest, when radioactively-labelled cDNA is hybridised with tumour cell mRNA. Unhybridised labelled cDNA can be recovered and used as a probe to screen a cDNA library constructed from non-tumour cells. New suppressor genes were identified in this way (Lee et al., 1991). This method also takes up considerably more time than that of, for example, differential display reverse transcriptase (DD-RT) see below, although, in comparison, it does have the advantage of being a positive selection technique. However, it does rely on the selection of a small amount of unique DNA from a large amount of initial DNA. For this technique to work effectively for the isolation of candidate genes, the pattern of tissue expression must be known for each gene. This is very difficult to achieve with anything but tumours or highly expressed tissue specific genes.
1.3.1.10 RFLP subtraction
RFLP subtraction (Rosenberg et al., 1994) purifies small restriction fragments from one genome (the tracer) from sequences which reside on fragments in a related genome (the driver). The genomic DNA is digested with HindIII, and then the driver is ligated to biotinylated adaptors, whilst the tracer is ligated to unlabelled adaptors. Both are PCR-amplified separately and then mixed together, the driver in 100 times excess of the tracer. After hybridisation, the biotinylated driver and its heteroduplexes are removed by streptavidin-coated beads, leaving the tracer-specific DNA. This technique can be used to generate DNA for library construction or to obtain markers near to a gene of interest. However, only tiny amounts of DNA will be generated and small under-represented sequences may well be lost as there is no enrichment. Sequences of interest may also be lost in the removal of heteroduplexes from the mixture as the tracer-specific DNA is not positively selected.
1.3.1.11 Differential Display-Reverse Transcription (DD-RT)
DD-RT is a method of analysing changes in gene expression in cells at different stages of differentiation, thus isolating genes expressed at a particular stage (Bauer et al., 1993). The mRNA is systematically amplified and the 3' termini visualised by electrophoresis on a denaturing polyacrylamide gel. An anchored oligo dT primer is used to anneal the mRNA polyA tails, in conjunction with a decamer oligodeoxynucleotide in solution. The decamer is arbitrary in sequence, so that it can anneal at different positions in relation to the first primer for PCR amplification. Thus, when the RNAs from two or more relevant cell types are compared, a small fraction of differentially expressed mRNAs are revealed. Probes can be made from the products, and the genes isolated by genomic library screening (Liang et al., 1993).
DD-RT is useful as it displays subsets of mRNA as short cDNA bands, enabling visualisation of cell mRNA composition. Alterations in gene expression can be simply detected between parallel samples as variable band patterns. The cDNA can easily be sequenced and compared to sequences in databases, or used for probing (Liang and Pardee, 1992). This technique has the advantage of being quick to apply. It is also non-selective, isolating all genes present at any given stage. Again this technique relies on knowledge of tissue expression patterns and times during which the gene of interest could be isolated, which may not be known or fully understood. For the isolation of genes associated with development during particular stages this technique can be very useful, but as a general technique it is too complex and technically demanding.
1.3.1.12 Genomic Mismatch Scanning
Genomic mismatch scanning (GMS) enables mapping in two related individuals. The genomic DNA of one individual only is methylated at GATC sites, while the other remains untreated. The two samples of DNA are then denatured, mixed, and hybridised in solution. The samples are digested at fully methylated and fully unmethylated sites leaving only whole heterohybrids. Base-mismatch hybrids are nicked using a single mismatch repair system. The DNA is then degraded by digestion, leaving mismatch-free heterohybrids. These hybrids should be from regions of identity by descent. This DNA can be labelled and used for probing DNA from the whole genome, yielding positive results with identity by descent and negative results at sites of meiotic recombination (Nelson et al., 1993). This methodology is very prone to failure as there are many separate enzymatic steps, which must initially be optimised in order to generate the complete methylation of one genome. GMS is performed on the whole genome, without reducing its complexity, thus at hybridisation not all homoduplexes will have time to reanneal. The under-represented sequences which differ between the two genomes are also unlikely to be detected as there is no enrichment for target sequences.
1.3.1.13 Representational Difference Analysis
RDA is a new technique developed by Lisitsyn et al. (1993) and will be discussed later in section 1.6.3.
1.3.2 Physical Mapping
Physical mapping relates genes and markers to their actual location on a specific chromosome. Physical mapping of chicken genes is difficult as cytological discrimination between the 29 MICs is difficult on chromosome spreads. Physical mapping in other species is now achieved using techniques such as fluorescent in situ hybridisation (FISH) of single copy molecular clones to metaphase chromosomes (Korenberg et al., 1992).
1.3.2.1 Pulse Field Gel Electrophoresis
PFGE is a technique for the separation of large segments of DNA, up to 1 Mb, using alternating changes in the angle of an electric field to separate the DNA on an agarose gel (Schwartz and Cantor, 1984). There are now many variations of the basic concept, the most common being CHEF (Clamped homogeneous field electrophoresis), in which the agarose gel is surrounded with a series of electrodes (Lognonne, 1993). Modern systems can resolve fragments from 10 kb to 10 Mb, helping to close the gap between the standard cloning systems in plasmid, phage and cosmid vectors and genetic linkage maps. However, PFGE still cannot resolve the chromosomes of larger genomes like the chicken, as even the smallest MIC is over 7 Mb in size (Bloom et al., 1993). Since the development of yeast artificial chromosome clones, however, PFGE has become essential to all mapping labs.
1.3.2.2 Radiation Hybrids
Initially hybrids were made by fusion of somatic cells with either single or total alien chromosomes. However, a more useful method of irradiation and fusion gene transfer (IFGT) was initiated by Goss and Harris (1975) to produce a resource for mapping genes, and was further refined by Cox et al. (1990). Goss and Harris (1975) discovered that genes could be rescued from cells killed by X-rays if the dose was increased so that double stranded breaks (dsb) occurred along the chromosome. A dose of 1500 rad is adequate to kill the cells whilst 4000 rad was sufficient to give a dsb between two loci. Cox et al. (1990) applied this technique to the production of radiation hybrids (RH) from a Chinese hamster-human somatic cell hybrid, containing a single copy of human chromosome 21. The somatic cell hybrid was irradiated with 8000 rad to kill the cell and produce dsb and then rescued by non-irradiated enzyme-deficient hamster cells. The dsb recombine with the chromosomes in the healthy cell to give RH clones containing fragments of human chromosome 21.
Walter et al. (1994) took this technique further to develop whole genome radiation hybrids (WG-RH). Human fibroblasts were irradiated with 3000 rad and rescued by thymidine kinase-deficient hamster cells. These WG-RH were grown in HAT (hypoxanthine/aminopterin/thymidine) medium to select for the human chromosomes and were initially analysed for the presence of chromosome 14 markers by FISH. In the forty-four cell lines produced, a marker retention of 20-50% was attained, so in theory a set of about 100 WG-RH should provide good coverage of the whole genome. WG-RH are a powerful resource, as they enable the mapping of YACs and ESTs which are non-polymorphic within the linkage map. The amount of recombination can be controlled by the radiation dosage (rad), as increased dose leads to increased dsb and therefore increased recombination. However, good coverage of the chicken genome in WG-RH might be difficult, again due to the large numbers of MICs which need large doses of radiation to induce enough dsb for significant recombination. It is possible that some of the MICs would remain whole within the RH clones and would quickly disappear. There is an added complication in that the RH clones will eject the alien chromosomes fairly quickly, so once made aliquots of each clone should be stored frozen while large amounts of DNA are prepared as a future resource.
1.3.2.3 Flow Karyotyping
Flow karyotyping is a technique used to separate individual chromosomes. A large number of chromosomes are isolated as a suspension, and stained with fluorescent dyes. The chromosomes are passed singly through a laser beam and sorted by their fluorescence pattern. The fluorescence pattern depends on DNA content, as the different dyes have differing affinities for certain sequences. The dye Hoechst 33258 binds preferentially to A/T sequences, whilst chromomycin A3 prefers G/C regions. This technique is very sensitive and has to be carefully optimised, only using the most pure chromosomes in order to prevent loss of resolution (Gray et al., 1975). Once single chromosomes are obtained they can be used as a resource for PCR amplification, cloning into single chromosome libraries and direct selection of yeast artificial chromosome clones (YACs) into single chromosome pools. However, although flow karyotyping has been used with great success for man and plants, the large numbers of MICs in the chicken genome prevent resolution beyond the first ten chromosomes. Flow cytometry can also be highly variable and needs to be verified by conventional cytogenetic methods, which can be problematical in the chicken.
1.3.2.4 Chromosome microdissection
Chromosome microdissection has been achieved with improved laser technology (Greulich, 1992). Chromosomes are held in an optical trap, whilst they are dissected by laser microbeam, and then removed by glass needle. This has the advantage of removing a specific piece of a specific chromosome. Microdissected fragments could be used in the construction of a contiguous YAC library.
Unfortunately, there is little information to aid selection of the correct chromosome in the chicken, as after the ten largest chromosomes the MICs are cytologically indistinguishable. This technique yields miniscule amounts of DNA (a few hundred femtograms to a picogram) which makes cloning of the dissected fragments difficult. However, PCR amplification of the fragments by DOP-PCR (Degenerate oligonucleotide primed-PCR) in a general and non-specific manner, using 22-mer degenerate primers in a dual annealing temperature protocol, can generate enough DNA for cloning the dissected chromosome (Arumuganathan et al., 1994). The actual region of the dissected fragment is still not that small, in the order of tens of megabases, which still leaves a significant gap from the 700 kb sequences achievable by conventional molecular techniques. This technique can generate enough DNA to be used as a probe for FISH to specific chromosomes.
1.3.2.5 Yeast Artificial Chromosomes (YACs)
YACs are constructed by inserting yeast chromosomal fragments into plasmid vectors. These yeast fragments include an autonomous replication sequence (ARS), centromere and telomeres, as well as selectable markers. These constructs are linearised and transformed into yeast cells where they replicate as synthetic chromosomes (Burke et al., 1987). Large fragments of exogenous DNA up to 2 Mb in length (Little, 1992) can be cloned faithfully into these vectors, although on average in the available chicken library they contain about 630 kb of inserted DNA (Toye et al., In Press).
YACs have been used in many cases to isolate genes, for example the Duchenne muscular dystrophy gene (Kunkel et al., 1985), where the gene is greater than 10⁶ bp in length. YACs have also been used for fine mapping around the murine Xist (X-inactive specific transcripts) sequence, which could not be achieved by interspecies backcross mapping (Heard et al., 1993). YAC clones for the Xic (X inactivation centre) region were isolated from the total genomic DNA library by probing with Xist. These isolates were then mapped by RFLP, giving localised marker loci, spanning a 1 Mb region. Construction of a chicken YAC library has been completed and this library is now being characterised in this laboratory (Toye et al., In Press).
Characterisation of YAC clones initially involves their alignment into a contig of a particular chromosomal region using hybridisation or PCR amplification of markers known to be present in that region. After verification of these results, end cloning of the YAC clones and their subsequent hybridisation can be used to check the contig alignment. Once a panel of YAC clones have been identified, these are subject to a number of techniques to refine the search for the candidate gene. This is initially achieved by targeting coding regions using CpG island mapping and exon trapping. CpG islands are short unmethylated CpG-rich sequences often positioned at the 5' end of vertebrate genes and can be identified by restriction mapping. They have been shown to be present in 14 out of 20 chicken genes selected at random from the EMBL DNA database (McQueen et al., 1996). Exon trapping utilises RNA splice sites to remove exons and insert them into vectors containing an intron. These constructs are transfected into COS7 cells and if the original splice sites can pair with those in the intron, mRNA is produced. This is then PCR amplified and cloned to isolate the original exon which can be investigated more fully (Buckler et al., 1991).
YACs are very important as cloning vectors because they can also be used for germ-line transmission of DNA into mouse embryonic stem (ES) cells. YAC clones containing the human hypoxanthine phosphoribosyltransferase (HPRT) gene (size 670 kb) were fused as yeast spheroplasts with an HPRT-deficient ES cell line and selected in HAT medium for the expression of HPRT. These transformed ES cells were then injected into mouse blastocysts and the subsequent chimaeric mice were shown to express the human HPRT gene (Jakobovits et al., 1993). Copy number dependent and position independent expression has also been shown by expression of the YAC derived tyrosinase gene in transgenic mice (Schedl et al., 1993). This technique has been adapted by Choi et al. (1993), by co-lipofection of the YAC clone into the ES cells in order to reduce the amount of yeast chromosomal DNA being introduced. Choi et al. (1993) also succeeded in creating chimaeric mice with a human heavy chain immunoglobulin gene which should lead to mice producing fully human monoclonal antibodies, a significant improvement on current technologies.
The use of YAC clones to produce transgenic animals has a further advantage as gene mutations can be engineered within the YAC by homologous recombination to create deletion mutants or knockouts. Constructs are designed, containing the mutant gene with an antibiotic resistance gene inserted in the middle, edged with antibiotic susceptibility genes. Once introduced into the yeast this construct will insert into the YAC. If homologous recombination has taken place the mutated gene with the antibiotic resistance will have inserted in the place of the original gene. In contrast, randomly inserted genes would still carry the ends encoding antibiotic susceptibility.
1.3.2.6 Sequence Tagged Sites (STS)
STS were proposed Olson et al. (1989) as an ideal way of integrating all markers into physical maps, by making sets of PCR-based markers. STS are 200-500 bp tracts of unique single-copy sequence which can be amplified by two 20-mer primers. Details of the primers, reaction conditions and product are available to all researchers, so any DNA clone can be tested for STS content. STS are particularly useful in the construction of YAC contigs, although they cannot be relied upon in isolation (reviewed by Ward and Davies, (1993)). However, Olson's original concept of an STS map, with a coverage of one marker per 100 kb of the human genome by 1995, has proved unrealistic. STS provide a relatively expensive procedure as they rely on the production of primer pairs and the sequencing of large numbers of clones. Although STS give some information about the genomic sequences they are not useful as anchor loci, and are only really useful in the characterisation of YAC or cosmid clones.
STS are currently being employed for sequence scanning in the pufferfish (Fugu rubripes) genome. This entails random sequencing of the well characterised cosmid library to give at least 50 STS per cosmid. The small size of this genome, 400-500 Mb means that there are a maximum of 8 genes per cosmid. This approach should lead to STS which are within the genes, and when compared to other genome databases these genes will be identified by homology. Once genes are identified, conservation of synteny can be examined (Elgar et al., 1996). Although the Fugu genome is considerably smaller than those of other vertebrates, including the chicken, it should still contain the same number of genes, estimated at 100,000. Thus randomly sequencing this genome should lead to the isolation of many genes, and comparative mapping will enable these to be mapped in other organisms.
1.3.2.7 Fluorescent In-Situ Hybridisation (FISH)
FISH enables the localisation of regions of DNA to a particular chromosome or chromosomes in metaphase or interphase cells following their detection by fluorescent label. Probes of any size can be used, from cDNAs or cosmids to YACs, but the bigger the clone the more fluorescence can be detected. The probe is usually labelled with biotin-16-dUTP and hybridised overnight to pre-treated slides of chromosome spreads. After washing at the required stringency, fluorescein-avidin is added and after a further washing biotinylated anti-avidin is added. This acts as a FISH "sandwich" to increase the number of binding sites for fluorescein-avidin, which is again added before the slides are sealed. The slides are viewed by fluorescence microscopy and the results photographed (Korenberg et al., 1992). A single copy gene should hybridise to both chromatids of an homologous pair of chromosomes, usually visualised as two fluorescent dots. However it is difficult to obtain good data from this technique in birds, due to the large number of chromosomes and the small amount of spread achieved with the metaphase chromosomes. It is also impossible to cytologically distinguish between the 29 pairs of MICs, so if a FISH probe hybridises to MIC, there is no way of deciding which chromosome it is. However, this technique is very useful for determining the chimaerism of the YAC libraries, and is being employed in our laboratory for this purpose. FISH can also show to which size of chromosome the probe hybridises.
1.4 Traits of interest in the chicken
Chickens exhibit a number of traits differing between the commercial lines available. These include associations of particular lines with increased productivity and disease resistance. This gives an initial starting point for the selection of birds in which a trait of interest is expressed.
1.4.1 Production Traits
The main production traits of interest in chickens are rapid weight gain and increased egg-lay. To this end, two types of commercial chickens are produced, broilers and egg-layers. These traits are often polygenic but may also be affected by other non-hereditary factors, which makes the identification of genes difficult. In order to map the genes controlling production traits, markers which are linked to the quantitative trait, termed Quantitative Trait Loci (QTL), are identified within a large population (about 1000 per sire). Quantitative traits are measurements of a phenotypic characteristic such as body weight, shank length (Dunnington et al., 1993) or abdominal fat deposition (Plotsky et al., 1993). The marker loci used are often derived from DNA fingerprints which are randomly distributed throughout the genome. QTLs can subsequently be used directly in the screening of commercial flocks for marker assisted selection (MAS), as they segregate with the trait of interest. The identification of genes controlling these complex production traits requires very large numbers of progeny and the QTL linkage is only as good as the number of markers in those regions. The final outcome of QTL mapping is usually identification of a number of highly significant areas of chromosomes which then need to be studied in greater detail using YAC contigs to identify putative candidate genes, for the trait under investigation.
1.4.2 Disease Traits
Chickens suffer from a number of diseases, some of which are fatal, all of which affect the growth, production and welfare of the birds. Thus disease resistance is an important economical consideration to both the poultry farmer and breeder. The most important diseases of chickens are those for which vaccination is now proving useless against very virulent strains and those which could affect man. Resistance to disease is fundamental to survival for all organisms. However, two species will not necessarily share the same mechanism of resistance to the same disease. Disease resistance will vary across a population, and is particularly clear in comparisons between the inbred chicken lines at the IAH, which have been studied by Bumstead et al. (1991). There are three basic mechanisms of genetic resistance: recognition of the pathogen by the immune system, under the control of the MHC complex; metabolic variants which affect the penetration or reproduction of the pathogen; and incorporation of pathogen DNA into the host genome which prevents infection.
1.4.2.1 The major histocompatibility complex (MHC)
A major element of disease resistance is the major histocompatibility complex (MHC) which in chickens lies on chromosome 16. Chromosome 16 is a MIC which has been shown by silver-staining to contain the nucleolar organiser region (NOR) containing the 5.8S, 18S and 28S ribosomal RNA genes (rDNA) for protein synthesis. Chickens only have one copy of rDNA consisting of a cluster of 145 40 kb repeats, which represents about 0.5% of the chicken genome (Bloom et al., 1993). The NOR was shown to exist on the same chromosome as the chicken MHC by trisomy mapping of a bird expressing B⁶, B¹³ and B¹⁵ antigens (Bloom and Bacon, 1985). Estimates of the size of the microchromosome range from 8 Mb upwards; however, the NOR occupies a large portion of the chromosome, perhaps as much as 6 Mb (Bloom and Bacon, 1985). The MHC region is therefore perhaps 2 Mb or 0.17% of the genome.
The MHC B complex in chickens was first identified by the ability of leukocytes to give strong graft rejection (Schierman and Nordskog, 1961). It was subsequently discovered to contain three loci, B-F, B-L and B-G (Pink et al., 1977), producing class I, II and IV antigens respectively. It has been shown that frequent recombination occurs between B-F and B-G regions, although none has been detected between B-F and B-L (Hála et al., 1976). Class I antigens are expressed on the surface of virtually all cells including leukocytes and erythrocytes, as a single transmembrane polypeptide chain folded into three extracellular domains associated non-covalently with [beta]₂ -microglobulin. Class I presents antigens for recognition by cytotoxic T cells. This region is homologous to HLA-A, B, and C genes in man. Class II antigens are expressed on antigen presenting cells (APC), B-cells and macrophages, as heterodimers with two non-variable immunoglobulin-like domains near the membrane and two variable domains, furthest from the cell. The class II antigen is recognised by T helper cells, which become activated to stimulate T cell proliferation. This region is homologous to the HLA-D genes in man. Class IV antigens are expressed on the cell surface of erythrocytes and are unique to avian species. They consist of two glycoproteins which are highly polymorphic, but of unknown function. However, they are distantly related to butyrophilin and myelin-oligodendrocyte glycoprotein (Kaufman et al., 1995). At present, no regions have been identified for class V loci (Sander, 1993). MHC haplotypes have traditionally been determined by serotyping each bird by haemagglutination of the B-G antigens (and in some cases B-F antigens), but give no information for the other MHC loci.
The chicken MHC region has now been more extensively explored and there are chromosome 16 linkage maps for both the Compton mapping reference population (Bumstead and Palyga, 1992) and the East Lansing mapping reference population (Levin et al., 1994b), as well as a resource of cosmid clusters (Guillemot et al., 1988). The class II genes are interspersed with the class I genes along a relatively small region of chromosome 16, the B-F/B-L region, represented in a cosmid cluster of 130 kb. There are at least 5 class II [beta] genes in total, two of which are known to be expressed and lie within the B-F/B-L region about 8 kb apart. The intron/exon structure for class II is similar to that in mammals, but with much smaller introns. The total gene sizes are less than 2 kb in comparison to 8-20 kb in man (Guillemot and Auffray, 1989). At present, only one class II [alpha] gene has been found (Kaufman et al., 1993), about 5 cM away from the B-F/BL region. There are four class I genes, two within the B-F/B-L region (Kaufman et al., 1993) which lie in a 20 kb segment of the chromosome. Between these two genes, in the B-F/B-L region, lies the TAP2 (Transporter associated with antigen processing-2) gene (Bumstead et al., 1994b). In the same B-F/B-L region are also located one of many B-G genes and the gene encoding the [beta] subunit of a GTP binding protein (Guillemot and Auffray, 1989). There are many B-G genes, a number of which are transcribed. However, their positions on the chromosome are not as yet defined, although one clone maps to the same linkage position as TAP2 (Mariani et al., In Press). Another gene mapping to this linkage position is the chicken homologue of G9a (BAT8), which is linked to the class III genes in man (Spike and Lamont, 1995). One of the few identified class III genes, C4, also lies within this region (Kaufman pers. com.).
There is a second MHC region, termed the Rfp-Y system (Briles et al., 1993), which was originally unlinked to the B-complex. This region contains two closely-linked class I and two class II genes (Miller et al., 1994a). Recent work (Miller et al., 1996) using trisomy mapping has shown that the Rfp-Y system contains two class I and three class II genes and lies on chromosome 16, in the same cosmid cluster as a class II MHC gene and the rDNA genes (NOR). Between the two class I [alpha] genes lies a lectin, type-C gene, Lec (Bernot et al., 1994).
Thus the MHC region in chickens is significantly smaller than that in mammals, although it appears to contain two regions of class I and class II genes, these are interspersed while the mammalian regions are separate.
The association of the MHC region, or genes associated with it, with disease resistance is shown in a number of diseases some of which are discussed below.
1.4.2.2 Salmonellosis
Salmonellosis is caused by the colonisation of the gut following ingestion of the Salmonella bacterium. This can lead to a systemic infection of the chicken with high morbidity and mortality by species such as S. typhimurium, S. gallinarum and S. pullorum. Alternatively chickens can be infected by S. enteriditis which causes little clinical infection in the birds, but is excreted into the eggs which when eaten uncooked by man can result in infection. At present control is by oral administration of antibiotics. Salmonellosis resistance in mice has been shown to be provided by the presence of a glycine at amino acid position 105 in a variant of the Nramp1 protein, formerly known as Ity (Vidal et al., 1993). The Nramp protein is expressed in macrophages as a putative transmembrane protein. However, although expressed in chickens, Nramp1 does not control resistance to Salmonellosis in the cross studied (Hu et al., 1996). Experiments have shown that chicken lines 6₁ ,WL and N were resistant to all serotypes tested, whereas lines 15I, 7₂ and C were susceptible (reviewed by Bumstead et al. (1991)). Crosses and backcrosses between resistant and susceptible birds showed that resistance is dominant and not linked to the sex or the MHC chromosomes, and is probably controlled by a single gene.
1.4.2.3 Infectious bursal disease virus
IBDV (Infectious bursal disease virus) is a birnavirus which mostly affects chicks 3-6 weeks old, causing lesions within the bursa, and subsequent immunosuppression. This usually leads to secondary infection by opportunistic pathogens, which causes increased suffering for the birds and great expense for the farmers. There are a number of different strains, some of which cause high mortality and are termed very virulent. Bacon (1987) found some resistance associated with the MHC; as haplotypes B² or B²¹ were more resistant than haplotypes B⁵,B¹⁵ or B¹². However, Bumstead et al. (1993) examined 11 inbred and partially inbred lines of chickens and determined by F₁ and F₂ crosses that resistance was partially dominant and did not involve the MHC. Although there is a vaccine against this virus, in 1989 a very virulent strain of the virus reached Britain and, despite vaccination with "hot" strains, has increased mortality by 1.5-2%. In 1994 the National Farmers' Union estimated that IBDV had cost the industry £15 million (Law and Payne, 1996). It would be very useful to identify genes for resistance to this disease as they could be used directly to screen birds for resistance or used to initiate a more informed approach to vaccine development.
1.4.2.4 Marek's disease virus
MDV (Marek's disease virus) is a herpes virus, causing a lymphoproliferative disease in chickens leading to paralysis and tumours. The initial disease phase is acute and occurs about 3 days PI involving virus replication in B lymphocytes, and as a consequence of an immune reaction to antigen particles the T lymphocytes become activated. These T lymphocytes disperse around the body giving a persistent viraemia. Some lymphocytes are transformed and proliferate as lymphomas in visceral organs. After two weeks a second cytolytic infection occurs in the feather follicle epithelia resulting in the shedding of cell-free virus into the environment (Payne, 1996). MDV can be controlled by administration of the Rispens vaccine or the HVT vaccine, which is a live, attenuated herpes-virus of turkeys. However, there are now a number of very virulent strains of MDV which are unaffected by vaccination and are increasingly causing high mortality, up to 80% in comparison to 10% associated with the classical from of MDV (Payne, 1996). In 1994 the National Farmers' Union estimated that MDV had cost the industry £3 million (Law and Payne, 1996), however this is a gross under-estimate.
The genetic differences within the B-complex have been associated with resistance to a number of diseases; the initial association of the MHC was with MDV resistance. The B²¹ haplotype has been shown by a number of experimenters to confer resistance to MDV (Hansen et al., 1967; Longnecker et al., 1976; Briles et al., 1977; Powell, 1984; Hedemand et al., 1993). However, the B¹⁹ and B² haplotypes tend to confer susceptibility. The B²¹ haplotype resistance is not associated with the bursa, but these birds have a higher virus-neutralising-antibody titre than susceptible birds which are immunosuppressed. At 7 days post-infection the virus titre falls in the blood and spleen whilst there is development of humoural antibody and cell-mediated immunity.
Non-MHC genes also have an effect on the bird's resistance to disease. This area has again been most intensively studied for MDV, where two lines of birds, from the Regional Poultry Research Laboratory (RPRL) line 6 and 7, have identical MHC haplotypes of B² , yet show very different resistances to MDV. Line 6 was shown by Stone (1969) to suffer from only 6% mortality in comparison to 100% of line 7 birds with MDV. It has been shown by thymic transplant from line 7 to thymectomised line 6 birds that lymphomas will occur in cells derived from line 7. However, no resistance is shown if the transplant is from line 6 to line 7. Transplants of bursa or spleen show no alteration in resistance, indicating that line 6 resistance is associated with its T cells (Powell et al., 1982). Line 6 birds show a high level of immune response which correlates with viral disappearance not seen in line 7 birds. It has also been shown that macrophages in uninfected line 6 birds have a higher phagocytic index than those of line 7 birds (Powell et al., 1983). The genes which control these differences between lines 6 and 7 are not fully understood, but a number of candidate genes have been suggested. These include Ly-4 (Payne, 1991), Bu-1, and Th-1 (Gilmour et al., 1976), alkaline phosphatase (Yotova et al., 1990), polyesterase (Bachev and Lalev, 1990), thermal regulatory genes (Yotova, 1988) and immune response and competence genes (Okado and Yamamoto, 1987). If the genes controlling resistance can be identified, breeders will be able to select genetically resistant stock and it should be possible to target vaccines or chemical prophylaxes to both the very virulent and normal forms of the virus. At present these candidate genes appear to have some effect on resistance but do not appear to be the major genes involved.
1.4.2.5 Newcastle disease
Newcastle disease, originally imported from the far east, is caused by a myxovirus which can infect chickens as well as turkeys and occasionally man. It has a 2-7 day incubation period culminating in lassitude of the birds, respiratory distress, diarrhoea and drooping wings. Egg production is greatly affected and there is high mortality in young chicks. Oral administration of live vaccine has significantly reduced the incidence of this disease.
1.4.2.6 Infectious bronchitis virus
IBV (Infectious bronchitis virus) is a highly contagious respiratory disease of chickens caused by a coronavirus. The respiratory infection reduces egg production and quality and can lead to secondary infections with other micro-organisms and thus death. This is controlled to some degree by vaccination. Resistance to IBV is also non-MHC derived. However, it does not appear to be controlled by a single gene as the levels of infection and organs affected are considerably different between the inbred lines of birds (Bumstead et al., 1991).
1.4.2.7 Avian leukosis and sarcoma virus
ALSV (Avian leukosis and sarcoma virus) are a group of retroviruses which are characterised by autonomous proliferation of blood cell precursors. They are acquired by infection with exogenous virus or vertical transmission from the hen. The infection leads to erythroblastosis at a few months of age or to lymphoid leukosis in the more mature birds. This can also result in tumours, as in Rous sarcoma virus infection. Resistance to infection induced by the MHC has also been observed for the ALVS group (reviewed by Schierman and Collins (1987)). Rous sarcoma virus-induced tumours have regressed dramatically in birds of haplotype B⁶ or B²;however, haplotypes B⁵ or B¹³ are susceptible. It has been determined that the genes influencing regression lie within the B-F region of the MHC and are also associated with inhibition of the development of metatastic tumours. There is a small amount of resistance to ALV shown in birds with B² haplotype whereas those with B⁵ haplotype have increased erythroblastosis and a higher incidence of haemangiocarcinomas. ALSV resistance is due to a lack of a cell surface receptor for the specific sub-group of virus reviewed by Bumstead et al. (1991). Recently the chicken gene tv-a, which confers susceptibility to infection by ALSV subgroup A, was isolated by Young et al. (1993).
1.4.2.8 Coccidiosis
Coccidiosis is caused by an apicomplexan protozoan, Eimeria, which inserts itself intra-cellularly into the gut wall, where it reproduces. It is spread by ingestion of contaminated faeces, and so is prevalent in litter-kept chickens. Infected chickens show a reduction in growth rate and feeding efficiency, which can lead to increased fatality. There has been some control of infection using attenuated vaccines such as Paracox, a vaccine based on an initial infection with small numbers of parasites which produces a strong protective immune response. MHC haplotype affects the resistance of birds to coccidiosis, where those with B²¹ haplotype are more resistant to Eimeria maxima than those with B¹⁵ haplotype. However this differs for the various species of Eimeria (Bumstead et al., 1991). Resistance to coccidiosis is also determined by non-MHC as well as MHC genes. Although these have not yet been defined, it has been shown that different lines of birds show differing resistance to coccidiosis. However, it again seems that the genes involved act differently for each parasite as, for example, those birds with resistance to E. maxima show susceptibility to E. tenella, (reviewed by Bumstead et al. (1991).
1.4.2.9 Fowl cholera
Fowl cholera is caused by the gram-negative, facultative aerobe Pasturella multocida. This disease has a sudden onset with fever leading to haemorrhagic septicaemia. Oral antibiotics can be administered, but this has little effect on those birds already showing the disease. Lamont et al. (1987) have also shown that MHC haplotype affects the resistance of birds to fowl cholera. Lines of birds homozygous for MHC haplotype and the F₁ and F₂ progeny were tested by intra-nasal administration of P. multocida. Homozygote B¹ haplotype birds demonstrated significantly higher resistance to the bacteria than B¹⁹homozygote birds.
It should be noted that resistance to different diseases associated with the MHC, is not generated by a single haplotype, thus suggesting an important role for non-MHC genes in the determination of resistance (Lamont, 1993).
1.5 Objectives
The genetic map of the chicken is now essentially complete, containing more than 800 loci at an average spacing of around 3 cM (Bumstead and Palyga, 1992; Levin et al., 1994b). A YAC library has now been constructed which contains 16,000 clones of an average insert size of 0.58 Mb giving a coverage equivalent to 10 copies of the genome (Toye et al., In Press). However, although it is now possible in principal to move from mapping information to the equivalent YAC clones, to do this effectively requires a means of efficiently increasing the density of markers in relevant regions of the map. Once this is achieved, it will be possible to isolate specific genes using the framework of the markers. In conjunction with the generation of markers, it is now possible to directly select for candidate genes, by positional cloning, which could code for traits of interest. Positional cloning is now used as the established approach to determine disease genes. It proceeds from an existing disease to identifying the responsible gene's chromosomal locus, then to recovering the gene, and finally to discovering the mutational alteration that accounts for the disease (Collins, 1992). Variants at a locus defined by a candidate gene could help to determine their effects on a trait of interest.
The aim of this project was to develop techniques for generating this high density of marker loci. Possible ways to approach this include adding known genes to the map whose position or possible function can be compared with that of other organisms; the addition of large numbers of markers to the map, by saturation or random mapping leading to enough markers such that there is one every centi-Morgan; and targeting markers to regions of the genome which contain genes of interest, in order to increase the marker density just within that region.
1.6 Approaches
In order to isolate the genes responsible for inherited traits it will be necessary to clone the chromosomal region containing the trait locus. In practice, this means isolating YAC clones containing the marker loci used to map the trait gene. In order to do this, refined mapping of the chicken genome using a large number of markers is needed. Three major areas of investigation have been undertaken:
i. Isolation of cDNAs which might code for candidate traits by ESTs.
ii. Total saturation of the genome with large numbers of anonymous markers by RAPD.
iii. Targeted isolation of markers, which could include candidate genes, in specific regions by RDA.
1.6.1 Expressed Sequence Tags (EST)
ESTs of genes have been used as an effective way of generating further loci on genomic maps which provide useful genetic information rather than just being anonymous markers. ESTs are the sequences generated from cDNA clones and act as type I anchor loci within the linkage map (O'Brien and Graves, 1991). The chance of cloning a candidate gene can to some extent be controlled by the choice of tissue for mRNA extraction which will generate the ESTs. Candidate genes can be isolated as ESTs and homology with known genes in other species identified and the inference of possible association to traits pursued. Comparative mapping of ESTs can determine possible regions in which a gene of interest may lie, as it is already mapped in that area in a different species (Adams et al., 1991). However, beyond using a specific tissue for mRNA extraction, there is no other method for targeting the required candidate gene.
In order to genetically map ESTs in chickens it was necessary to find polymorphisms between the parents of the Compton reference mapping population, this is usually achieved by RFLP analysis. However, this is time consuming and not always productive. Ultimately with the construction of a WG-RH library all chicken ESTs will be mapped, as there is no need for polymorphisms between the two lines. Using ESTs derived either from tissue known to express the gene product of interest, or from comparative mapping synteny, it is unlikely that the candidate gene will be isolated, but this will at least lead to further placement of informative loci on the chicken linkage map.
1.6.2 Random Amplified Polymorphic DNA markers (RAPD)
RAPD analysis should lead to the saturation of the genome with markers to a level of at least one marker every 0.5 cM, without the requirement of previous genetic information and using few expensive oligonucleotides. RAPD markers are usually generated by the amplification of random DNA segments with single short arbitrary primers. An oligo is used as both forward and reverse primer in PCR amplification, at a low annealing temperature (in the region of 36°C) (Williams et al., 1990). This procedure generates large numbers of bands for each individual's genomic DNA (resolved by gel electrophoresis) and may identify multiple polymorphisms between individuals. Polymorphisms detected by RAPD are due to point mutations, which affect the binding of the primer to the DNA and also to insertions and deletions between primer binding sites. Although these markers are anonymous, once a large number have been generated, high density genetic maps can be constructed. This system should rapidly generate large numbers of closely spaced markers. Each primer should be capable of detecting a number of polymorphisms leading to the identification of multiple informative loci for a small amount of initial resource. The benefits of RAPD technology include its simplicity and ease of use in the laboratory. It does not require any previous sequence data, but will give easily interpreted results. It is also less costly in time than other methods, for example RFLP. However, changes in enzyme manufacturer, primer or enzyme concentration, or thermal cycling equipment, can give inconsistent results (MacPherson et al., 1993; Meunier and Grimont, 1993). RAPD markers have been used for gene cloning, medical diagnostics and trait introgression in breeding programs (Williams et al., 1990). Levin et al. (1993) used RAPD markers to generate new markers on the Z chromosome of the chicken in order to identify sex-linked traits. RAPD markers have also been useful in determining phylogenetic relationships between species as demonstrated by Barral et al. (1993) with the Shistosoma genome.
1.6.3 Representational Difference Analysis
Representational difference analysis (RDA) uses subtractive and kinetic enrichments to purify restriction endonuclease fragments present in one population of DNA fragments but not in another (Lisitsyn et al., 1993). It does not rely on knowledge of the sequences involved but can be used to compare complex DNAs elucidating the sequence differences between them. In comparison to subtractive hybridisation, this method is far quicker and more successful, as the complexity of DNA is first reduced by the representation step, and the regions of interest are enriched. A representation of the genome is made by digestion with a specific endonuclease, ligation of enzyme-specific initial adaptors (termed R) and amplification of fragments by PCR. This generates a pool of fragments referred to as amplicons. Two species of DNA are used; DNA containing the target sequence, the "tester" and the non-target-containing DNA, the "driver". The adaptors are subsequently removed from the amplicons, and new adaptors (termed J) ligated only to the tester amplicons which are then mixed, melted and reannealed with excess non-target DNA (driver) amplicons. During treatment with Taq DNA polymerase, only those amplicons with adaptors can be filled in at the 3' ends. When the samples are amplified by PCR, exponential amplification only occurs with the tester homoduplex DNA. Single-stranded DNA is not amplified, and any heteroduplexes are only amplified linearly. This hybridisation and amplification step can be repeated many times alternating between two sets of adaptors (termed N and J) to greatly increase the enrichment of the target DNA. Disadvantages of this system mainly concern the resource of target and non-target DNA, which particularly in some experiments, should come from the same individual, or at least a related individual.
More recently, Lisitsyn et al. (1994) have refined this technique by using either congenic strains or two-generation crosses to generate genetic markers linked to a trait of interest. They used Genetically Directed Representational Difference Analysis (GDRDA) to target three polymorphisms to within less than 1 cM of the mouse nude locus on chromosome 1. Thus, RDA can be used to isolate candidate genes directly or to target specific regions of genomic maps, in order saturate these regions with markers for ease of YAC contig alignment.
RDA has been used to isolate a number of candidate genes from a variety of diseases which may be caused by viruses, including the identification of human herpes virus 6 from patients with multiple sclerosis (Challoner et al., 1995), a flavivirus from patients with hepatitis strain GB (Simons et al., 1995) and a herpes-like virus from AIDS patients with Kaposi's sarcoma (Chang et al., 1994). RDA has also been applied to the isolation of candidate genes for different cancers; Schutte et al. (1995) identified a tumour suppressor gene deleted in pancreatic carcinoma, while Hino et al. (1995) found 4 candidate genes for renal carcinomas in the Eker rat. However when Tsuchiya et al. (1994) tried to use RDA to isolate genes specific to the Yoshida sarcoma (YS), not expressed in the long term survival Yoshida variant, they discovered that the two were not as closely related as had been believed and generated nine genes specific to YS.
RDA has also been adapted to the isolation of candidate differentiation sequences from tissue using cDNA, initially by finding the two genes RAG-1 and RAG-2 which are known to be responsible for activating V(D)J recombination (Hubank and Schatz, 1994). More recently the technique has been applied to the isolation of sequences present on a single chromosome. This has been successful in isolating sequences specific to the bovine-Y chromosome (Wigger et al., 1996), mouse chromosome 10 (Baldocchi et al., 1996) and wheat chromosome 6 (Delaney et al., 1995).
RDA is a powerful tool which can be used to isolate sequences on specific chromosomes, candidate genes and previously undetected viruses. This technique depends on a number of enzymatic steps, which could lead to failure. However, with the decrease in DNA complexity and enrichment of target sequences, RDA could also prove highly successful for chickens as well as mammals and plants.

2.1 Materials
Chemicals, stock solutions, media, oligonucleotides and the thermal profiles used for PCR reactions are detailed in Appendix A.
2.2 Bacterial strains and plasmids
Details of bacteria and plasmids used are shown in Table 2.1.
Table 2.1 Descriptions of bacterial strains and plasmid.

2.2.1 cDNA Library IAHchB1
The cDNA library, IAHchB1 in pCDM8, was used as a source of clones for sequencing and mapping. The library was constructed as detailed in Tregaskes and Young (1997), from RNA extracted from a 17-day embryonic whole bursa and has insert sizes of 0.5 to 3.5 kb. Thus any random clone could contain part or all of any gene expressed in the bursa. This library would therefore include ubiquitously expressed genes, as well as those specific to the tissue. The Bursa of Fabricius is a major organ in the immunological system of the bird mainly concerned with the multiplication and development of B lymphocytes, which are involved in the control of disease through antibody production. Thus expressed genes specific to this tissue could be of great interest as they could code for proteins associated with the immune system or the development of B lymphocytes. Any of these genes could code for proteins involved in host resistance, and so should be fully characterised.
2.2.2 Transformation with IAHchB1 by electroporation of Escherichia coli MC1061/P3
A single colony of Escherichia coli MC1061/P3 was picked and grown up overnight in 5 ml L-broth (160 rpm, 37°C). Two ml of this overnight culture were used to seed 200 ml of L-broth which was then incubated (160 rpm, 37°C) until the OD₆₀₀ measured 0.5. The culture was placed on ice for 30 min, with occasional swirling, and then transferred to a pre-cooled sterile 250 ml centrifuge bottle. The cells were pelleted by centrifugation at 2500 g for 15 min. The supernatant was discarded, and the pellet resuspended in 200 ml pre-cooled (4°C) sterile distilled water. The cells were replaced on ice for 30 min, and then centrifuged as before. The cell pellet was resuspended in 100 ml cold sterile distilled water, and returned to the ice for another 30 min. The cells were centrifuged as before, the supernatant removed and the pellet resuspended in 5 ml of pre-cooled sterile filtered (using a 0.2 µm pore filter) 10% glycerol. This cell suspension was transferred to a 50 ml Falcon tube and left on ice for 30 min. The cells were then centrifuged at 700 g, 15 min. The supernatant was removed and the pellet resuspended in 500 µl of 10% glycerol and stored on ice until used.
Aliquots of 100 µl of competent cells were added to varying amounts (10 ng, 50 ng and 100 ng) of the IAHchB1 library plasmid DNA ( initial concentration 500 µg/ml) in pre-cooled tubes and electroporated, using a Biorad Gene Pulser apparatus with pulse controller at 2.5 kV, 200 [Omega] and 25 mF. One ml of SOC medium was immediately added to the cuvette, and the entire contents were transferred to a 5 ml Sterilin tube containing 1.5 ml SOC. The tubes were incubated (160 rpm, 37°C) for one hour before plating onto pre-warmed L-agar plates, containing 7.5 µg/ml tetracycline and 12.5 µg/ml ampicillin. The plates were incubated overnight at 37°C. Colonies were selected at random, and preparations of plasmids made from these.
2.2.3 Cloning RDA products into pGEM-T vector
All Representational Difference Analysis (RDA) products were cloned into the pGEM-T (Promega) TA cloning vector following the manufacturer's protocol. Approximately 8 ng of un-purified RDA re-amplification product was ligated to 50 ng of pGEM-T at 15°C for 3 hours. Once ligated, 2 µl (11.6 ng) of the product was transformed into 50 µl of E. coli JM109 competent cells (Promega). Initially, the tubes were kept on ice for 20 min before heat-shock treatment at 42°C for 45 sec, and then returned to ice for 2 min. 900 µl of SOC broth was added with subsequent incubation with agitation at 37°C for one hour. The cells were then plated onto L-agar plates containing 100 µg/ml Amp, 40 µg/ml X-GAL and 46 µg/ml IPTG. After incubating the plates overnight at 37°C they were placed at 4°C for one hour to enhance blue/white colony identification.
2.3 Chicken Strains and Crosses
The inbred lines of chickens used for all the experiments are white leghorn chickens kept at Compton. The individual lines described here, lines N, 15I, 6₁ and 7₂ were all imported from the Regional poultry research laboratory (RPRL), East Lansing. The coefficients of inbreeding for all the lines used are detailed in App. A. All the inbred lines produce two generations a year, the female siblings are artificially inseminated by pooled semen samples from the male siblings at two weeks after the start of lay, in order to synchronise the flock production. Line 15I has been closed since its importation in 1962 and had been inbred for 17 generations by 1978 . Lines 7₂ and 6₁ have been closed since their importation in 1972 and had been inbred for 7 generations by 1978 . Line N has been closed since its importation in 1982, it is not known how many generations of inbreeding have occurred.
Chickens from the Compton Mapping Reference Population were derived from a backcross from line N and line 15I birds, as described by Bumstead and Palyga (1992). These lines are inbred and differ in their resistance to salmonellosis, line N being resistant, and line 15I susceptible. One of the F₁ progeny (B989), a line N x line 15I female, was backcrossed to a line 15I male parent (B988). This mating produced 113 backcross progeny, which have been extensively used to construct a genetic map in this population.
A second population of birds, the Marek's Disease (MD) reference population, has been established using line 6₁ and line 7₂ . These two White Leghorn lines show differences in their susceptibility to MD: line 6₁ is resistant to MD, whilst line 7₂ is susceptible. In this population F₁ birds were backcrossed to the line 7₂ parent to give 85 progeny for the generation of a new linkage map and the mapping of resistance. F₁ birds were also mated to give an F₂ population of 42 birds for further analysis of disease resistance. All the birds used were infected with MD virus and the infection characterised in terms of viraemia by quantitative PCR and plaque assays, as well as mortality and tumour development
Quantification of Marek's disease virus by PCR has been developed (Bumstead et.al. In press). The PCR assay used two MDV specific primers to amplify a 279 bp product, which was fluorescently labeled to enable quantifiable detection by ABI Genescan gel electrophoresis. Primers to chicken interferon were used as a control assay for cell number and PCR effectiveness. The PCR assay was calibrated and its linear range determined by comparing the amount of PCR product for a range of dilutions of an MDV-transformed cell line. The relationship of PCR value to the number of viral genomes was also determined by applying the PCR assay to a series of dilutions of a plasmid containing the template sequence. This was carried out in addition to the usual viral plaque assay, with which it correlated well. The PCR assay was then tested on samples of lymphocytes from pure line 6₁ and 7₂ birds as well as F₂ birds of a cross between these two lines. The PCR assay gave repeatable and consistent results which in fact correlated better with the fate of the birds than did the plaque assay results.
There is no data for the phenotypes of the F₁ birds with respect to Marek's disease resistance, as all available birds were used to generated the F₂ population. However, the phenotype data for the pure lines and the backcross birds is shown in App. A. Approximately half the backcross birds show some resistance to Marek's disease according to the quantitative PCR values.
2.4 DNA Preparation
2.4.1 Genomic DNA Extraction
DNA was extracted from 0.5 ml of chicken blood collected from the wing vein using 3% sodium citrate as an anti-coagulant. The blood was treated with 1.5 ml of 1% saponin in PBS, mixed and left at room temperature (RT) for 10 min before washing with PBS. The nuclei were then pelleted by centrifugation at 1000 g for 5 min. This was repeated twice to ensure that the nuclei were free of cytoplasmic contamination. After centrifugation, the pellet was resuspended in 2 ml TE with 10% SDS. Protein and RNA were removed from the pellets by proteinase K digestion (20 mg/ml) for two hours at RT, followed by addition of RNase A (100 mg/ml) and mixed by rotation at RT overnight. The samples were phenol-extracted twice and chloroform:iso-amyl alcohol extracted once, before precipitation with 0.5 ml 2 M sodium acetate and 8.75 ml ethanol at -20°C overnight. The DNA was pelleted by centrifugation at 7800 g for 15 min, washed with 2 ml 70 % ethanol, air-dried and resuspended in 2 ml TE.
2.4.2 Plasmid DNA Preparation-ABI Protocol
Plasmid DNA was prepared using the following modified mini alkaline-lysis/PEG precipitation procedure. Overnight cultures of 5 ml of each clone were grown up in Terrific broth containing 7.5 µg/ml tetracycline and 12.5 µg/ml ampicillin. After removal of 850 µl for glycerol stock preparation, the remaining culture was pelleted by centrifugation at 1000 g for 3 min. Glycerol stocks were made by the addition of 150 µl glycerol to the removed cultures, which were then vortexed briefly and snap frozen in a dry-ice and ethanol mixture before storage at -70°C. The cell pellet was resuspended in 200 µl of GTE buffer, and lysed by addition of 300 µl of 0.2 N NaOH and 1% SDS. The tube was inverted and incubated on ice for 5 min. The solution was then neutralised with 300 µl of 3 M potassium acetate (pH 4.8) and inverted, before being placed on ice for 5 min. Cellular debris was pelleted by centrifugation for 10 min at 7000 g. The supernatant was removed and digested with 20 µg/ml RNase A at 37°C for 20 min and then extracted twice with 400 µl chloroform. The extracted DNA was precipitated with an equal volume of isopropanol at RT and immediately pelleted by centrifugation for 10 min at 7000g. After washing with 70% ethanol, the pellet was dissolved in 32 µl Milli-Q water and further precipitated by the addition of 8 µl 4 M NaCl and 40 µl 13% PEG₈₀₀₀, with incubation on ice for 20 min. The DNA was pelleted by centrifugation at 4°C for 15 min. After washing in 70% ethanol the pellet was dried and resuspended in deionised water at 1 µg/µl. This preparation was stored at -20°C.
2.4.3 Plasmid DNA Preparation-Hybaid maxiprep Protocol
Individual bacterial colonies were picked into two 10 ml of cultures of L-Broth, containing 100 µg/ml ampicillin and incubated with agitation overnight at 37°C. The following day 850 µl of each was removed and a glycerol stock prepared as previously described. The rest of the culture was centrifuged for 5 min at 1000 g to pellet the cells before resuspension in 250 µl Milli-Q water per tube. The combined 500 µl was then placed in a 1.5 ml microfuge tube and spun at 7000 g for 1 min, the supernatant removed and the tube re-spun to remove any remaining liquid. The pellet was then resuspended in 200 µl of pre-lysis solution and vortexed before addition of 400 µl of alkaline lysis solution. The tubes were inverted 15 times and left to stand at RT for 5 min before addition of 300 µl of ice-cold neutralising solution and re-vortexing for 5 sec. The tubes were left on ice for 5 min and then centrifuged for 5 min at 7000 g. The supernatant was transferred to a clean 2 ml tube and 900 µl of Hybaid binding buffer (containing guanidine thiocyanate) added. The tubes were again inverted 15 times and left to stand for 3 min at RT. The upper aqueous layer was discarded and the binding beads transferred to the spin filter and centrifuged for 3 min at 7000 g. The eluent was removed and 500 µl wash solution added to the spin filter before centrifugation as before. The eluent was again removed and the filter centrifuged for a final 5 min to ensure all liquid was removed. The filter was then transferred to a new tube and 100 µl TE was carefully stirred into the beads to elute the DNA followed by centrifugation for 5 min at 7000 g. This gave an approximate DNA concentration of 0.4 µg/µl.
2.5 Restriction Endonuclease Digestion
DNA samples were digested with various restriction endonucleases according to the manufacturer's instructions (of 10 U of enzyme/µg of DNA). See App. A for details of enzymes used.
2.6 Electrophoresis
2.6.1 Agarose Gel Electrophoresis
Plasmid DNA and PCR products were resolved by electrophoresis at 4 V/cm for one hour in 1 x TBE, 1% agarose gels with 10 ng/ml of ethidium bromide. Genomic DNA digests were resolved by electrophoresis at 1 V/cm overnight in 1 x TAE, 0.6% agarose gels with 30 ng/ml of ethidium bromide. Bands were visualised using a UV transilluminator, and the resulting pattern recorded using a GDS5000 digital camera system (Mitsubishi).
2.6.2 ABI Genescan electrophoresis
2.6.2.1 Creation of a Matrix File
The ABI Genescan system enables the electrophoresis and detection of fluorescent products. A matrix file was made so that background fluorescence could be reduced in the Genescan gel to enable good resolution of the fluorescent products. Prior to using the Genescan for analysis, the matrix file was created to ensure optimum fluorescent detection. An identical gel to those used for Genescan analysis, a 12 cm, 5.5 % "Long ranger", 1 x TBE gel, was set up and run for 5 ¼ hours at 800 V, on filter set B ([lambda]= 531, 545, 560, 580 nm) in the ABI electrophoresis apparatus. The dye primer matrix standard kit P/N 401114 was used, loading two samples for each dye colour. The samples were loaded, 5 µl per lane, alongside samples of 0.5 µl of Genescan ROX 2500 standard which were either native, or denatured with the addition of an equal amount of deionised formamide and heated before loading.
Raw data was collected using the Genescan (ABI) collection and analysis programs. The matrix file was constructed according to ABI instructions. Once the gel had run, the tracking was checked and a new matrix created for all four dye colours. This matrix was entitled "5.5% LR 1 x TBE Matrix" and was used in all subsequent Genescan reactions.
2.6.2.2 RAPD-PCR Electrophoresis
All the RAPD-PCR samples were denatured with the addition of an equal volume of deionised formamide, mixed with 0.5 µl of Genescan ROX 2500 standard, heated to 94°C for 2 min and kept on ice until loading. The samples were loaded on a 12 cm, 5.5 % Long ranger gel which was run in the ABI electrophoresis apparatus for 6-8 hours at 800 V on filter set B. The Genescan collection software was set up according to ABI instructions.
2.6.3 Sequence Gel Electrophoresis
Samples were loaded onto a 6% polyacrylamide sequencing gel (ABI 373A Sequencer), and electrophoresed in 1 x TBE for 14 hours at 2500 V. The apparatus was set up according to ABI protocols using Filter set A ([lambda]= 531, 560, 580, 610 nm) and the data collected using the 373A collection package.
2.7 PCR Amplification
2.7.1 Plasmid insert PCR Amplification
PCR amplification was carried out using diluted plasmid DNA (1:10) approximately 100 ng, in a 50 µl PCR reaction mix of 10 x buffer, 100 µM dNTPs, 2 mM MgCl₂, 25 pmol of each primer (oligonucleotides 12 and 16 for pCDM8 and oligonucleotides 147 and 151 for pGEM-T, see App.A). The reaction mix was overlaid with 50 µl of mineral oil before the addition of 2.5 U Taq DNA polymerase after the tubes had reached the denaturation temperature. Amplification was for 25 cycles with an annealing temperature of 55°C. The thermal profile in program 1 was followed (App. A).
2.7.2 Colony PCR of RDA products
Colonies were picked and screened for inserts by colony PCR. Each colony was picked directly into a 50 µl PCR reaction mix of 10 x buffer, 100 µM dNTPs, 2 mM MgCl₂, 1U Taq DNA polymerase, 25 pmol of each primer (147 and 151), and amplified for 25 cycles with an annealing temperature of 55°C. PCR products were analysed by agarose gel electrophoresis. Each picked colony was also streaked onto L-agar plates, containing 100 µg/ml ampicillin, in an ordered array and also inoculated into 1 ml of L-Broth, containing 100 µg/ml ampicillin, for incubation with agitation at 37°C overnight. The overnight cultures were spun down at 7000 g for 1 min and resuspended in L-Broth with 20% glycerol, followed by snap-freezing in ethanol and dry-ice, and storage at -70°C.
2.8 Gel purification of PCR products
Clones containing two plasmids were initially PCR amplified by colony PCR described above using primers 147 and 151. A 40 µl sample of each product was separated by electrophoresis on a 2% low melting point agarose gel. The insert DNA was excised as a band from the gel and placed into a 0.5 ml tube. Once the weight was obtained, 3 M NaCl was added to a final concentration of 30 mM and the agarose melted at 68°C for 10 min. The temperature was then reduced to 39°C and agarase (Sigma) added to a concentration of 50 U/ml. After incubation for one hour 60 µl of water was added along with 100 µl 4 M ammonium acetate and 200 µl propan-2-ol. The tubes were centrifuged for 30 min at 7000 g, the supernatant discarded and the pellet washed with 70% ethanol. The DNA pellet was resuspended in 40 µl of water and reamplified as detailed in the pGEM-T sequencing protocol.
2.9 Automated Sequencing
Clones were partially sequenced, from each end of the insert, using the Applied Biosystems 373A DNA automated sequencing system. Dye-deoxy terminator chemistry was used to label the products of Taq DNA polymerase cycle sequencing. Two different ABI kits were used as the technology was developed from the PRISM^TM ready reaction dyedeoxy^TM terminator cycle sequencing kit to the PRISM^TM ready reaction dyedeoxy^TM terminator cycle sequencing FS kit.
The PRISM^TM ready reaction dyedeoxy^TM terminator cycle sequencing kit was used following the ABI protocol to sequence clones in pCDM8. One microgram of double-stranded DNA template was used in each 20 µl cycle sequencing reaction. Oligonucleotides 12 and 16 (see App. A) were used as primers for the reaction. In each sequencing reaction 1 µl of cDNA was mixed with 9.5 µl of terminator premix, 3 µl of one primer and Milli-Q water up to a total volume of 20 µl. This was then overlaid with 50 µl mineral oil and placed into a Hybaid Omnigene PCR machine. The thermal cycling reaction was as described in App. A for ABI Seq. The samples were then extracted with a phenol: H₂O:chloroform mix (68:18:14) at room temperature, and precipitated with 2M sodium acetate and ethanol, to remove excess dye.
Clones in pGEM-T were sequenced using ABI PRISM^TM dyedeoxy^TM terminator cycle sequencing ready reaction FS kit (Perkin Elmer). The protocol was similar to that used previously except that only 8 µl of reaction mix was added. To obtain 1 µg of double stranded DNA each clone was initially PCR amplified in a standard reaction, using primers 147 and 151 before precipitation of the product with an equal volume of 4 M ammonium acetate and propan-2-ol at RT. This precipitated product was centrifuged for 30 min at 7000 g, resuspended in 5 µl Milli-Q water and used directly for sequencing. The PCR product was sequenced using 7 pmol of one of two nested primers: oligonucleotides 156 and 155. Each 20 µl reaction was precipitated on ice for 10 min with the addition of 2 µl of 3 M sodium acetate (pH 4.6) and 50 µl 95% ethanol, centrifuged at 7000 g, and air-dried. The pellet was resuspended in 4 µl of a 5:1 deionised formamide: 50 mM EDTA mixture and denatured at 90°C for 2 min. These samples were loaded and analysed as described earlier.
2.9.1 Sequence Analysis
The 373A software enables direct determination of the sequence. Sequences were edited using SeqEd (ABI) and converted into GCG format. Previously sequenced genes with homology to the clones screened in this project were identified by GCG (Program manual for the Wisconsin package, version 8, September 1994, Genetics Computer Group, 575 Science Drive, Madison, Wisconsin, USA 53711) FASTA analysis, in the EMBL database.
2.10 Southern Blotting
Agarose gels were blotted using a method based on that of Southern (1975). Gels were washed in 500 ml blotting solution A (1.5 M NaCl, 0.5 M NaOH) twice for 30 min and then 500 ml of blotting solution B (0.02 M NaOH, 1 M CH₃COONH₄) twice for 30 min. The gel was placed on a 3 mm filter paper wick on a platform over a reservoir of 1 l of blotting solution B. Hybond-N (Amersham) membrane was carefully laid over the gel, followed by three layers of filter paper. A stack of paper towels were placed over this, and held in position with a weight. Blotting took place overnight. The filter paper and towels were then removed, the nylon membrane marked and air-dried before baking under vacuum for two hours at 80°C.
2.10.1 Probe preparation and hybridisation
Southern blots were soaked overnight in Milli-Q water, and placed into Hybaid tubes. Blots of less than 30 cm² were probed in 20 ml plastic universals. Prehybridisation buffer (50% formamide) was added at 42°C, 0.02 ml/cm² for blots in Hybaid tubes and 0.25 ml/cm² for blots in 20 ml plastic universals. Three universals were placed into each Hybaid tube during hybridisation. The tubes were placed in a preheated (42°C) Hybaid oven and rotated overnight. Two methods of probe preparation were used; Nick translation (Gibco BRL) and Prime-It RmT random primer labelling kit (Stratagene). Probes were labelled with [[alpha]-³²P] dCTP (3000 Ci/mmol) (Dupont NEN), 10 µCi/µl (0.37 MBq/µl).
2.10.2 Nick translation system (Gibco BRL)
Blots were probed either with labelled PCR products (4 µl) or labelled whole plasmids (1 µl) containing about 1 µg DNA. The following basic protocol was used; a mixture of 2.5 µl solution A, 19.5 µl water, 2 µl of [[alpha]-³²P] dCTP and 2.5 µl of solution C were added to the template DNA (see App. A for details of solutions). The mixtures were incubated for one hour at 15°C. The reaction was terminated with 2.5 µl of solution D and 300 µl of salmon sperm DNA (500 µg/ml). The probes were boiled at 100 °C for 7 min and kept on ice for at least 10 min before adding to the blots in the Hybaid tubes.
2.10.3 Prime-It RmT random primer labelling kit (Stratagene).
Blots were probed either with labelled PCR products (2 µl) or labelled whole plasmids (1 µl of 1:40 dilution) containing about 25 ng DNA. The Prime It reaction mix was resuspended in Milli-Q water to a final volume of 46 µl. Template DNA from the pCDM8 clones was added to the reaction mix and boiled for 5 min before addition of 0.37 MBq of [[alpha]-³²P] dCTP and 12 U of Magenta DNA polymerase. The probes were incubated for 10 min at 37°C before 2 µl of stop mix and 100 µl salmon sperm DNA (10 mg/ml) were added. The probes were then boiled at 100°C for 5 min and cooled on ice for 10 min before adding to the Hybaid tubes.
RDA colony PCR products of pGEM-T clones were labelled in a quarter of the Prime-It reaction mix. Seven ng of colony PCR product was added to 10.5 µl of resuspended reaction mix before boiling. The subsequent additions of Magenta DNA polymerase, stop solution and salmon sperm DNA were also a quarter of those detailed previously. However the usual amount of 0.37 MBq of [[alpha]-³²P] dCTP was added to label each product.
Clones were also used to screen chicken YAC library colony blots prepared by Ayo Toye (Toye et al., In Press). Probe labelling and hybridisation was carried out in the manner previously described. One reaction mix tube was used with 1.85 MBq [[alpha]-³²P] dCTP to probe two blots.
After addition of the probes all tubes were returned to the Hybaid oven and left to rotate at 42°C overnight. The blots were then washed with 0.1 x SSC, 0.1% SDS at 55°C twice for 20 min and 0.1 x SSC, 0.1% SDS at 65°C twice for 25 min. Blots in the Hybaid tubes were washed in 0.1 ml/cm² , while blots in universals were washed in 1 ml/cm² of wash buffer. The blots were removed and air-dried before autoradiography at -70°C for at least four days.
2.11 Mapping Techniques
2.11.1 Restriction Fragment Length Polymorphism Analysis
Polymorphisms were detected in the parents of the Compton Mapping Reference Population by probing Southern blots, termed C blots, with the clone of interest. C blots represent the "alternative" parental genomic DNA (bird C50, a line 15I female, and bird B189, a line 15I x line N female) following digestion with seven enzymes: BamHI, EcoRI, XbaI, RsaI, TaqI, MspI and SacII. These were chickens of the same lines as the original parents, which were used in order to conserve the original parental DNA (bird B988, a line 15I male and bird B989, a line N x line 15I female). If a polymorphism was detected after autoradiography, the clone of interest is used to probe a second Southern blot, termed A, B or D blot. The B blot represents the parental and the first 56 backcross progeny genomic DNAs, the A blot the next 27 progeny and the D blot the last 30 progeny, as described by Bumstead and Palyga (1992), digested with the appropriate enzyme (i.e. that which showed a polymorphism in the F₁ parent).
Occasionally, no polymorphisms were observed with the seven enzymes used in the C blots. In these cases a number of other restriction enzymes were used to digest the alternative parental genomic DNA; these included PstI, EcoRV, HaeIII, HindIII and NcoI. Observed polymorphisms determined which type of restriction blot to probe with the clone of interest.
If no polymorphisms were observed in this reference family, the clones were used to probe Southern blots of the two lines used for the MDV reference population, termed M blots. M blots represent the genomic DNA from line 6₁ and line 7₂ digested with seven enzymes: BamHI, EcoRI, XbaI, RsaI, TaqI MspI and SacII. If a polymorphism was detected for a specific enzyme, the clone was used to probe a Southern blot of genomic DNA from the 47 backcross progeny digested with that enzyme, termed MV blot. If further segregation data were required a P blot of the last 37 backcross progeny or a N blot of the 53 F₂ progeny could be probed.
2.11.1.1 Mapping Segregation Analyses
Data from Restriction fragment length polymorphisms (RFLP) and other analyses which segregated alleles in the Compton mapping reference mapping population, or in the MDV reference population were analysed using MapManager software (Manly and Cudmore, 1988). Loci were determined to be linked when the Logarithm of odds ratio (LOD) score was greater than 3.0.
2.11.2 RAPD Analysis
2.11.2.1 RAPD PCR Reaction
Random amplified polymorphic DNA (RAPD) polymorphisms were screened using fluorescently-labelled (FAM and HEX) short primers. Initially the two labelled oligonucleotides (78 and 85) were used to establish a working protocol for RAPD analysis using a line 15I bird (B981) and a line N bird (B984). Reaction conditions and annealing temperatures were optimised. The genomic DNA used in these reactions was prepared as described by Bumstead and Palyga (1992) and diluted to a concentration of 50 ng/µl. Each RAPD reaction mix contained 2 mM MgCl₂ , 10 x PCR buffer, 100 µM dNTPs, 1.25 U Taq DNA Polymerase, genomic DNA, a primer and Milli-Q water in a total volume of 25 µl. Oligonucleotide 85 was used at a concentration of 119 pmol/25 µl with 50 ng of genomic DNA, whilst oligonucleotide 78 was used at 108 pmol/25µl with 63 ng of genomic DNA. All reactions were set up on ice and overlaid with 30 µl mineral oil (Sigma), before denaturation at 95°C for 3 min followed by addition of Taq DNA polymerase. The thermal profile RAPD37 was used, (see App. A). After PCR, all reactions were stored at 4°C, until the RAPD profile was analysed by Genescan.
Once reproducible RAPD products were obtained, the oligonucleotides were used to generate RAPD with the Compton Mapping Reference Population. A group of the first 20 progeny were used for the analysis at a dilution of each DNA of 1:10 which gave a concentration of 50-60 ng/µl DNA.
In order to utilise the fluorescent primers to maximum effect, the template DNA was pre-digested prior to amplification (Levin et al., 1993) using a variety of enzymes: EcoRI, XbaI, RsaI, TaqI, HaeIII and MspI. As a method for generating polymorphisms this was initially tested on DNA from the two lines of birds to determine whether any polymorphisms were generated. Another method of generating further polymorphisms using the fluorescent primers is to combine RAPD primers in the reaction (Levin et al., 1994a). This was attempted using four non-fluorescent primers from Kit 1 Advanced Biotechnologies Ltd. in different combinations. Initial products were separated by agarose gel electrophoresis, to determine viability before analysis by Genescan.
2.11.2.2 Genescan Analysis
After electrophoresis, the gel data was automatically analysed by the Genescan software (ABI). This gave electropherograms for each lane and dye colour. These were compared, polymorphisms identified and the marker segregation scored in the progeny. The RAPD polymorphism was mapped as previously described in the methods and a chromosomal position assigned by linkage.
2.12 Representational Difference Analysis
2.12.1 RDA comparison of line N and line 15I
RDA was used to determine differences between two inbred lines of chickens which differed in their susceptibility to salmonellosis. The original protocol of Lisitsyn et al. (1993) was adapted. Fig 2.1 shows a diagram of RDA, the changes in oligonucleotides used throughout the experiments should be carefully noted. Genomic DNA from a line N (B984 the tester) and a line 15I (C60 the driver) were digested separately with BamHI. Two µg of each were then ligated to 0.5 nmol of each of the RBAM oligonucleotides 70 and 68 overnight at 15°C in a total volume of 30 µl. After dilution to 1 ml with TE (pH 8.0), 40 µl aliquots were removed to tubes containing 400 µl of 10 x SB buffer, Milli-Q water and 120 nmol dNTPs overlaid with 110 µl mineral oil. When the tubes reached a temperature of 72°C, 15 U of Taq DNA polymerase were added and after extension for 5 min at 72°C, 0.5 nmol of RBAM oligonucleotide 68 was added. PCR was carried out for twenty cycles with a combined annealing and extension step at 72°C for 3 min (program RDA1-see App. A). The products of two identical tubes were combined and phenol:chloroform:isoamyl-alcohol (25:24:1) extracted before propan-2-ol precipitation and resuspension in 100 µl TE (pH 8.0). The products, termed amplicons, were verified by agarose gel electrophoresis. The tester and driver amplicons were treated separately.
All amplicons were digested with BamHI and the adaptors removed using QiaQuick PCR purification columns (Qiagen). 20 µg of amplicon was made
fig1
up to 100 µl with Milli-Q water and mixed with 5 volumes of PB buffer. This was placed on the column and centrifuged at 7000 g for 45 seconds. The eluent was removed and the column washed with 7 M NaClO₄. This was removed by centrifugation at 7000 g for 45 sec and the column washed with 0.75 ml PE buffer, again removed by centrifugation, and spin-dried before placing into a fresh collection tube. Addition of 50 µl of 10 mM Tris-HCl (pH 8.5) to the column eluted the amplicon after centrifugation at 7000 g for 45 sec. Approximately 10 µg of tester and 150 µg of driver amplicons were treated in this way. At this stage the DNA concentration of the amplicons was determined by measuring the OD₂₆₀ of a 1:100 dilution and calculating the total amount of DNA.
After removal of the original adaptors, tester amplicons were religated to new adaptors (JBAM oligonucleotides 66 and 69, see App. A), 0.5 nmol of each, overnight at 15°C in a total volume of 30µl. After dilution to 100 µl with TE (pH 8.0), three different ratios of tester and driver amplicons were hybridised:
All the samples were phenol-chloroform extracted as before and precipitated with 1 M ammonium acetate in propan-2-ol. The pellets were resuspended in 4 µl of 3 x EE buffer (Straus and Ausubel, 1990) (pH 5.5), vortexed and overlaid with 35 µl of mineral oil. After denaturation for 4 min at 95°C, 1 µl of pre-heated 5M NaCl was added and the temperature lowered to 67°C for 20 hours of hybridisation. As a positive control, 4 µg of tester amplicon was reannealed to itself, and as a negative control, 40 µg of adaptor-free driver amplicon was reannealed to itself, as described for the tester:driver hybridisation.
After hybridisation, the oil was removed and the products diluted in 390 µl TE (pH 8.0) with 4 µl tRNA. Selective amplification of two 40 µl aliquots from each hybridisation was performed in tubes containing 10 x SB buffer, Milli-Q water and dNTPs overlaid with 110 µl mineral oil as before. When the tubes reached a temperature of 72°C, 15 U of Taq DNA polymerase were added and after extension for 5 min at 72°C, 0.5 nmol of primer JBAM 69 was added. The PCR was carried out using PCR program RDA1 (see App. A). Product was taken and cloned into the pGEM-T vector, as described earlier. The remaining 200 µl of PCR products were phenol-chloroform extracted before propan-2-ol precipitation and resuspension in 10 µl TE (pH 8.0). Single-stranded products were removed at 30°C for 30 min from a 20 µl sample by exonuclease digestion with 20U Mung bean nuclease (NEB), which was subsequently inactivated at 95°C for 5 min after dilution to 80 µl with 50 mM Tris-HCl (pH 8.5). A subsequent round of re-amplification was then performed on the two 40 µl aliquots as previously described for selective amplification using primer JBAM 69 (program RDA1 App.A).
The contents of duplicate tubes were combined and co-precipitated with 10 µg tRNA in propan-2-ol after phenol-chloroform extraction. The resuspended products were digested and the adaptors removed as previously described for the original amplicons. The products were then ligated to new adaptors (NBAM oligonucleotides 67 and 65, see App. A) as described previously. 200 ng of the RDA product (determined by spectrophotometry at OD₂₆₀) was ligated with 0.5 nmol of each oligonucleotide and the ligated products subsequently diluted to 80 µl with TE (pH 8.0) and 0.1 µg/µl tRNA. However only 100 ng of RDA product was reannealed to 40 µg of the driver. After hybridisation, the products were diluted with 390 µl TE (pH 8.0) and the content of two duplicate tubes were selectively amplified using 0.5 nmol of NBAM 65 as the primer (PCR program RDA2, App.A). After this amplification the contents of duplicate tubes were precipitated with tRNA in propan-2-ol after phenol-chloroform extraction. The products were resuspended in 20 µl TE (pH 8.0), single stranded DNA was removed as before and the samples were diluted to 80 µl with 50 mM Tris (pH 8.5). Re-amplification of 40 µl of each product was performed as before using 0.5 nmol of primer NBAM 65 (PCR program RDA2, App.A). Prior to cloning the amplified products, single stranded DNA was removed and the products reamplified using primer NBAM 65. These products were then also cloned into the pGEM-T vector as described in the methods.
2.12.1.1 Screening and characterisation of RDA clones
Genomic DNA from a line N and that from a line 15I bird were separately digested with BamHI separated by electrophoresis and Southern blotted onto Hybond-N membrane as described earlier. Each 20 cm² blot was placed into a 20 ml plastic universal for hybridisation. The blots were probed with colony PCR products as described previously. Those clones detecting variants between the two lines were subsequently used to probe Southern blots of BamHI-digested DNA from the Compton mapping reference population (Bumstead and Palyga, 1992). Any segregation patterns were mapped as previously described in the methods.
DNA was prepared from the cloned RDA products detecting multiple fragments or polymorphic fragments using either the ABI miniprep method or the Hybaid maxiprep kit, as described previously. All clones were digested with NcoI and NotI and electrophoresed though agarose gels, to verify the insert sizes.
2.12.2 RDA targeted to chromosome 16
RDA was carried out as previously described with the following differences. To generate markers specific for chromosome 16, a single bird (a tester) was chosen which had inherited line N alleles for all pre-existing markers on this chromosome. For comparison, the driver sample incorporated pooled amplicons from sixteen siblings at equi-molar concentrations, all of which had inherited line 15I alleles for all pre-existing markers on this chromosome.
2.12.2.1 BamHI Representation
Driver and tester BamHI amplicons were generated and taken through one round of hybridisation whilst testing different conditions.
After cloning the products were screened as before by probing Southern blots of BamHI digests of line N and line 15I birds. Clones detecting polymorphisms were again used to determine the segregation of alleles within the Compton mapping reference population.
2.12.2.2 TaqI Representation
TaqI amplicons were also produced using TaqI-digested genomic DNA from the single tester and 16 driver birds. Products were cloned from all three rounds of hybridisation into pGEM-T. Colonies were picked from all the transformations into microtitre plates, as well as being picked into a PCR reaction mix for colony PCR. The microtitre wells contained 150 µl of TB containing 100 µg/ml ampicillin. 92 colonies were picked leaving four wells blank as controls and the plate was incubated with agitation at 37°C overnight. The plates were scored the next day and 40 µl of 80% glycerol was added to each well and mixed. The plates were individually snap-frozen on dry-ice blocks, before storage at -70°C.
After cloning, the products were screened as before by probing Southern blots of TaqI digests of line N and line 15I birds with the colony PCR products. Clones detecting polymorphisms were again used to determine the segregation of alleles within the Compton mapping reference population.
2.12.2.3 NheI Representation
NheI was also used to digest genomic DNA from the single tester and 16 driver birds to produce larger fragments for the generation of amplicons.
2.12.3 RDA targeted to Marek's Disease Resistance Genes
RDA was carried out as described previously, except that the birds were chosen from the F₂ MDV reference population of 42 birds. Again, a single bird was chosen as the tester, a pure line 6₁, which had been tested and was resistant to MDV. The maximum number of birds possible was chosen to form the pool of driver amplicons, based on their susceptibility to Marek's Disease Virus.
Fourteen F₂ birds (shown in App. E) fell within the criteria for susceptibility. Genomic DNA from these birds was digested with BamHI, ligated to RBAM oligonucleotides 68 and 70 and amplified as previously detailed with RBAM oligonucleotide 68 to give amplicons. Driver amplicons were mixed in an equi-molar ratio as before. The RBAM oligonucleotide adaptors were removed from all the amplicons and JBAM oligonucleotides 66 and 69 ligated to the tester amplicon as before. All hybridisations were carried out at 67°C for 48 hours.
Products were cloned from each round of hybridisation and screened as before by probing Southern blots of BamHI digests of line 6₁ and line 7₂ birds. Clones detecting polymorphisms were used to determine the segregation of alleles within the backcross, F₁ with line 7_2, population.

3.1 Introduction
One method of isolating the gene or genes responsible for inherited traits is to map expressed genes. These are isolated as clones from cDNA libraries and mapped as Expressed sequence tags (ESTs). That is the clone's sequence or part of it is determined and identified as novel in the chicken by comparison to databases. Once mapped, ESTs provide information such that either the candidate gene can be isolated directly or the relative position of the EST to the candidate gene can be inferred. ESTs are highly informative as markers because they show the positions of expressed genes. Thus they can be used for comparative mapping between species to identify regions of conservation of synteny between the chromosomes, and so sparse areas in one species can be investigated for the genes contained in that area in another species. ESTs can also be investigated as candidate genes in a region which is linked to a disease resistance gene. ESTs act as type I anchor loci (O'Brien and Graves, 1991), and so their positions can be usefully compared across strains or between species.
The choice of tissue from which the library is generated can be crucial to the relevance of the ESTs to the candidate gene of interest. For example ESTs derived from a bursal cDNA library could code for proteins associated with the immune system or the development of B lymphocytes and thus be implicated in host resistance to disease. These ESTs could be useful as candidate genes as well as placing more marker loci on the genetic map.
3.2 Results
3.2.1 Mapping ten novel chicken genes
A total of 239 colonies were picked from a plated out cDNA library and 86 clones with single inserts were selected from digested plasmid DNAs for further investigation. The inserts ranged in size from 0.5 to 3.5 kb and all 86 clones were sequenced and the data obtained using the ABI Seq Ed program was exported into Wisconsin GCG format. Sequence comparison to the EMBL database by FASTA analysis was carried out on the VAX computer. Homology to known genes was taken to be significant if identity was above 60% for a region of at least 75 bp of the sequence. Table 3.1 summarises the clones sequenced, and their homologies. A more detailed table is shown in App. B.
Table 3.1 Summary of cDNA clones and FASTA homologies.

Homologies were identified to eight previously cloned chicken genes, and ten novel chicken genes (see Table 3.2). Homology of these clones showed on average around 70% identity with the EMBL sequence. Variations of 67% identity over 533 bp (clone h86) and 83% identity over 105 bp (clone h60) were accepted as showing homology.
A number of the clones only identified sequence homologies for one of the sequenced ends. The ends of the clones were identified according to the primers used (PC or T7), as there was no information as to which was the 3'-end. Clone h24 showed 82% identity for 323 bp at the PC end to bovine GDID4, while clone h42 showed 75% identity over 208 bp at the T7 end to CNBP. Clone h54 showed 83% identity over 150 bp to a mouse EST at the PC end, while clone h60 showed 83% identity over 105 bp to a human clone also only at the PC end. Clone h71 showed 76% identity over 157 bp at the T7 end to HMG1, while clone h78 also showed identity at the T7 end of 81% over 106 bp to human 60S. However, some clones showed sequence identity at both sequenced ends, clone h29 identified homology to ITPR2 for both ends of 80% over a total of 696 bp. Clone h56 showed identity of 76% to an anonymous human clone, at both ends over 702 bp. Clone h58 showed identity across 537 bp of 78% to PSF for both sequenced ends and clone h86 showed identity to COL4a1 of 69% over 1129 bp sequenced from both ends.
Table 3.2 Homology to chicken genes

Once homology had been established for each of the novel chicken genes the sequenced ends were submitted to the EMBL sequence database. Table 3.3 shows the accession and plasmid numbers of all ten genes.
Table 3.3 Accession numbers and sequence data of ten Novel chicken genes.

3.2.2 RFLP Mapping
Eight of the ten novel chicken genes identified RFLP for one or more of the seven restriction endonucleases used, between the alternative parents of the mapping population. However no polymorphisms were detected for the plasmids 1854 or 1846 (the COL4A1 and 60S genes respectively). These were tested on the MDV reference population, and a polymorphism was found for plasmid clone 1846 (see Table 3.4).
Table 3.4 RFLPs detected for the ten novel chicken genes.

A typical result is shown in Fig. 3.1, which shows the polymorphism detected by clone h58. DNA from the segregating progeny of the Compton mapping reference population were digested with the enzyme yielding the clearest polymorphism and subsequently probed with the labeled clone.
Clone h24 showed polymorphisms only with the TaqI enzyme and was subsequently used to probe a B blot of the first 56 birds of the Compton mapping reference population (Fig. 3.2). This showed the segregation of a 2 kb band present in the line N x 15I parent DNA which was absent in the line 15I parent. Fig. 3.3 shows the segregation of a 9 kb band present in the line N x 15I parent DNA digested with XbaI, in a B blot, probed with clone h29. Fig. 3.4 shows the segregation of MspI alleles for clone h42, the 6 kb band present only in the line N x 15I parent DNA. Clone h54 hybridises to a 3 kb band only in the line N x 15I parent DNA when digested with MspI and Fig. 3.5 shows its segregation in the B blot.
Clone h56 hybridised to a 2 kb band present in the line N x 15I parent DNA, absent from the line 15I parent DNA, digested with RsaI. However, when hybridised to D and A blots of 57 birds of the Compton mapping reference population this band was not polymorphic (see Fig. 3.6). A smaller band, of less than 2 kb did appear to be polymorphic and was scored for segregation, but remains unlinked in the linkage map.
Clone h58 showed a particularly clear polymorphic band with DNA digested by TaqI, this enzyme was used to digest the DNA of the first 56 birds of the Compton mapping reference population. This showed the segregation of the 3 kb band, see Fig. 3.7 inherited from the N line. Clone h60 hybridised to two polymorphic bands in the parental DNA digested with MspI. However, when the D and A blots were probed with this clone the polymorphism was only detected in DNA from birds 69-83 (Fig. 3.8).
Fig. 3.9 shows the segregation of a HindIII 2 kb polymorphic band in the D and A blots probed by clone h71. Clone h78 did not detect any polymorphisms between the two parent birds, line 15I and line N x 15I, with 12 different enzymes. This clone was then used to probe a blot of line 6₁ and line 7₂ DNA digested with the panel of seven enzymes. This showed a variant band for TaqI, which was used to digest DNA from the MDV reference population. Fig. 3.10 shows the segregation of a band in the TaqI MV blot probed with clone h78.
The segregating progeny were scored for each polymorphism (results are tabulated in App. B). Analysis was carried out using MapManager ver. 2.6.5 (Manly and Cudmore, 1988), and the position of the EST within the chicken linkage map was identified. Fig. 3.11a shows the linkage map for chromosome 1, with genes CNBP, HMG1, Mouse1 and Man1 highlighted. Mouse1 is 26.8 cM from locus L18A, but is tightly linked to CNBP (7.2 cM), which in turn is only 19.6 cM from Man1. Man1 and HMG1 are separated by locus CMYC28 and are 42 cM apart. CNBP, Mouse1 and Man1 have all been mapped by inversion of the original segregation data, indicating that the polymorphism is from the line 15I chromosome in the F₁ parent of the cross. Fig 3.11b shows the linkage group for chromosome 15 with the PSF locus, which is linked to the TUBA (alpha-tubulin gene) by 14.5 cM. Fig. 3.12a shows the linkage group for chromosome 6 with the GDID4 locus highlighted which is situated 30.9 cM away from the nearest marker COM125. Fig. 3.12b shows the position of ITPR2 on chromosome 7 situated 28.6 cM from locus COM49. Two clones, Man2 and 60S at present remain unlinked (UNL) to any other loci. This mapping information is summarised in Table 3.5.

Table 3.5 Map positions of cDNA clones

Comparative mapping information was also obtained for the ten novel chicken genes by searching: The Genome Database (GDB) at http://www.hgmp.mrc.ac.uk/gdb/gdbtop.html and The Mouse Genome Database (MGD) at http://mgd.hgmp.mrc.ac.uk/mgd.html.
3.3 Discussion
3.3.1 Analysis of clones from the bursal cDNA library
A large proportion of the clones sequenced, 20%, showed strong homology to tetracycline resistance genes. This amber-mutated tetracycline resistance gene is present in the P3 episome of the bacterial strain MC1061/P3, which is suppressed by the supF suppressor tRNA gene encoded in the vector pCDM8, which aids selection for transformed cells. The P3 episome originated from R factor 1822 of Pseudomonas aeruginosa (Olsen and Shipley, 1973). Unexpectedly, it appears that this amber mutated tetracycline resistance gene has inserted into the pCDM8 vector in almost 20% of the successful transformations in this particular library. This problem has not been seen in similar libraries and it would be possible to eliminate clones containing this gene by pre-screening.
Sequence identity to previously identified chicken and mammalian genes was seen in 24% of the clones. Of the clones showing homology, two show identity to the mammalian genes, GDID4 and ITPR2, which are molecules involved in the signal transduction pathway. Clone h24 is homologous to GDI, the bovine GDP dissociation inhibitor. This protein is a GDP dissociation inhibitor (GDI) for the ras related rho subtype proteins (Fukomoto et al., 1990). This gene has also been identified in both murine and human genomes and has been designated GDP dissociation inhibitor, clone 4; marker symbol GDID4 in the human genome. GDID4 has been mapped to human chromosome 12 in region 12p12.3 (Adra et al., 1994), but at present remains un-mapped in the murine and bovine genomes. Clone h29 is homologous to the rat inositol 1, 4, 5-triphosphate receptor, type 2 (ITPR2) (Mignery et al., 1990). Inositol triphosphate is a small water soluble molecule which releases calcium ions from intracellular stores. ITPR2 receptors lie on the compartment surface and open gated calcium ion channels when activated. ITPR2 has also been located on human chromosome 12 at region 12p11 (GDB).
Clone h42 is homologous to the cellular nucleic acid binding protein (CNBP). This gene codes for a protein which binds to the sterol regulatory element (SRE) to repress transcription, while itself being regulated by sterols. It is a zinc finger protein of the Cys/Cys-His/Cys family, with an affinity for the octa-nucleotide SRE motif (Rajavashisth et al., 1989). This gene is found on human chromosome 3q13.3 (GDB) but is at present un-mapped in the mouse (MGD).
Clone h58 is homologous to the human gene for polypyrimidine tract-binding protein (PTB)-associated splicing factor (PSF), known by its marker symbol SFPQ. This protein is an RNA binding protein with properties identical to PTB. It binds the polypyrimidine tracts of pre-mRNA to form splicing complexes for the removal of introns (Patton et al., 1993). At present this gene is un-mapped in man (GDB) and unknown in mouse.
Clone h71 is homologous to rat high-mobility group (non-histone chromosomal) protein 1 gene (HMG1) (Paonessa et al., 1987). HMG1 protein has three structural domains, two of which are structurally similar and contain a repeated region of 85 aa, which can bind DNA, termed an HMG domain. It binds to DNA with little sequence specificity, but recognises both four-way junction and kinked DNA and can induce bending. However, the protein's function is currently unknown (reviewed by Grosschedl et al. (1994). It has been characterised in some eukaryotic organisms but its gene remains un-mapped in the human, cat, hamster, pig and rat databases. However, HMG1 has been mapped to mouse chromosome 5 position 52.0 cM.
Clone h54 shows homology to the mouse clone M1537, a cDNA clone isolated from mouse liver. Clone M1537 is located on mouse chromosome 3 (Warden et al., 1993), but the function of the corresponding gene product is currently unknown. Clone h60 is homologous to cDNA clone (hm02h02) generated from a HepG2 cell line derived from a liver cancer (Okubo et al., 1992). Both the function of this gene product and the gene's position in the human genome are currently unknown. Clone h78 is homologous to the human ribosomal phosphoprotein P2 gene. It codes for one of three proteins which are associated with 60S ribosomes (Rich and Steitz, 1987). None of the P-protein genes have been mapped in other animals.
3.3.2 RFLP Analysis and Mapping of clones
One of the identified novel chicken genes (h86) showed no polymorphisms in either mapping family, with any of the twelve restriction endonucleases tested (see App. A). RFLP searches do not always reveal polymorphisms, as they are reliant on either a base change in the restriction recognition sequence, or a substantial deletion or insertion which would alter the length of the fragment detected. A possible means of mapping this gene would be by single stranded conformational polymorphisms (SSCP) (Orita et al., 1989b). This technique has been used in similar circumstances for the detection of polymorphism in five out of eight bovine genes, for which no polymorphisms had previously been found (Neibergs et al., 1993). Clone h86 is homologous to collagen [alpha]1 chain type IV gene (COL4A1). Type IV collagen is a major constituent of basement membranes, comprising two [alpha]1 and one [alpha]2 chains (Brazel et al., 1987). Thus COL4A1 is a housekeeping gene. The murine and human homologues do show RFLPs which map to human chromosome 13q34 and mouse chromosome 8, 5 cM distance from Insr, the insulin resistance gene (GDB, MGD).
One of the clones (h56) remains unlinked on the chicken linkage map, this is because the original polymorphism detected was absent in the Compton mapping reference population. There appeared to be another polymorphism segregating in the progeny which was scored, but this was not genuine as it was unlinked on the linkage map. Almost every clone placed in the database is now linked to at least one other locus. Less than 2% of loci are unlinked. Clone h56 is homologous to an anonymous human clone identified by exon trapping whilst searching for genes associated with meningioma. It is located in the human genome on chromosome 22q12 less than 80 kb from the neurofibromatosis 2 gene (NF2), a tumour suppressor gene (Xie et al., 1993).
The TaqI polymorphism which segregated in the MDV reference population for the 60S clone is not at present linked to any loci. At present only 70 loci are mapped in this population, probably covering less than half the genome. As more loci are placed on the linkage map it should become possible to identify the positions for new markers.
The polymorphism detected by clone h60 only segregated in 15 birds. These birds had a different line N x 15I female parent (bird 2818) from the rest of the Compton mapping reference population. This is rare since the birds are all generated from the same inbred lines, however (as shown in App. A) line N has a coefficient of inbreeding of only 0.52, much lower than the other inbred lines. While the main portion of the Compton mapping reference population is generated from a single F₁ bird, and hence from a single line N chromosome, it appears that the other F₁ bird used as a female parent for 15 of the birds in the backcross differ in some of their line N alleles. This has been seen occasionally in previous hybridisations with the DNA from these birds.
Overall seven of the ten novel chicken genes were successfully mapped to four chromosomes. This is a good result as it has previously been estimated that 50% of cDNA clones detect polymorphisms and can be mapped (Bumstead et al., 1994a).
ESTs can be a very powerful means of identifying genes since there are already very many characterised mouse and human genes. The function of some of these genes is well understood, and can be inferred on the identification of a new chicken homologue. Thus genes, whose function is known in other organisms, can be further investigated as candidate genes for particular traits. However, success for this comparative approach is not guaranteed, as for example one allele of Nramp1 confers resistance to salmonellosis in mice (Vidal et al., 1993) but not in chickens (Hu et al., 1996). ESTs can also be used in comparative mapping between species, to identify a putative candidate gene-bearing region. Thus if two ESTs in the chicken map are linked to each other and their linkage is conserved in another well mapped species such as mouse, this may indicate the presence of genes relevant to the trait of interest located between these loci in the mouse map.
Although there are many benefits of EST mapping, including comparison of synteny and the identification of potential candidate genes, it is at present laborious and difficult to map the sequenced clones. For this reason mapping of EST clones was not pursued further as a means of saturating the current linkage map. However with the generation of a chicken radiation hybrid map EST mapping will become more efficient. Radiation hybrid maps are based directly on chromosomal linkage and do not require the identification of polymorphisms. If a radiation hybrid map can be established it may be possible to infer the position of candidate genes for disease resistance by comparison of synteny of loci between the chicken map and those maps of humans or mice.
However, an easier method is needed to saturate the linkage map with markers. This is necessary for positional cloning of any candidate gene, as there need to be enough markers in order to align YAC clones into a contig, for subsequent physical mapping.

4.1 Introduction
After investigating ESTs as ideal genetic markers, it was decided that a less laborious approach was needed to saturate the linkage map. Random amplified polymorphic DNA (RAPD) analysis is well suited, as it requires no previous sequence knowledge and theoretically produces large numbers of polymorphisms which can be used to saturate the linkage map with markers. RAPD products are generated by the binding and subsequent PCR amplification from a single short arbitrary primer. This technique was developed by Williams et al. (1990) and Welsh and McClelland (1990) to utilise the decreased specificity in binding at low annealing temperatures for short oligonucleotides (usually less than 10 nucleotides). Large numbers of fragments are amplified by this method, some of which are polymorphic. RAPD offers the possibility of generating large numbers of polymorphic markers relatively cheaply which can then be mapped.
RAPD markers have been used in many species for a variety of investigations: gene cloning, medical diagnostics and trait introgression in breeding programs (Williams et al., 1990). Levin et al. (1993) used RAPD markers to generate new markers on the Z chromosome of the chicken in order to identify sex-linked traits. RAPD markers have also been useful in determining phylogenetic relationships between species as demonstrated by Barral et al. (1993) with the Shistosoma genome.
RAPD markers have already been generated for use in chicken mapping by Levin et al. (1993) and Plotsky et al. (1993). There are now 68 RAPD markers on the East Lansing chicken linkage map (Burt et al., 1995), 16 of which are located on the Z chromosome (Levin et al., 1993). This shows that RAPD analysis is a viable way of generating markers for the chicken genome, however, in Levin's work RAPD only detected a low number of polymorphisms. Increasing the resolution of detection could however increase the number of potential polymorphic markers. To attempt to improve the yield of polymorphisms I have used fluorescently labeled primers and resolved the RAPD fragments by polyacrylamide gel electrophoresis (PAGE) using a 373A ABI Sequencer and analysing the products using Genescan software.
4.1.1 Choice of RAPD Primer
Two primers were chosen initially, one which had already generated polymorphisms as a 10-mer and one which was derived from the only characterised chicken middle repetitive element (CR1). Oligonucleotide 85 was a 9-mer version of the 10-mer oligonucleotide OPP-02 used by Levin et al. (1993) 5'-labeled with FAM. Oligonucleotide 78 was a 9-mer selected using PRIMER software (version 0.5, Whitehead Institute for Biomedical research 1991) from the 3' end of the CR1 element sequence in the [beta]-globin cluster as described by Reitman et al. (1993), 5'-labeled with HEX. The initial protocol was optimised for these fluorescently labeled primers. CR1 elements are common throughout the chicken genome and therefore primers derived from within these repeats would be expected to find particularly large numbers of binding sites.
CR1 elements are non-long terminal repeat (LTR) retrotransposons (Burch et al., 1993), with a pol-like open reading frame encoding reverse transcriptase which is responsible for dispersal of the element throughout the genome. The CR1 element contains two highly conserved regions but is very polymorphic in the rest of its sequence, making it an ideal sequence from which to generate RAPD primers. There are an estimated 100,000 copies of CR1 distributed throughout the chicken genome (Vandergon and Reitman, 1994), which will ensure that RAPD markers generated from CR1 primers will also be spread throughout the genome. However these elements are not randomly distributed but occur mostly in G-C rich regions of the genome (Olofsson and Bernardi, 1983), which tend to be the most gene-rich regions (Bird, 1986). Thus primers generated from CR1 should be capable of giving RAPD marker loci which will be distributed throughout the genome, probably near to gene-rich regions. Levin et al. (1994a) also designed primers from the CR1 element, from the 5' end to the consensus 3' end of the CR1 repeat adjacent to the ev21 (endogenous virus locus) integration site. However, oligonucleotide 78 shared no homology with those in the study of Levin et al. (1994a).
4.2 Results
4.2.1 Creation of the Matrix File
The matrix file was created to reduce the effects of interactions of fluorescent tags and eliminate background interference which can occur on analysis of Genescan gels as detailed in the Methods, section 2.6.2.1. This ensures that the resolution of the products is as clear as possible, by first analysing all the ABI fluorescent products and defining limits for the detection of each colour in a matrix file. The matrix file was defined over 1000 scan points containing a minimum of three good peaks. These were obtained from the gel shown in Fig. 4.1. together with data from the ROX 2500 standard. Matrix values were obtained as shown in Table 4.1 below. The matrix file was saved under the filename: 5.5% LR 1 x TBE matrix and checked for validity against another Genescan gel. Background noise was greatly reduced on the application of this matrix to the RAPD gels and the matrix was used for all gels.
Table 4.1 Matrix values used to create "5.5% LR 1 x TBE matrix" gel file.

14.2.2 RAPD Analyses using fluorescent primers
Oligonucleotide 85 was used to generate RAPD products for the two inbred line birds B981(line 15I) and B984 (line N), as well as the alternative parent birds C54 (line 15I) and D551(line N x line 15I) and twenty progeny from the Compton mapping reference population (see Methods section 2.11.2.1). These products were electrophoresed on an ABI Genescan gel with 1.7 µl of each 25 µl PCR reaction loaded with 1 µl denaturing buffer and 0.7 µl Genescan ROX 2500 standard in each lane as detailed in the Methods section 2.6.2.2. The gel image is shown in Fig 4.2. Many bands can be seen in the relatively narrow size range selected, some of which are clearly polymorphic between the two inbred lines. The arrow indicates the polymorphic band mapped as locus COM142. Electropherograms for the two lines were aligned and polymorphisms identified for 12 of the RAPD products; 4 fragment present only in the line N birds and 8 fragments present only in the line 15I birds (see Fig. 4.3). The scan number and ROX 2500 standard position were used to define the polymorphic band in each lane. However only one of the RAPD products was polymorphic for the parents of the Compton mapping reference population, and this was the only one to segregate in the progeny, giving a chromosomal location for the locus. This locus, designated COM142, lies on chromosome 5 (LOD>3), 33.2 cM from the transferrin receptor gene (TFRC) and 28.2 cM from the microsatellite locus MCW135 (see Fig. 4.4). Oligonucleotide 78 was also tested for its ability to reveal polymorphisms in the two pure lines, the alternative parents and the first 20 progeny. The products were analysed (see Methods section 2.6.2.2) by ABI Genescan gel with 3 µl of each 25 µl PCR reaction loaded with 1.5 µl denaturing buffer and 0.5 µl Genescan ROX 2500 standard in each lane (see Fig. 4.5). Although many bands were again generated, no polymorphisms were observed between the two lines or the parent birds.
2345In order to try and identify polymorphisms with primer 78 the template DNA was pre-digested prior to amplification (Levin et al., 1993) using the enzymes: EcoRI, XbaI, RsaI, TaqI, HaeIII or MspI. The products were analysed on an ABI Genescan gel, 2 µl of each 25 µl PCR reaction was loaded with 1.6 µl denaturing buffer and 0.4 µl Genescan ROX 2500 standard in each lane. The results of RAPD analysis on the two inbred lines are shown in Fig. 4.6. Although the numbers of products generated from the different pre-digestions varied, no clear polymorphisms were detected between the two lines.
It is possible to identify further polymorphism by combining multiple RAPD primers in a single reaction (Levin et al., 1994a). To explore this, oligonucleotide 85 was used in equimolar amounts with combinations of four primers from Kit 1 (Advanced Biotechnologies Ltd). Reaction products were initially generated from DNA of the two inbred lines and the alternative parent (D551:line N x line 15I) and visualised by agarose gel (see Fig. 4.7). Using two primers gave RAPD products only in certain combinations (see Fig. 4.7 section 5). When three primers are combined (see Fig. 4.7 section 3) almost double the number of bands are generated with two primers. However this analysis also did not reveal any polymorphisms for these primer combinations.

4.3 Discussion
Oligonucleotide 78 generated no polymorphisms which was unfortunate as in a comparable experiment by Levin et al. (1994a) only one out of ten primers failed to generate any polymorphisms. Oligonucleotide 85 was designed from a primer which was known to give RAPD polymorphisms (Levin et al., 1993) and thus acted as a control for the RAPD analysis. Oligonucleotide 85 successfully generated 12 polymorphisms between the inbred lines tested here.
It was expected that the two primers chosen for fluorescent labeling, oligonucleotides 78 and 85, would have generated large numbers of polymorphic products, because the resolution of the products in this system is much better than by agarose gel electrophoresis. Detection of larger numbers of products should mean the detection of increased numbers of polymorphisms. Primers previously designed from the CR1 element have successfully identified RAPD polymorphisms (Levin et al., 1994a), it is surprising that oligonucleotide 78 did not.
The generally low level of polymorphisms of RAPD products seen here have been found in other studies. Thus, in the study of Levin et al. (1994a) out of a total of ten primers, one primer failed to generate a product under a variety of conditions, whilst three failed to show any detectable polymorphisms when used singly. However primers which did generate polymorphic bands, revealed more than one polymorphism per primer. In the study of Woodward et al. (1992) RAPD-PCR analysis was carried out by screening two different strains of mouse with 481 arbitrary 10-mer primers. Of these, 75 primers identified a total of 95 polymorphisms, 76 of which were mapped. Rothuizen and Van Wolferen (1994) in their study of RAPDs in the canine genome, screened sixteen breeds of dog with six primers, five of which identified a total of 11 polymorphisms. Although some primers identifying polymorphisms may be very productive, a large number need to be screened in order to find those capable of showing multiple polymorphisms.
Primers identify RAPD polymorphisms when the binding sites differ between the two samples tested. Primer binding depends on a number of factors, the sequence of the binding site, the concentration of magnesium ions and dNTPs and the annealing temperature. In these RAPD reactions the concentration of magnesium ions and dNTPs is that of normal PCR reactions, ensuring optimum specific binding. However, the annealing temperature (37°C) is significantly lower than for typical PCR reactions. This increases non-specific binding, which reduces the number of potential polymorphisms which could be detected. Primer extension is highly specific for the 3'-end but not the 5'-end of the primer, which again will cause a decrease in the detection of polymorphisms which occur in the binding sequence at the 5'-end of the primer. So polymorphisms between two samples would only rarely be detected by RAPD analysis if generated by a point mutation. However, if the variation between inbred lines was due to a significant change e.g. a deletion, RAPD analysis would easily reveal polymorphisms.
Primer combinations have been successfully used to identify further polymorphisms. Levin et al. (1994a) used ten primers in paired combinations, to generate a further twelve polymorphisms. On average, 1.1 polymorphisms were generated per single primer and 0.3 polymorphisms per primer pair combination (Levin et al., 1994a). Rothuizen and Van Wolferen (1994) in their study, also used paired combinations of six primers, fifteen paired combinations generated a further 44 polymorphisms. However this method failed to generate further polymorphic products when tried for combinations of oligonucleotide 85 and a number of random 10 mers.
Another alternative for identifying more polymorphisms with a single primer is the use of pre-digested template DNA. This utilises the RFLPs present in the template to reveal other polymorphisms when detected by RAPD. Levin et al. (1993) used the restriction endonucleases EcoRI and HaeIII to pre-digest their genomic template. This generated a further polymorphism, in the case of one primer, in each of the pre-digested templates. With another primer, two new polymorphisms were generated from EcoRI and HaeIII digested template. However when oligonucleotide 78 was used for RAPD analysis on predigested samples from the two inbred lines, it again produced no polymorphisms. It is possible that oligonucleotide 85 may have succeeded in identifying more RAPDs in this way, but this was not tested.
Using fluorescently labeled RAPD products, analysed by the Genescan program greatly increases the number of PCR product bands which can be resolved per gel in comparison to other studies (Woodward et al., 1992; Levin et al., 1994a; Rothuizen and Van Wolferen, 1994). Using this system it is possible to resolve more than 50 bands in comparison to a conventional system where a maximum of 30 bands can be resolved (Rothuizen and Van Wolferen, 1994). The ABI system also allows for easy analysis of the RAPD segregation patterns. This is due to better resolution obtained by a Long Ranger gel, and the ability to analyze the data using ABI Genescan software. It is also possible to use different fluorescently-tagged primers in various combinations, both during the PCR reaction and the electrophoresis. This enables the analysis of large numbers of samples simultaneously.
Although a total of twelve polymorphisms were generated between the two inbred lines only one segregated in a clear and reproducible manner. This is a similar number to results obtained previously by Woodward et al. (1992) and Levin et al. (1994a). The segregating locus was mapped within the Compton mapping reference population. It has also been reported that changes in enzyme manufacturer, primer or enzyme concentration, or thermal cycling equipment, can give rise to inconsistent results (MacPherson et al., 1993; Meunier and Grimont, 1993; Schierwater and Ender, 1993). Thus RAPD analysis is not a reliable way of easily and cheaply producing large numbers of polymorphic loci with which to saturate the linkage map in chickens. Because RAPD markers are not targeted it would be necessary to produce a lot of markers, however this cannot be done easily, so RAPDs are not a viable solution to the problem of how to increase marker density on the chicken map. It was decided not to continue with RAPD analysis for these reasons.
RAPD analysis has classically been used for saturation mapping of the genome. However, RAPD analysis has also been successfully targeted to specific chromosomes. Giovannoni et al. (1991) used bulked segregant analysis of F₂ tomatoes to target a region on chromosome 11, by using two pools of DNA from individuals homozygous for opposing alleles on that chromosome, all other loci being inherited in a random manner. 200 primers were used and 3 RAPD markers were identified, two of these were tightly linked to the region of interest. Bulked segregant RAPD analysis was also shown to be useful in targeting the region containing the downy mildew resistance gene in lettuce (Michelmore et al., 1991). RAPD markers were successfully placed in a 50 cM area around the resistance gene. This use for RAPD analysis has recently undergone a revival as shown by Horvat and Medrano (1996) who succeeded in targeting the region surrounding the high growth (hg) locus on mouse chromosome 10 again using RAPD markers with bulked segregant analysis.
RAPD analysis has also proved useful for population genetics and other studies requiring identification of variation. RAPD analysis was used to determine a consensus dendrogram for chickens and turkeys (Smith et al., 1996), using a set of 60 primers to identify polymorphisms between four chicken breeds and two turkey populations. As expected the lowest genetic distances were within species, while the highest were between species. Barral et al. (1993) also used RAPD to analyse genomic diversity, of the Schistosoma spp. both between and within species. RAPD analysis produced from 0 to 20 amplified fragments, which enabled the identification of one useable interspecific polymorphism for 5 species and one useable intraspecific polymorphism for 3 strains.
An extension of the RAPD method for the identification of species or strain specific RAPD polymorphisms has incorporated the use of pooled DNA samples in order to simplify the screening process.
Bailey and Lear (1994) used pooled samples of 10 thoroughbred and 10 Arabian horses in order to identify RAPD primers which would give breed specific alleles. This technique has also been applied to the identification of primers which show intraspecific polymorphisms between Bos indicus and B. taurus cattle (Kemp and Teale, 1994).

5.1 Introduction
The investigations described in chapters three and four as means of generating markers have not proved useful. Therefore, an alternative method was needed to generate large numbers of targeted markers quickly and easily. Representational difference analysis (RDA) appeared to fulfill these criteria. Thus the potential of RDA was investigated.
RDA has not previously been used with the chicken genome, so as an initial test of the method two inbred lines of chickens were compared. This provides a relatively undemanding test of the technique, since very many differences will exist between birds of the two lines. The lines chosen, line N and line 15I, are the parent lines of the Compton mapping reference population and hence polymorphic markers identified in these lines can be readily incorporated in the genomic linkage map. The lines also differ in their susceptibility to salmonellosis and further markers suitable for mapping progeny of this cross will be useful in mapping this trait.
5.2 Results
5.2.1 Comparison of line N and line 15I
Amplicons for the tester and driver birds were generated from BamHI digested genomic DNA and the adaptors removed as described in the Methods section 2.12.1. A 10 µl sample of each amplicon was analysed by agarose gel electrophoresis (see Fig. 5.1) to confirm PCR amplification. The BamHI amplicons ranged in size from less than 0.56 kb to just over 2.3 kb. After removal of the original adaptors, tester amplicons were re-ligated to new adaptors (JBAM oligonucleotides 66 and 69, see App. A).
Three different ratios of tester and driver amplicons were hybridised:
a) 4 µg of tester : 40 µg of driver,
b) 1 µg of tester : 40 µg of driver,
c) 0.4 µg of tester : 40 µg of driver.
After hybridisation the samples were selectively amplified by PCR using program RDA1 (see App. A) while removing samples for analysis by agarose gel electrophoresis at each denaturation step from 13 until 20 cycles. The best ratio, giving the greatest amount of product was 4:40 of tester to driver. Fig. 5.2 shows that amplified product was visible from the 17^th PCR cycle for sample for this ratio of tester:driver, with products ranging from 0.5 to 2 kb. Product from the sample was taken after the 20^th PCR cycle and cloned into the pGEM-T vector. The remaining RDA products were purified and any single stranded DNA removed before a second round of PCR (the re-amplification step) was performed. These products were analysed by agarose gel electrophoresis and are shown in Fig. 5.3. The size of the RDA fragments had decreased to less than 1 kb during the replication process, with a mean of approximately 0.6 kb.
After removal of the old adaptors by digestion, the RDA products were ligated to new adaptors (NBAM oligonucleotides 67 and 65, see App. A). For the second round of hybridisation 100 ng of RDA product was reannealed to 40 µg of the driver. After hybridisation two duplicate tubes were selectively amplified using NBAM 65 as the primer (PCR program RDA2 App.A). Samples were removed from the PCR reaction after every other PCR cycle and analysed by agarose gel electrophoresis. Fig. 5.4 shows 10 µl samples taken from PCR cycles 10-24. It was decided to continue only with products generated at 25 amplification cycles, as these gave the most amplified products. Single stranded DNA was removed from the amplified products which were then reamplified using primer NBAM 65, shown in Fig. 5.5 ( see Methods section 2.12.1). These products were then cloned into the pGEM-T vector as described in the methods (2.2.3).
5.2.2 Colony PCR of RDA products
Individual colonies of the transformed bacteria were analysed by PCR to identify insert size and generate products for subsequent probes (see methods section 2.7.2). Colony PCR products were analysed by electrophoresis on a 1%, TBE agarose gel, as shown in Fig. 5.6. The primers used were 100 bp from the cloning site and so any product less than 220 bp was generated from plasmid without insert and therefore discarded.
5.2.3 Screening and characterisation of RDA clones
Genomic DNA from a line N and a line 15I bird were each BamHIdigested and Southern blotted onto Hybond nylon membrane and probed with the colony RDA PCR products. A total of 87 clones, with an average insert size of 288 bp, were initially investigated (see Table 5.1).
Table 5.1 Results from initial hybridisation of RDA clones to Southern blots.

Some of these, approximately 59%, did not hybridise to the Southern blots and were not further investigated. Those that did hybridise showed a variety of banding patterns. Some clones hybridised to single bands, but most hybridised to a number of bands, usually including an unique small band present only in the tester (see App. C for details). Fig. 5.7 shows the hybridisation patterns for a number of the RDA clones. Clone hR7 is a typical example of a non-polymorphic clone which hybridises to a small fragment in both the driver and tester (designated with the symbol @ in App. C). Clones hR17 and hR2 are typical examples of the types of hybridisation patterns obtained from these RDA products. Several apparently duplicate clones were generated, but most were non polymorphic.
5.2.4 Mapping RDA Clones
A total of five clones were used to probe segregating progeny blots in order to map them. The segregation of the RFLPs of clone hR2 within the first 56 progeny are shown in Fig. 5.8. For this clone four segregating bands were scored which map together on chromosome 16, designated loci COM101-104. Fig. 5.9 shows the segregation of hR17 within 57 of the progeny. Two bands are present in the line N birds and segregate identically, while there was no hybridisation to the line 15I birds. This clone was mapped to chromosome 16 designated locus COM155. Clone hR69 hybridised to three segregating bands which are tightly linked on the W chromosome (loci COM194-196). Clone hR51 hybridises to two segregating bands and maps to chromosome 4 (loci COM179 and COM181). Clone hR29 hybridises to six bands, two of which show polymorphisms, but at present remains unlinked on the map.
5.2.5 YAC Hybridisation
Clone hR17 was used to screen the chicken YAC library (Toye et al., In Press). The YAC colony blots consist of a grid of spotting squares. Each YAC clone is spotted onto the square twice in the orientations shown. To verify hybridisation on the colony blots the clone must hybridise to two spots within the spotting square in a specific orientation (see Fig. 5.10).

5.2.6 Nucleotide Sequences
All 15 polymorphic clones were sequenced from each end. Many clones contained repetitive elements including clone hR17 which contained a series of seven ATTAG repeats in its 145 bp sequence (see Fig. 5.12).

5'-GGATCCATTAGAGCCCGTCAGATCCCATTAGAGTCCATCAGATCCC
 ATTAGAGCCCGTCAGATCCCACTAGAGCCCGTCAGATCCG
ATTAGAGCCCATTAGGGCCAATTAGAGCCCATCAGATCCT
ATTAGAGCCCGTCGGATCC

Sequence data from clone hR17, using primer GEMF, showing all 145 nucleotides with seven underlined pentameric repeats. Recognition sequences for BamHI are shown in italics.
Clone hR20 hybridised in the same RFLP pattern as clone hR17, however, it proved to be 50 bp longer containing a total of 12 ATTAG repeats. Clones hR2, hR8 and hR11 were identical at the 155 primer end, but differed in length hR11 being approximately 170 bp, hR8, 226 bp, and hR2 longer than 400 bp. Clone hR8 contains five TGGTTGGGTTGG dodeca-nucleotide repeats within its sequence. None of these clones identify any known sequences within the EMBL database.
Some of the other clones identify sequences containing alpha satellite type repeats; for example clone hR33 shows 60% identity with a sequence from Xenopus laevis which contains these repeats. Clone hR33 contains 22 ATGGG repeats within its 340 bp sequence. Clone hR36 shows a similar degree of identity to a human alpha satellite sequence and actually contains seven CAATGG repeats within its 200 bp sequence. Clone hR29 contains at least two sets of repeats within its sequence of 410 bp; 25 repeats of CATTG and 18 repeats of AATGGG.
The only identity found for any sequence was with clone hR32. This clone showed 60% identity with the pentameric CATTG repeat present in the telomeric region. This clone contained 33 repeats in 530 bp of sequence.
5.3 Discussion
These results show that this technique can easily be applied to the avian genome. In comparisons between the two inbred lines (N and 15I) BamHI polymorphisms are normally detected for 7% of random genomic clones, the use of RDA has increased this percentage to 47%. Products were cloned from both rounds of hybridisation, although most clones were obtained from the second round as many of those from the first round did not hybridise to chicken DNA. The reason for this is unclear, as the clones were selected by blue/white selection and subsequent PCR amplification of inserts. However the transformation efficiency was 100 fold less than predicted by the manufacturer (Promega), although self ligation was very low and the positive control gave approximately 50% white colonies, as predicted. A similar problem occurred when Delaney et al. (1995) used RDA to target mapping of rye chromatin in wheat. Approximately 50% of their clones also failed to hybridise to tester and driver DNA, it was assumed that this was due to external contamination of the samples.
Due to a technical error the 3 x EE buffer used during the annealing stage was not adjusted to the correct pH. Thus the pH was 5.5, when it should have been pH 8.0 according to the methods of Straus and Ausubel (1990). DNA which has previously been denatured, depurinates rapidly at low pH values and so should preferentially be kept at a neutral pH value. Lindahl and Nyberg (1972) showed in their experiments that in a simulated physiological buffer using previously denatured DNA the rate of depurination was much slower at pH 7.4 than at pH 5.0. Due to the long annealing times at the relatively high temperature of 67°C it would be expected that at pH 5.5 an amount of DNA would become depurinated in these experiments. However, as good results were obtained it was decided to continue with the same buffer throughout all subsequent experiments, so that controlled comparisons could be made.
RDA generates two types of polymorphic products. One type is totally absent in the driver, present only in the tester, an example of which is clone hR17. The other, more common type, occurs when a restriction site is absent in the driver, probably due to a point mutation, but present in the tester. This means that after digestion the tester contains a small fragment which is present as a large fragment in the driver. When the digested fragments are PCR amplified, the smaller fragments will be preferentially amplified in comparison to the larger fragments. Thus the driver amplicons will not contain some of the large fragments which are present as small fragments in the tester amplicons. These RDA products therefore hybridise to small fragments in the digested tester DNA while hybridising to large fragments in the digested driver DNA.
A number of the RDA clones contained repetitive sequences, in fact eight out of the 15 clones which identified polymorphisms contained repeats of 5 to 12 nucleotides in length. A similar proportion of RDA clones containing repetitive sequences has also been found by Delaney et al. (1995), where 11 out 43 clones contained repetitive sequences. It would have been possible to remove these repetitive sequences by prescreening the RDA library with repetitive DNA, for example the chicken middle repetitive element (CR1).
RDA generated large numbers of products which were cloned and screened. At present only five have been typed on the Compton mapping reference population, with ten still needing to be typed. One clone remains unlinked as it has only been typed on part of the Compton mapping reference population.
Clone hR17 was used to probe the gridded chicken YAC library as it hybridises to two small bands only present in the tester. This clone also showed strong linkage to chromosome 16, the MHC-bearing micro chromosome. This linkage was confirmed as four of the YAC clones identified had previously been isolated by hybridisation to MHC genes.
This test of the RDA technique has provided large numbers of clones, almost 50% of which are polymorphic between the two inbred lines. Some of these clones have been mapped and have provided an extra 10 loci for the chicken genetic map. It appears RDA is very useful tool for the production of genetic markers and this has shown that it can be equally applied to the avian genome, as the mammalian one. RDA generates significantly more clones which detect polymorphisms in comparison to other mapping techniques, such as EST mapping and RAPD analysis.

6.1 Introduction
Once RDA had been established as a viable technique for use in the avian genome a second comparison was made targeting loci to chicken chromosome 16. This provides a more stringent test of the method, and will provide marker loci in a region of particular interest.
Others have targeted specific chromosomes in different organisms using RDA with success. Lisitsyn et al. (1994) used congenic strains of mouse to target a 15 cM region (equivalent to 30 Mb in the mouse) using a pair of strains as the tester and another pair as the driver DNAs. A total of 18 clones were characterised from six experiments, fifteen of which were eliminated in the first round of screening. Seven hybridised to multiple band patterns, seven detected fragments in both tester and driver amplicon blots and one failed to hybridise. In this experiment only fragments which were present exclusively in the tester amplicons were used for subsequent linkage analysis. One of the remaining clones was also discarded as it detected multiple loci. The two remaining clones were mapped and one maps to within 3 cM of pudgy while the other maps to a 7 cM interval consistent with tottering. Lisitsyn et al. (1994) also targeted a region of mouse chromosome 11 containing the nude locus using the DNA from 12 F₂ intercross progeny which had crossovers between the nude locus and closely linked markers as the drivers and DNA from the parental strain as the tester. After three rounds of hybridisation only two bands were visible and these were cloned. One hybridised to multiple band patterns while the other was mapped to within a 1.3 cM region. In fact this clone was subsequently proved to lie within 0.2 cM of nude. A different representation gave a further two markers also within 0.2 cM of nude.
After the success of Lisitsyn et al. (1994) a number of other groups have used RDA to target specific chromosomal region. Delaney et al. (1995) targeted a region of rye chromosome 6 which contained a resistance gene. Initially a wheat-rye line containing this region was used as the tester against a pure wheat line, in another experiments the role of tester and driver were reversed. Three separate experiments were performed which resulted in over 108 cloned fragments, of which eight were potentially useful. Five of these clones gave seven loci which mapped to chromosome 6. Baldocchi et al. (1996) used RDA to target mouse chromosome 10 in the region of the jcpk locus. They altered the original protocol by size selection of amplified fragments in order to enrich each size pool. They generated a total of 39 clones, 31 of which were polymorphic although six of these were duplicates. Of these seven were linked to jcpk within 6.5-9.1 cM interval.
Chicken chromosome 16 is a microchromosome which carries the major histocompatibility complex (MHC) (Pink et al., 1977), the chicken's only nucleolar organiser region (NOR) (Bloom et al., 1993), and the Rfp-Y complex (Miller et al., 1994b) and is therefore of considerable immunological interest. Within the Avian Genetics group we are currently attempting to construct a YAC contig of this microchromosome and additional genetic markers will be very valuable for this purpose. Estimates of the size of the microchromosome range from 8 Mb upwards (Kaufman et al., 1990), however the NOR occupies a large portion of the chromosome, perhaps as large as 6 Mb (Bloom and Bacon, 1985). The targeted region is therefore only approximately 2 Mb or 0.17% of the genome. If the RDA method is able to identify markers specific to this microchromosome it is likely that it can be generally applied to other specifically targeted regions, such as those bracketing mapped trait genes. To generate markers specific for chromosome 16 a bird was selected from the Compton mapping reference population which had inherited line N alleles for the MHC, NOR and Rfp-Y regions on this chromosome. DNA from this bird was compared with pooled DNA from sixteen of its siblings, all of which had inherited line 15I alleles for the MHC, NOR and Rfp-Y regions. This choice of birds is shown diagrammatically in Fig. 6.1.
At the same time as targeting chromosome 16, the optimum ratio of driver to tester amplicons for hybridisation was investigated as well as the optimal hybridisation time needed to generate polymorphic RDA clones and the effects of using three different restriction endonucleases: TaqI, BamHI and NheI. These give representations of the genome generated by a frequent cutter (TaqI), an intermediate cutter (BamHI) and an infrequent cutter (NheI). TaqI recognises a four base sequence which occurs more frequently than other restriction endonuclease sites in the chicken genome, giving fragments of more than 9 kb to less than 0.6 kb. BamHI recognises a hexanucleotide sequence which is present in a number of chicken genomic sequences giving fragments in the range of more than 26 kb to less than 2 kb. NheI recognises a hexanucleotide sequence which is represented infrequently in the chicken genome, giving fragments ranging from more than 26 kb to less than 4 kb.
6.2 Results
6.2.1 BamHI Representation
Amplicons were generated individually from each of the 16 driver DNA samples and the single tester sample. Equal amounts were determined by spectrophotometry readings as well as by agarose gel visualisation, shown in Fig. 6.2. The amplicon fragments varied in size from more than 2.3 kb to less than 0.6 kb. Four different hybridisation conditions were tested using 1.14 µg of tester amplicon in:
1) a driver to tester ratio of 35:1 and a re-annealing time of 20 hours;
2) a driver to tester ratio of 35:1 and a re-annealing time of 48 hours;
3) a driver to tester ratio of 70:1 and a re-annealing time of 20 hours;
4) a driver to tester ratio of 70:1 and a re-annealing time of 48 hours.
After hybridisation duplicate 40 µl aliquots were taken from each diluted sample and selectively amplified.
This amplification was visualised by agarose gel electrophoresis of 10 µl samples, shown in Fig. 6.3. Apparently equivalent amplification occurred for 1&2&3each of the test conditions and also the tester control. The duplicate samples were combined, purified and resuspended in 40 µl TE (see Materials in App. A). Single-stranded DNA was removed from 20 µl aliquots, as described previously in the methods, which were subsequently diluted to 200 µl in 50 mM Tris-Cl (pH 8.5). Re-amplification of 40 µl aliquots was carried out in a 400 µl volume. The RDA products generated were cloned into pGEM-T, as detailed previously in the methods (2.2.3). However due to a technical problem in cloning, 113 out of 120 clones contained two copies of the plasmid. This would have caused problems in sequencing, as the primers would preferentially amplify across the plasmid containing no insert, so the clones were gel purified to obtain insert DNA for sequencing, as detailed in the methods (2.8).
6.2.2 TaqI Representation
Amplicons were initially generated by PCR amplification after the ligation of adaptors (RTAQ oligonucleotides A and B, see App. A) and subsequent pooling of the driver amplicons mixed in an equi-molar ratio as for the BamHI representation. Fig. 6.4 shows the distribution of the amplicons by agarose gel. The amplicon fragments varied in size from more than 2.3 kb to less than 0.6 kb. The RTAQ oligonucleotide adaptors were removed from all the amplicons and new adaptors (JTAQ oligonucleotides C and D, see App. A) ligated to the tester amplicon only as detailed previously in the methods. Hybridisations were carried out at 67°C for 48 hours (except for the first, which was only undertaken for 38 hours).
In the first round of hybridisation 0.4 µg of tester amplicon was added to adaptor-free driver amplicon in a driver:tester ratio of 200:1. A positive control of 0.4 µg tester amplicons and a negative control of 40 µg adaptor free driver amplicons were hybridised alongside the samples. After selective amplification with JTAQ oligonucleotide C, removal of single-stranded DNA as before and re-amplification by JTAQ oligonucleotide C, a sample of the first round products was cloned.
A second round of hybridisation was carried out using the remainder of the first round product. Adaptors were removed from the hybridisation products and new adaptors (NTAQ oligonucleotides E and F, see App. A) ligated, then a fresh 80 µg of driver amplicon was mixed with 50 ng of the first round product in a driver:tester ratio of 1240:1. After the same amplification steps as before but using primer NTAQ E, a sample of the products was cloned and the remainder used for further rounds of hybridisation.
The adaptors were removed from the second round hybridisation product and again new adaptors (JTAQ oligonucleotides C and D, see App. A) ligated.
Again, 80 µg of driver amplicon was mixed with 100 pg of the second round product in a driver:tester ratio of 4x10¹¹:1 (as recommended by Lisitsyn et al. (1993) for the third round of hybridisation. After similar amplification steps as before, using primer JTAQ C, a sample of the products was cloned into the pGEM-T vector.
RDA products from the re-amplification stage of each round of hybridisation are shown in the agarose gel in Fig. 6.5. As can be seen from the figure, with each subsequent round of hybridisation the number of large fragments in the product decreases, whilst a large pool of fragments less than 0.6 kb remain.
6.2.3 NheI Representation
RNHE oligonucleotides (G and H, see App. A) were ligated to the NheI derived fragments and after dilution, 40 µl aliquots were taken for amplification of amplicons using RNHE oligonucleotide G. The amplicons were visualised by agarose gel, shown in Fig. 6.6. Amplification was poor and unequal between the driver samples, so this representation was not pursued further.
4&5&66.2.4 Screening and characterisation of RDA clones
All clones were initially prepared by colony PCR, as detailed in the methods (2.72). BamHI derived clones were screened by hybridisation to Southern blots of BamHI digested line N and line 15I genomic DNA to determine those which were polymorphic between the lines. Similarly TaqI derived clones were screened by hybridisation to Southern blots of TaqI digested line N and line 15I genomic DNA. A total of 185 clones have been investigated; 120 from BamHI RDA products and 65 from TaqI RDA products. Table 6.1 shows the numbers of variant clones generated from each different set of hybridisation conditions and how many were successfully targeted to chromosome 16. Further details of all these clones can be found in App. D.
Table 6.1. Characterisation of RDA clones targeted to chromosome 16 generated by different hybridisation conditions.

However clones were also generated from the BamHI derived tester control which was allowed to self-anneal. This surprisingly generated eleven clones detecting polymorphisms between the two lines, these are not shown in the above table. However three of these were duplicates of clones generated from the other variations in hybridisation conditions.
Chi-square analysis showed no significant difference in the proportion of clones which showed multiple band patterns for the different hybridisation conditions. However there was a significant increase in the proportion of clones which identified polymorphisms when hybridisation was carried out for 48 rather than 20 hours (P< 0.05). There was no significant effect on proportion of clones identifying polymorphisms for the different driver:tester ratio.
The BamHI RDA products identified polymorphisms between the parents of the mapping population in 52 out of 120 clones investigated. Multiple banding patterns were seen for 48 of the 120 clones and in some cases the numbers of bands present may have prevented the detection of variants. This was particularly true for the TaqI derived clones, which made identification of variants difficult. TaqI RDA products identified polymorphisms between line N and line 15I parents for 17 of the 65 clones investigated. However 60% of all the clones hybridised to multiple bands in Southern blots. Some typical examples of the hybridisation patterns obtained from the BamHI derived clones are shown in Fig. 6.7. Some typical examples of the hybridisation patterns generated by the TaqI derived clones are shown in Fig. 6.8.
6.2.5 Mapping RDA Clones
The mapped clones generated from this experiment are listed in Table 6.2. A total of 20 clones from BamHI derived RDA products have been mapped so far and seven of these identify a total eight loci which lie on chromosome 16. Another seven clones identify a group of 16 closely linked loci on chromosome 4, a further clone is linked to chromosome 22 and another to chromosome 33, while four are unlinked at present. Two clones, hM98 and hM178 identify more than one locus on more than one linkage group. Several of the clones have only been screened on part of the mapping reference population, which will have contributed to the high proportion of unlinked loci. Fig. 6.9 shows the segregation of a clone hM37 in 56 progeny of the Compton mapping reference population, which is typical. This clone identifies a single locus, on chromosome 16 which has been designated COM154. However, one of the eight unique clones generated from the BamHI derived tester control which was allowed to self-anneal, clone hM152, 789also mapped to chromosome 16. This is an interesting if unexpected result, as this clone was derived without subtraction by driver DNA.
Five TaqI derived clones have been analysed so far of which three have been mapped and are located on chromosome 16.
Table 6.2. Locus number and linkage groups for BamHI (hM) and TaqI (hF) derived clones.

The current linkage map for chromosome 16 is shown in Fig. 6.10. The boxed markers are those placed on the chromosome by RDA, those in double boxes were generated from BamHI and TaqI targeted RDA clones. Fig. 6.11 shows the linkage map for chromosome 4 with the boxed markers generated by RDA. Again markers which are boxed with a double line were generated by chromosome 16 targeted BamHI derived RDA clones.
6.2.6 Nucleotide Sequence
Sequence analysis of the polymorphic clones from the BamHI derived RDA products identified several clones which contained repetitive sequences. Comparisons of DNA sequences with GenEMBL database were made using 1011GCG-FASTA software (Wisconsin package, version 8, September 1994, Genetics Computer Group, 575 Science Drive, Madison, Wisconsin, USA 53711). Homologies to known genes were not identified for any of the 89 clones sequenced. A number of clones (58%) contained a variety of repetitive sequences, an example of which is shown in Fig. 6.12.

5'-GGATCCAATGGGACTCAATGGGGCCCAATGGGATTCAATGGGACCCAATGGG
ACCAAATGGGACTCAATGGGATTCAATGGGGCCCAATGGGATTCAATGGGATT
CAATGGGGCCCAATGGGACTCAATGGGACTGAATGGGATTCAATGGGATT
CAATGGGATTCAATGGGATTCAATGGGAACCAATGGGAACCAATGGGACTGAA
TGGGAGTCAATGGGATTCAATGGGACCGAATGGGATTCAATGAGACTCAATGGG
AAACAGTGGGATTCAATGGGATTCAATGGGACCAATGGGATTCAATGAGACTC
AATGGGAACCAATGGGATTCAATGGGAGTCAATGGGACTCAATGGGGTTCAATG
GGAATCAATGGGACAATGGAACTCATGGGACCAATGGGACTCAATGGGATCC

Sequence data from clone hM110, using primer GEMF, showing all 421 nucleotides with 33 underlined heptameric repeats. Recognition sequences for BamHI are shown in italics.
The differences in the size of the polymorphic variants identified by these clones are too large to be explained by the variation in copy number of these repeats, which were also seen in non-polymorphic clones. However, the presence of these repeats may have affected the selection of these clones during the RDA process by affecting the kinetics of hybridisation. A number of these repeats were similar to minisatellite type repeats.
There was only a small proportion of redundancy in the RDA product libraries made. A total of six clones were represented between two and four times in the BamHI derived libraries and three clones were represented twice in the TaqI derived libraries. Duplicate clones are represented individually in App. D, but not in the results.
6.2.7 YAC Hybridisation
One clone (hM37) was used to screen the chicken YAC library, as it had identified a single locus (COM154) on chromosome 16. The gridded YAC colony blots are shown in Fig. 6.13 and should be interpreted as detailed in Chapter five (5.2.5). Clone hM37 hybridised to 10 YAC clones. Two of the YAC clones identified, also hybridised to MHC IIß probes (shown in Table 6.3) providing independent confirmation of the position of this locus on chromosome 16.
Table 6.3 Identification of Quad, 96-plate and AceDB clone numbers for YAC clones hybridised to clone hM37.

6.3 Discussion
In order to target RDA for the isolation of marker loci on the MHC chromosome the single tester was compared to 16 full siblings, which had inherited line 15I alleles for the MHC, NOR and Rfp-Y regions. This number of driver samples should ensure that over 99% of the tester genome is represented in the driver amplicons and thus could be removed by the subtractive selection within the RDA procedure.
The clones generated from the BamHI derived RDA products were cloned using a PCR product, TA-cloning vector: pGEMT (Promega). Five separate transformations were performed from a single kit, which produced 93% of clones containing two copies of the plasmid which was not expected. It is unlikely that the two plasmids exist independently as only one plasmid can enter a cell during transformation, so it is likely that the two plasmids have ligated to form one large double plasmid containing a single insert. This did not give problems in the isolation of DNA by miniprep methods, although the amount of insert was considerably reduced. However, in order to sequence the clones, it was necessary to gel purify the PCR-amplified insert. Thus all the sequence data that has been generated has come from samples which have been PCR amplified a total of four times. This could have generated Taq DNA polymerase errors, which are incorporated at an estimated frequency of one base substitution in 9000 nucleotides and one frameshift error every 41000 nucleotides (Tindall and Kunkel, 1988), but as most of the sequenced clones are less than 500 bp and were sequenced from both ends this is unlikely to have significantly affected the data.
Targeted RDA appears to have been successful, since 39% of the polymorphic clones identified by the BamHI RDA and 60% of the small sample of clones mapped for the TaqI RDA were located on chromosome 16. Since the target area of this chromosome is less than 2 Mb this suggests that the method can be applied successfully at the scale necessary to assist building YAC contigs for regions of interest identified through linkage mapping.
A comparison was made between hybridisations using amplicons generated by BamHI and TaqI restriction endonucleases. These two enzymes will give two different representations of the genomes in this targeted area and may enable a better coverage of the region with markers. It would be expected that clones derived from TaqI amplicons would be more likely to be polymorphic, as TaqI polymorphisms are particularly prevalent in the genome (Bumstead and Palyga, 1992). However 60% of the TaqI RDA clones detected multiple band patterns in Southern hybridisation analysis, making it difficult to distinguish whether those clones are polymorphic. This may explain why the relative number of polymorphic clones appears to be lower with TaqI derived clones than for BamHI.
Representation of the genome by the restriction endonuclease NheI was not successful and was abandoned. However, it might have been possible to improve the results by increased extension time during the amplicon preparation, enabling the larger fragments generated with this enzyme to be reproduced. Likewise an increase in the RDA selective amplification extension time as detailed by Lisitsyn et al. (1993) for the enzyme HindIII would also improve this reaction. Additional representations of the genome using different restriction endonucleases could potentially enhance the coverage of the targeted region.
Seven clones from this targeted RDA lie on chicken chromosome 4. This may be due to the inheritance by chance of line 15I alleles for number of marker loci on this chromosome by ten of the 16 driver birds. It should also be noted that 11 out of 15 of the RDA loci which segregate on this chromosome are in the inverted phase to the original data i.e. the original polymorphism appears in the line 15I allele not the line N allele of the F₁ parent. It would have been possible to discard these markers before mapping, however they have generated some potentially useful markers on chromosome 4.
A relatively high proportion of the RDA loci are currently unlinked as they have only been analysed in half the Compton mapping reference population. It is likely that these clones will show linkage once they have been typed on the whole Compton mapping reference population.
The proportion of variant clones was significantly higher after hybridisation over 48 hours than after 20 hours of hybridisation (chi-square P<0.05). This presumably reflects the time necessary for sequences at low copy numbers to reanneal. There appeared to be no difference between the products obtained for the range of DNA ratios used.
Many RDA clones detected multiple band patterns on hybridisation to Southern blots. This was not solely due to the presence of internal repetitive sequences within the clones, as only half such clones hybridised to multiple band patterns. The multiple band patterns presumably reflect replication of these regions within the genome.
To confirm the RDA hybridisation and subsequent PCR amplification during the experiments a sample of 0.4 µg tester with adaptors was used as a positive control throughout all the stages of the BamHI representation. Sample products were cloned and some of the clones generated were used to probe the line N and line 15I southern blots. This gave some unexpected results. It would be expected that these clones would represent a random selection of small fragments generated purely from the tester, it would not be expected that they would, in general, be polymorphic between the line N and line 15I birds as no subtractive selection had occurred. Three of these clones were duplicates of those generated by the full RDA technique with subtraction by driver amplicons. It is however surprising that one of these eleven clones (hM152) maps to chromosome 16. There seems to be no satisfactory explanation for this result, except that it could have occurred by chance, as any polymorphic clone has a 0.17% chance of mapping to this targeted region of chromosome 16.
These results show that in addition to identifying differences between two inbred lines of birds RDA can also be used to target particular regions of chicken chromosomes, in this case the MHC bearing microchromosome 16. Although not all the clones generated detected variants and only some of the clones targeted to chromosome 16 were located on that chromosome, there was significant and substantial enrichment for clones detecting variants in the targeted area. This shows RDA can be targeted to regions of less than 2 Mb and hence it should be possible to increase marker density in regions containing candidate gene

7.1 Introduction
With the success of RDA targeted to a specific chromosome detailed in chapter 6, it seemed that it should prove possible to target RDA towards the direct identification of a trait gene or genes. This would be most likely to be successful for traits caused by a single gene, or a small number of genes of large effect. Ideally with a clear phenotype and fully penetrant. RDA targeting of a single fully penetrant gene would then involve targeting a region of similar size (2 Mb) to that on chromosome 16 (targeted in Chapter 6). As a model for this RDA was applied to the identification of novel genes affecting resistance to Marek's disease.
MDV is a herpes virus, causing a lymphoproliferative disease in chickens leading to paralysis and tumours. The initial disease phase is acute and involves virus replication in B lymphocytes, following which the T lymphocytes become activated and infected. These T lymphocytes disperse around the body giving a persistent viraemia and in genetically susceptible birds some lymphocytes are transformed and proliferate as lymphomas in neural and visceral organs (Payne, 1996). MDV can be controlled by administration of the Rispens vaccine or Turkey herpes virus (HVT) a related herpes-virus of turkeys. However there are now a number of very virulent strains of MDV which are poorly controlled by vaccination and are increasingly causing high mortality, up to 80% in comparison to 10% associated with the classical from of MDV (Payne, 1996).
In order to investigate direct targeting of a trait gene an experimental population was needed which had been fully phenotyped for the trait and which had been used to generate a linkage map, enabling mapping of RDA generated loci. Populations of experimental birds present at IAH Compton were examined for these qualities. It was decided that the most clearly demonstrated segregation between resistant and susceptible birds by measurable pathogenesis was shown in the Marek's disease virus (MDV) reference population. This population is based on a cross between two inbred lines, the highly resistant line 6₁ and the highly susceptible line 7₂. There have been known to be at least two genetic components to MD resistance, one of which is within the MHC, however these two lines of birds share the same MHC haplotype B² and hence in this cross only non-MHC genes must be responsible for the difference in susceptibility. The MDV reference population includes a backcross (6₁x7₂)x7₂ as well as an F₂ cross, both of which have been fully typed for their resistance to Marek's disease (MD).
7.2 Experimental Design
In order to target MD resistance by RDA it was necessary to identify resistant and susceptible birds. Within experimental populations of birds MD infection was measured by four criteria, which were initially assessed on pure line 6₁ and line 7₂ birds. Resistant birds, after infection with MDV, showed low levels of cell associated viraemia, measured both by plaque assay and by quantitative PCR. For the plaque assay, infected cells were cultured on monolayers of chicken kidney cells and the number of plaques counted. In resistant birds this was in the region of 0.9 pfu per 10⁷ cells. MDV quantitative PCR amplifications were performed on samples of 10⁶ lymphoid cells (Bumstead et al., In Press) to determine levels of virus at three time points, 4, 11 and 39 days post infection (PI). The PCR values obtained were values of the amount of fluorescent PCR product generated. Resistant birds showed no detectable PCR product during the whole experiment. Also resistant birds failed to develop any neural or visceral tumours and did not die during the course of the experiment.
For comparison, the birds most susceptible to MD were also identified. The pfu count in these birds was high, greater than 3.2 pfu per 10 ⁷ cells. The MDV quantitative PCR values were also high: an initial (4 days PI) MDV PCR-A value of greater than 3658; a second (11 days PI) MDV PCR-B value of greater than 6136; and a third (39 days PI) MDV PCR-C value of greater than 12297. Susceptible birds died during the experiment and contained on post mortem either visceral or neural and visceral tumours.
Tester amplicons were generated from DNA from a genetically resistant bird and compared with driver amplicons generated from pooled DNA from genetically susceptible birds. Data from the MDV reference population was carefully studied in order to determine which birds were the most susceptible. Using the four criteria already detailed above 14 birds were chosen from the F₂ cross as those which were most susceptible to MD.
7.3 Results
7.3.1 RDA using BamHI representation
Representations of the driver and tester DNA were made as described in the methods (2.12.3), using RBAM oligonucleotides (70 and 68, see App. A) to generate the initial amplicons. Fig. 7.1 shows the BamHI amplicons generated from the 14 driver samples and the single tester sample. These PCR products range in size from 2.3 kb to less than 0.6 kb. After pooling equal amounts of the driver amplicons three rounds of hybridisation were performed, with the adaptor changes, similar to those described in the chromosome 16 targeted RDA using TaqI (2.12.2.2) but using the oligonucleotides used for the BamHI digested DNA i.e. new adaptors (JBAM oligonucleotides 66 and 69 see App. A) were re-ligated to the tester amplicons before the first hybridisation. NBAM adaptors (oligonucleotides 67 and 65 see App. A) were ligated to the first round hybridisation product and subsequently JBAM adaptors were ligated to the second hybridisation product. Fig 7.2 shows the RDA products after the first, second and third 1&2rounds of hybridisation respectively. Interestingly after the second round of hybridisation a small number of discrete bands can be seen. Some of the products from each round were sampled and cloned into the pGEMT vector, as described in the methods (2.2.3).
7.3.2 Screening and Characterisation of RDA clones
The inserts present in a sample of the clones were amplified by colony PCR and a sample of clones used to probe Southern blots of BamHI digested genomic DNA from a line 6₁ and a line 7₂ bird. The restriction fragment polymorphisms identified by ten of these clones are shown in Fig. 7.3. These clones show a variety of hybridisation patterns indicating the presence of one or two copies of some clones (for example clone hV11) and multiple copies of other clones (for example clone hV29). As in previous RDA comparisons a substantial proportion (32%) of the clones detected clearly visible polymorphisms between the two lines (see Table 7.1)
Table 7.1 Characterisation of RDA clones targeted to resistance to Marek's disease.

7.3.3 Mapping RDA Clones
A total of eight clones have been used to probe Southern blots of the MDV mapping panel. A typical mapping blot is shown in Fig. 7.4 (clone hV32) segregating in 47 of the backcross population. To date seven of these have identified linkage groups within the incomplete genetic map developed in this cross. The eighth locus (hV29) is linked to a microsatellite marker, but this group of two loci has not been assigned a linkage group, so remaining unlinked. The seven linked loci lie on the linkage groups corresponding to chromosomes 1, 16 and W. Four of these clones identify loci on chromosome 1, three of which are closely linked (hV32, hV2 and hV11). A diagram of the 34chromosomes is shown in Fig. 7.5, with the RDA clones identified in boxes. The segregation data for all the RDA clones is tabulated in App. E.
7.3.4 Analysis of association of RDA clones to MD resistance.
The association of each locus with mortality was assessed by chi-square analysis, the results of which are shown in Table 7.2.
Table 7.2 Chi-square analysis for targeted RDA clones.

The table shows the contingency chi-square analysis for association of the line 6₁ allele at each locus with survival. The calculated probabilities for clones hV32, hV11 and hV20 are all significant at P<0.010, with the association of clone hV32 and the "death" pseudo-locus being very highly significant at P<0.005. However, since there is a firm prediction that the line 6 allele would be associated with resistance, as is the case for these loci, it would be reasonable to treat these probabilities as for a one tailed distribution i.e. a further halving of the probabilities. Each of the three closely linked loci on chromosome 1 shows a significant association with survival in the predicted phase when mortality was included in the linkage analysis as a pseudo-locus (death). The proposed position for this pseudo-locus was linked to the RDA loci with LOD=2.1.
A more refined analysis of the association with death of the RDA clones mapped to chromosome 1 was performed using the quantitative analysis 5function of Map Manager QT version [beta]8. The interval mapping is derived from the work of Haley and Knott (1992), Martinez and Curnow (1992) and Zeng (1993). The QT [beta]8 program determines the regression coefficient for the effect of the QTL between pairs of flanking positions on a specific chromosome. As a measure of significance the program calculates the distribution of values for all other positions and from this distribution identifies significance levels. In this case survival was the quantitative trait, and multiple regression analyses were determined at 1 cM intervals on chromosome 1 with no control for other QTLs. Since the "death" is a categorical trait, this approach is approximate but provides the best indication of the likely position of the QTL. Using the interval mapping with permutation test (Churchill and Doerge, 1994), data for 500 permutations at 2 cM intervals was obtained for likelihood ratio statistics values of suggestive linkage (GL), significant linkage (SL) and highly significant linkage (HSL) as proposed by Lander and Kruglyak (1995). Chromosome 1 was analysed for 500 permutations, in 2 cM bins for association with death, giving values of GL=7.8, SL=15.2, HSL=22.9. These values were then used as thresholds for interval mapping of chromosome 1 to show the associations for each locus with death by regression coefficients. Interval mapping of chromosome 1 is shown diagramatically in Fig. 7.6, again with regions of suggestive linkage and significant linkage (Lander and Kruglyak, 1995). This showed that there was suggestive to significant linkage between hV32, hV11 and death on chromosome 1. Data from interval mapping is shown in Table 7.3 for probabilities and additive effects, at a confidence limit of P=0.05.
Interval mapping with permutation analysis was also applied to the other RDA clones to test for their association with survival. However, the strongest other association found was with clone hV5, giving a value of P=0.0112, with a positive additive effect of +0.33 i.e. in the unexpected phase. Interval mapping for this and other non-chromosome 1 loci is imperfect since there are not multiple loci on a well defined linkage map for these regions.6
Table 7.3 Probability of association with survival for loci mapped on chromosome 1.

The above table shows the loci in their order on the chromosome with the probability of association derived from permutation analysis with no allowance for other QTLs. If all the probabilities are halved, then hV11, hV32 and microsatellite marker ADL 358 all show very strong associations with death of P=0.00058, P=0.00019, P=0.00077 respectively. The additive effect of each of these loci is greater than -0.44 which means that almost half the variation in mortality is caused by this region. Loci hV20 and MCW145 also gave significant probabilities, after correction, to one tailed, of P=0.0069 and P=0.0055 respectively.
Data for the association of each set of quantitative MDV PCR values with the loci on chromosome 1 was also analysed by interval mapping with permutation analysis. The first PCR values were obtained 4 days PI and gave permutation threshold values of GL=8.2, SL=13.1, HSL=17.9, for chromosome 1. Interval mapping of the loci on chromosome 1 in relation to the PCR-A values, is shown in Fig. 7.7. Loci, hV32 and hV11 have low 7probabilities (P=0.05279 and P=0.0391 respectively) of association with PCR-A values and are well below suggestive linkage. The second PCR values were obtained 11 days PI and gave permutation threshold values of GL=9.6, SL=15.2, HSL=23.8, for chromosome 1, interval mapping is shown in Fig. 7.8. The likelihood ratio statistics of "death" and ADL358 show values between suggestive and significant linkage to PCR-B. These strong associations with PCR-B values are also shown by the two-tailed probability values from P=0.0228 to P=0.00118 for the loci: Death, ADL358, ADL362, MCW109, MCW61, ADL268, CHB6, and MCW68. The third PCR values were obtained 39 days PI and gave permutation threshold values of GL=9.2, SL=15.4, HSL=25.6, for chromosome 1, interval mapping is shown in Fig. 7.9. Again suggestive to significant linkage was shown by "death", as well as loci hV11, hV32 and ADL358 to PCR-C values. These strong associations with PCR-C values are also shown by the two-tailed probabilities from P=0.02576 to P=0.00021 for the loci: hV20, hV11, hV32, death, ADL358, ADL362 and MCW61.
These results show that loci hV32, hV11 and hV20 are strongly associated with amount of MDV PCR product as well as with survival. Viral PCR values and subsequent mortality are strongly correlated, however, if the RDA loci only affected mortality there would not be any relationship between the MDV PCR values and the loci. However, since these loci show association with both PCR value and survival, it seems likely that the gene located in this interval effects mortality through a reduction in viral levels by a possible reduction in proliferation.
7.3.5 Segregation of hV32 in the F₂ population
Clone hV32 was also used to probe a blot of BamHI digested F₂ progeny. The segregation of clone hV32 was analysed by chi-square, giving a one-tailed value of 3.16 (P>0.05), showing that there was no significant association with death. Interestingly only five of the 14 selected driver birds had inherited line 7₂ alleles for locus hV32. Details of the segregation can be found in App. E.
89
7.4 Discussion
RDA was targeted to the non-MHC loci responsible for resistance for Marek's disease (MD) which are present in the resistant inbred line 6₁ chickens, but absent in the susceptible inbred line 7₂ chickens. These two lines differ in their susceptibility to MD by the inheritance of genes which are outside the MHC region, as they share the same MHC haplotype. Tester DNA was taken from a pure line 6₁ bird and compared with driver DNA from 14 susceptible F₂ birds. Thus the DNA in the pool of 14 F₂ birds will be similar to that of the line 6₁ bird for any random locus, except in regions of MD resistance where these birds would have inherited only line 7₂ alleles.
The statistical analyses performed on the eight targeted RDA loci to determine their association with survival apply a greater level of stringency than is normally used in conventional mapping. Although these clones have been placed on the linkage map they have been targeted to MD resistance and do not represent a saturation coverage of the genome in the manner of conventional mapping loci. The fact that one of these clones (hV32) showed a very strong association with survival and that two other clones (hV11 and hV2) were closely linked to this clone suggests that the MD resistance targeting was successful.
Since the RDA loci on chromosome 1 show association with both mortality and quantitative PCR product values, this implies that the influence of this locus is due to an effect on viral replication or proliferation, rather than directly on the suppression of developing tumours. It would appear that this locus is affecting viraemia directly, presumably by restricting viral replication specifically or by increased immune response, thus decreasing viraemia. It has previously been shown that there is a strong correlation between reduced viral levels, as measured by PCR, and subsequent mortality (Bumstead et al., In Press). Thus with decreased viraemia there will be less transformed lymphocytes and therefore a decrease in damage to the immune system leading to reduction of the numbers of tumours, probably due to the relative increase in capacity by immune response.
It would be interesting to see if the clones hV32 and hV11 show a similar association to death from MD in other genetic crosses of inbred lines of chickens. This could clarify the association of these loci with MD survival, although it is unlikely that these particular loci would be polymorphic in another cross.
There are other polymorphic RDA clones which have not yet been applied to the MDV mapping panel, which should provide an increase in the numbers of loci linked to MD resistance, and further refine the position of the gene. In order to isolate candidate genes for this region of interest greater numbers of loci which are closely linked to death will be required in this area, so that physical mapping by YAC contig analysis can be performed. With the generation of these clones a region has been identified and it is now necessary to continue characterising the clones already generated.
Many microsatellites are not polymorphic in this cross (about 75%) and they are not evenly distributed, so although around 350 have been investigated in total only one at present lies in this region, despite particular efforts to map microsatellites in this region. This microsatellite (ADL358) shows a strong association to mortality. However, there are no other microsatellites available for this area of chromosome 1 which are polymorphic in this cross. Therefore, in order to refine the position of the gene for MD resistance it will be necessary to use other methods, particularly targeted RDA to produce more markers.
Here RDA has been shown to be useful, both in targeting a specific chromosome and in targeting regions containing specific genes. In the latter case RDA has identified a region containing a gene, or possibly a few very closely linked genes. The application of RDA was helped by conventional mapping data in order to place the targeted markers within a specific linkage group and to relate their position to other markers. However, RDA decreased the need for very large numbers of markers on the map and thus appears a much more efficient method of generating markers linked to the putative gene. Given the success obtained in this fairly complex situation RDA could also be applied to much more complex situations where larger numbers of genes influence the trait of interest.
Thus RDA provides a very useful tool for relatively direct identification of markers linked to genes of interest. This will be very helpful in reducing the numbers of markers needed when new chicken crosses are made, as other loci will only be needed to define the initial linkage groups in order to relate them to the reference map. RDA can then produce targeted markers enabling refined mapping only of the region of interest.

8.1 The Chicken
The chicken is a good animal for genetic study because of their genome size and structure, the ease of DNA preparation from nucleated red blood cells, the large numbers of available inbred lines, the segregation of disease and production traits between lines, the breeding potential of large families of birds and the ease with which genetic markers could be used in industry. With the advent of the chicken linkage map in 1992 (Bumstead and Palyga), the chicken genome was open to more complete investigation. Linkage maps have now been developed for two other populations (Levin et al., 1994b; Crooijmans et al., 1996) and all three are being integrated to provide the best possible coverage of the genome. A recent advance includes the development of a chicken yeast artificial chromosome (YAC) library (Toye et al., In Press) containing more than 10 copies of the genome. Currently a chicken whole genome radiation hybrid library (WG-RH) is being developed at the IAH Compton, with the expectation that this will provide a resource for mapping ESTs, cloned genes and YACs. Alongside the biological resources, chicken bio-informatics has also expanded, with the introduction of the public database, ChickGBASE which is accessible via the world wide web as ChickMap (http://www.ri.bbsrc.ac.uk) or as US. Poultry Gene Map (http://poultry.mph.msu.edu/).
However, although resources for chicken mapping are increasing there are still a number of problems which have to be addressed. The chicken linkage map is not densely covered with markers, and very few genes have currently been mapped. There is still no simple way of mapping non-polymorphic clones, although the development of WG-RH should assist with this. Also it is still difficult to correlate the linkage groups with the actual chromosomes, this particularly applies to the MICs.
8.2 Genome Mapping
Mapping the chromosomal DNA, aimed at the isolation of trait genes, of any species, is a time consuming and laborious process,. It has only been relatively recently that technology has enabled geneticists to move from "functional" mapping, in which the protein responsible for a disorder is identified first before working back to the gene, to "positional" mapping (Collins, 1992), in which the position of the gene responsible for the disorder is identified in the genomic DNA, subsequently allowing its protein product and finally its function to be determined. Functional mapping has been successful in the isolation of a number of genes, including those causing Phenylketonuria (Kwok et al., 1985) and G6PD deficiency (Persico et al., 1986). However, functional mapping is highly dependent on the nature of the effect of the gene. In contrast positional mapping is almost independent of the gene function and has increasingly been used to locate and isolate trait genes, of which perhaps the first, was the gene for Chronic granulomatous disease (Royer-Pokora et al., 1986) and one of the latest the gene for Hermansky-Pudlak syndrome, a disorder of cytoplasmic organelles (Oh et al., 1996).
The process of identifying a trait gene by positional cloning begins with a genetic map which has a number of loci, one of which is closely linked to the trait of interest. The region surrounding the linked locus must have a high enough density of loci to enable the alignment of YACs. The YAC contig is a crucial step towards physical mapping and gene identification as once this is achieved genes within the YACs can be analysed. There is no perfect method for the identification of genes within YAC clones, so a variety of techniques are employed to increase the chance of isolating the trait gene. Approaches to gene identification include: exon trapping (Buckler et al., 1991), CpG island mapping (Bird, 1986), open reading frame cloning (Weinstock et al., 1983), direct selection of cDNAs (Lovett et al., 1991), STS mapping and sequence scanning reviewed by Ward and Davies (1993). Candidate genes are then isolated by sub-cloning and further analysis usually resolves which is the gene of interest. However, none of these procedures can be carried out until there are enough markers on the linkage map, such that at least one will be linked to the trait and that there are at least two flanking markers to facilitate the isolation of YACs and their alignment.
8.3 Saturation Mapping
One manner of generating enough markers for the alignment of YACs and the identification of candidate gene regions is simply to add very large numbers of marker loci at random to the genetic map. This has been the approach of geneticists in the absence of a reliable method of targeting. However, ideally saturation of the genome with markers needs to be achieved to a density on the scale of the length of the average YAC clone, which in the chicken is approximately 0.63 Mb (Toye et al., In Press), since this will largely remove the need for difficult and expensive chromosomal walking from YAC to YAC. In the chicken a genetic distance of 1 cM is approximately equivalent to 0.5 Mb of DNA (Levin et al., 1994b), so markers will be needed one per cM to allow direct identification of YACs containing trait genes.
8.4 Random amplified polymorphic DNA (RAPD) analysis
Random amplified polymorphic DNA (RAPD) analysis has been investigated in this work as one method of increasing marker density on the chicken linkage map. The advantage of this technique is its uninformed approach to the genome i.e. no previous sequence data is required. RAPD analysis has provided large numbers of markers for the genetic maps of many organisms including some in the chicken (Levin et al., 1993).
However, chicken DNA does not appear to give many polymorphisms either in the present study or that of Levin et al. (1994a) where out of a total of ten primers, one primer failed to generate a product under a variety of conditions, whilst three failed to show any detectable polymorphisms when used singly.
It is possible to target RAPD analysis by determining which primers give polymorphisms between pooled samples by bulked segregant analysis. This has been successfully achieved for RAPD markers on chromosome 11 in tomatoes (Giovannoni et al., 1991) and markers near to the downy mildew resistance gene in lettuce (Michelmore et al., 1991). However, this can only be achieved if it is possible to identify large numbers of polymorphisms between the samples.
The number of RAPD polymorphisms generated in this work by a single primer was low, so it was decided not to proceed with this technique. The lack of polymorphisms also ruled out the possibility of targeted RAPD markers. This method of marker generation does not provide the large numbers of markers needed for saturation of the genome in order to isolate trait genes.
8.5 Microsatellite Markers
The most frequently used technique for randomly placing markers on a linkage map is microsatellite analysis (Weber, 1990). This is the technique of preference for mapping the human genome. However, microsatellite analysis requires knowledge of the DNA sequence surrounding the di-, tri-, or tetra- nucleotide repeats. Thus it is time consuming and costly in initial outlay, but once primers have been synthesised the cost of mapping is much reduced. This system works well in man as there are many microsatellites distributed throughout the genome. However, the chicken genome has fewer microsatellites estimates range to as low as 7,000, almost 10 fold less than in mammals and they appear not to be uniformly distributed throughout the genome (Toye, 1993). Also in chickens primers which work in one cross have less than 50% chance of identifying polymorphisms in another cross (Khatib et al., 1993). Within the Compton mapping reference population microsatellites appear to be even less polymorphic (Bumstead pers. com.). This is not a very productive method of random mapping, but it does give anchor loci which can be used between different chicken crosses to identify linkage groups.
8.6 Amplified fragment length polymorphisms (AFLP)
The relatively new technique of amplified fragment length polymorphism (AFLP) (Vos et al., 1995) is now being used within the Compton mapping reference population to generate markers to saturate the map. Like RAPD, AFLP does not require any previous sequence knowledge, however AFLP are more stable than RAPD and give much better uniformity of resolution. Even more significantly each reaction identifies five to ten polymorphisms between pairs of inbred chicken lines, thus resulting in large numbers of mapped loci. These polymorphic markers are currently being used at IAH Compton in the development of linkage maps of resource crosses. AFLP seem to be the ideal method of saturation mapping, as they require no prior knowledge of the genome and produce large numbers of polymorphisms cheaply and efficiently. The anonymous nature of AFLP markers means that they are difficult to relate between crosses and hence to candidate genes mapped in the reference population. Also there are some technical difficulties entailed in the isolation of individual AFLP markers as probes to identify YACs.
Within the mammalian species; mice and humans, saturation mapping has achieved its goals of a high density of markers over the whole genome. However, the chicken linkage maps have far fewer markers and a targeted approach to mapping would have a greater chance of finding candidate genes.
8.6 EST mapping
An initial approach to mapping, by targeting genes expressed in specific tissues was the use of expressed sequence tags (EST). These are generated by sequencing clones from a cDNA library and comparing the sequence to DNA databases in order to find homologies with known genes. Targeting is limited as it can only be achieved in the selection of a specific tissue from which the cDNA is generated and requires knowledge of protein expression patterns for the trait studied. In the present study seven ESTs were successfully placed on the chicken linkage map. These were derived from a bursal cDNA library which could potentially have contained genes involved disease resistance.
ESTs can also be used as markers for comparative mapping, since if synteny is conserved between two species, it should be possible to infer the presence of genes lying within the conserved region for the species of interest. Thus it should be possible to move from one species map to another and back again in order to compare the position of genes located within certain chromosomes, and so this may take advantage of the detailed gene maps of humans and mice. This comparison of synteny between species can only be studied for ESTs, which is why they are such useful markers. An example of synteny between chickens, man and mouse is in the region of the gene controlling Salmonella resistance in mice, Nramp1 (Vidal et al., 1993). Nramp1 is linked to the FN1 and VIL genes in all three species. Synteny of the gene order FN1, Nramp1 and VIL is conserved on mouse chromosome 1, human chromosome 2q and chicken chromosome 7 (Girard-Santosuosso et al., In Press).
At present the principal drawback of the EST approach in chickens is the difficulty of mapping the sequenced genes, it may be possible to overcome this problem in the future by using WG-RH.
8.7 Subtractive Hybridisation
In the past other approaches to targeted mapping have concentrated on subtractive methodologies. The most common of these is subtractive hybridisation, where two types of tissue are compared by subtracting the cDNAs in one tissue from the cDNAs in the tissue of interest. This is a positive selection technique as the cDNAs of interest are subsequently radioactively labeled and used to screen a cDNA library, from which clones are isolated. These clones will only be expressed in the tissue of interest (Lee et al., 1991). However, this method is totally dependent on knowledge of the expression pattern for the gene of interest, and so is not always applicable, it is also technically extremely demanding.
8.8 Differential display-reverse transcription (DD-RT)
Another method for the analysis of changes in gene expression in cells at different stages of differentiation is Differential display-reverse transcription (DD-RT) (Bauer et al., 1993). DD-RT isolates potentially useful genes expressed at a particular stage, by amplification of mRNA from two or more cell types. These are compared between the tissue types and the differences isolated. Once again however, this method is dependent on knowledge of the expression of the gene as it is targeted at differences between the mRNAs for the selected tissues. DD-RT is also technically demanding and laborious as large numbers of products need to be compared. A new method also based around isolating differences between cDNA samples, cDNA RDA has proved efficacious and may offer a less laborious alternative (Hubank and Schatz, 1994).
8.9 RFLP subtraction
Another method of subtraction utilises the whole genome, based on known genotype or phenotype differences between two genomes. RFLP subtraction purifies small restriction fragments from one genome (the tracer) away from sequences which reside on fragments in a related genome (the driver) (Rosenberg et al., 1994). In order to purify the unique tester fragments the driver DNA is labeled by ligation to biotinylated primers. Thus only tester:tester hybridised fragments will remain in solution, to be cloned. However, again there is no positive selection for the tester samples and hence the method is technically difficult. There is also no enrichment of the tester samples and the complexity of the total genomic DNA will reduce the rehybridisation of low copy sequences.
None of these methods give a systematic approach to targeted mapping. Most are based on knowledge of the expression of the gene either in a particular tissue or at a specific time during development. In 1993 Lisitsyn et al. developed a method based on these subtractive methodologies which did not require knowledge of expression patterns within tissues. This method is Representational difference analysis (RDA)
8.10 The RDA Approach
Representational difference analysis is a targeted approach to mapping based on information from either the genotype or phenotype of a pedigree family. RDA has many advantages over previous methods particularly as the target DNA is positively selected and subsequently enriched during the reaction. This work has shown that RDA can equally be applied to the chicken genome generating large numbers of clones, of which 50% are polymorphic.
RDA can be used in a simple comparison of two genomes (line N vs. line 15I), or can be targeted towards specific areas. Lisitsyn et al. first performed Genetically Directed Representational Difference Analysis (GDRDA) in 1994 to target three polymorphisms to within less than 1 cM of the mouse nude locus on chromosome 1. This work has used genotype information from the original loci on the chromosome 16 linkage group to target part of that chromosome. This has proved very successful placing, ten new loci in a region that is less than 0.17% of the entire genome.
Data generated from the phenotype can also be successfully used to target areas of the genome. This work has targeted regions of the chicken genome containing genes conferring resistance to Marek's disease from analysing the phenotypes of the F₂ birds subjected to infection with Marek's disease virus. This succeeded in targeting regions of the genome associated with death from MD, particularly one area on chromosome 1.
Although RDA is a very powerful technique which does not require knowledge of tissue specific gene expression it does have some disadvantages. The greatest problem involves finding DNA species which are suitably matched for the analysis. The quality of the samples is also of concern as contamination by chemicals and RNA will cause problems with subsequent enzyme reactions. Any DNA contamination present in the tester samples will be amplified and preferentially selected for, which could lead to erroneous results. Tester and driver samples should be treated separately to avoid any cross contamination which could also significantly affect the procedure. There are many technical steps involved with RDA and if any one of these fails to work correctly the numbers of positive results will be greatly reduced.
8.11 The Future of Mapping
Many strategies for mapping genes have been developed since the advent of recombinant DNA technology. However, most of these have used a random or saturation mapping approach to positional cloning. Random mapping is a useful tool for generating loci on a new genetic map but it is consuming of both time and resources. This approach also does not necessarily include the use of anchor loci which can be used to relate linkage groups in a variety of crosses as well as relating chromosome synteny between species. A new approach is needed for targeted mapping and Representational difference analysis may fill this role. Specific genes can be targeted with only minimal information from their phenotype to determine a selection of driver and tester samples. As with the MD targeted RDA described here, this may be able to directly identify a region containing the trait of interest.
RDA may not remove the need for random or saturation mapping, but works alongside conventional methods of microsatellite analysis and AFLP analysis to provide genetic linkage to traits of interest. This culminates in an increased marker density in a region of interest, preferably one every centi-Morgan. Thus enabling physical mapping with the alignment of YAC clones and ultimately the isolation and characterisation of the trait gene.
RDA will be an important future resource, as with each new cross generated a new linkage map will be needed and RDA will provide the necessary markers targeted to the trait of interest. It will be difficult to place large numbers of markers on each new linkage map but by using RDA, specific genes can be targeted and only those areas saturated with markers. This should decrease the time taken to map new genes, as well as decreasing the cost. However, RDA should not be used in isolation, but is of most benefit when combined with traditional methods of generating markers, especially anchor loci. Anchor loci will establish the identity of the linkage group concerned, enabling comparisons between the resource linkage map and the reference map. However, it should also be possible to move indirectly from radiation hybrid panels or in situ hybridisation positions to the reference map. This would lead to further clarification of the area by markers already present on the reference map, thus enabling a higher density of markers for the identification of YACs.
8.12 Future Work
Targeted mapping by RDA has succeeded in generating 16 extra loci on chromosome 16. Two of these clones have hybridised to 23 chicken YAC clones some of which have previously been identified by MHC probes. These YACs have been aligned into a contig of chromosome 16 using AceDB. With the alignment of YAC clones the process of characterising them and establishing a physical map of the chromosome can proceed. Ultimately all the genes within the YAC contig will be identified and eventually characterised. This will lead to a greater understanding of the chicken MHC region, its evolution and the significance of the Rfp-Y region (Briles et al., 1993).
The physical mapping of chromosome 16 could also enable the genomic organisation and structure for a microchromosome to be elucidated and compared with that of a macrochromosome. This work could lead to an increased understanding of the evolution of microchromosomes which are a conserved feature of the avian genome, but absent in mammalian genomes. If further markers are required for YAC contig building there are another 14 mapped RDA markers which can be used. There are also another 27 polymorphic BamHI derived RDA clones targeted to chromosome 16 which have yet to be mapped. Only five TaqI derived RDA clones have been mapped to date, but there are a further 13 polymorphic clones to be mapped and 174 clones still to be characterised.
Targeted mapping of MD resistance by RDA was successful in identifying four loci associated with resistance, one of which was closely linked on chromosome 1. In order to increase the marker density in this region of interest there are at least another 16 cloned polymorphic RDA markers which should be mapped on this population. These are likely to generate further loci on chromosome 1, some of which could be more closely linked directly to the trait. There are also a further 30 RDA clones which have yet to be characterised, and should produce at least another 12 polymorphic clones, as approximately 40% are polymorphic. Thus with the application of another 28 targeted markers, it would be expected that this region on chromosome 1 would become saturated with loci. When all the products from this experiment have been exploited it would be possible to perform other RDA experiments to generate more targeted markers.
It would be possible to perform many other RDA experiments to refine the targeting of MD resistance. For example using the same tester and driver samples, but changing the enzyme used, TaqI has already proved useful in the chicken genome.
Investigations of other chicken inbred lines which show resistance to MD will enable the function of the resistance gene found in this study to be assessed finally confirming the association of the RDA locus with a resistance gene. With the generation of more crosses, further RDA experiments can be targeted to the identification of other resistance genes, after first excluding the present one.
The success and technical robustness of RDA demonstrated here shows that it could also be used to target other disease resistance genes currently under investigation at IAH Compton. For example there are crosses between inbred lines of chickens which differ in their resistance to salmonellosis and a cross between two inbred lines which differ in their resistance to IBDV is in preparation. All these resource populations will require initial linkage maps generated by conventional means, using microsatellite markers to define links to the reference map. These populations will test the ability of targeted RDA to generate loci linked to the trait of interest, as they could be due to more complex situations than for MD resistance. Thus the populations should be examined closely to identify birds which show good segregation for resistance to the disease in question. This will enable targeted RDA, by using DNA from resistant tester birds and comparing it to DNA from susceptible driver birds.

Abbott, U.K. and Yee, G.W.: Avian Genetics. In: King, R.C. (Ed.), Handbook of Genetics. Plenum, New York, (1975) pp. 151-200.
Aburatani, H., Stanton, V.P. and Housman, D.E. (1996) High-resolution physical mapping by combined Alu-hybridisation/PCR screening: Construction of a yeast artificial chromosome map covering 31 centimorgans in 3p21-p14. Proceedings of the National Academic Society 93, 4474-4479.
Adams, M.D., Kelley, J.M., Gocayne, J.D., Dubnick, M., Polymeropoulos, M.H., Xiao, H., Merril, C.R., Wu, A., Olde, B., Moreno, R.F., Kerlavage, A.R., McCombie, W.R. and Venter, J.C. (1991) Complementary DNA sequencing: Expressed sequence tags and human genome project. Science 252, 1651-1656.
Adra, C.N., Kobayashi, H., Rowley, J.D. and Lim, B. (1994) Assignment of the human GDID4 gene , a GDP/GTP exchange regulator, to chromosome 12p12.3. Genomics24, 188-190.
Arthur, R.R. and Straus, N.A. (1983) DNA sequence organisation and transcription of the chicken genome. Biochimica et Biophysica Acta741, 171-179.
Arumuganathan, K., Martin, G.B., Telenius, H., Tanksley, S.D. and Earle, E.D. (1994) Chromosome 2-specific DNA clones from flow sorted chromosomes of tomato. Mol. Gen. Genet 242, 551-558.
Avner, P., Amar, L., Dandolo, L. and Guénet, J.L. (1988) Genetic analysis of the mouse using interspecific crosses. T.I.G. 4, 18-23.
Bachev, N. and Lalev, M. (1990) A study of the relationship between some corporal dimensions and live weight of turkeys. Zhivotnov'Dni-Nauki27, 31-34.
Bacon, L.D. (1987) Influence of the major histocompatibility complex on disease resistance and productivity. Poultry Science 66, 802-811.
Bailey, E. and Lear, T. (1994) Comparison of thoroughbred and Arabian horses using RAPD markers. Animal Genetics 25, 105-108.
Baldocchi, R.A., Tartaglia, K.E., Bryda, E.C. and Flaherty, L. (1996) Recovery of probes linked to the jcpk locus on mouse chromosome 10 by the use of an improved representational difference analysis technique. Genomics33, 193-198.
Barral, V., This, P., Imbert-Establet, D., Combes, C. and Delseny, M. (1993) Genetic variability and evolution of the Schistosoma genome analysed by using random amplified polymorphic DNA markers. Molecular and Biochemical Parasitology 59, 211-222.
Bateson, W. and Saunders, E. (1902) Experimental Studies in the physiology of heredity. Rep. Evol. Com. R. Soc. Lond. 1, 87-124.
Bauer, D., Müller, H., Reich, J., Riedel, H., Ahrenkiel, V., Warthoe, P. and Strauss, M. (1993) Identification of differentially expressed mRNA species by an improved display technique (DDRT-PCR). N.A.R. 21, 4272-4280.
Bernot, A., Zoorob, R. and Auffray, C. (1994) Linkage of a new member of the lectin supergene family to chicken Mhc genes. Immunogenetics39, 221-229.
Bird, A. (1986) CpG-rich islands and the function of DNA methylation. Nature321, 209-213.
Bloom, S. and Bacon, L. (1985) Linkage of the major histocompatibility (B) complex and the nucleolar organiser in the chicken. The Journal of Heredity 76, 146-154.
Bloom, S., Delany, M. and Muscarella, D. (1993): Constant and variable features of avian chromosomes. In: Etches, R. and Verrinder Gibbons, A. (Eds.), Manipulation of the Avian Genome. CRC Press, Guelph, pp. 39-59.
Boise, L.H., Gonzaléz-Garcia, M., Postema, C.E., Ding, L., Lindsten, T., Turka, L.A., Mao, X., Nunez, G. and Thompson, C.B. (1993) bcl-x, a bcl-2- related gene that functions as a dominant regulator of apoptotic cell death. Cell 74, 597-608.
Brazel, D., Oberbaumer, I., Dieringer, H., Babel, W., Glanville, R.W., Deutzman, R. and Kuhn, K. (1987) Completion of the amino acid sequence of the [alpha] 1 chain of human basement membrane collagen (type IV) reveals 21 non-triplet interruptions located within the collagenous domain. European Journal Of Biochemistry 168, 529-536.
Briles, W., Stone, H. and Cole, R. (1977) Marek's disease: Effects of B histocompatibility alloalleles in resistant and susceptible chicken lines. Science 195, 193-195.
Briles, W.E., Goto, R.M., Auffray, C. and Miller, M.M. (1993) A polymorphic system related to but genetically independent of the chicken major histocompatibility complex. Immunogenetics 37, 408-414.
Bruford, M.W., Hanotte, O. and Burke, T. (1994) Minisatellite DNA markers in the chicken genome. II. Isolation and characterisation of minisatellite loci. Animal Genetics 25, 391-399.
Buckler, A.J., Chang, D.D., Graw, S.L., Brook, J.D., Haber, D.A., Sharp, P.A. and Housman, D.E. (1991) Exon amplification: A strategy to isolate mammalian genes based on RNA splicing. Proc. Natl. Acad. Sci. USA.88, 4005-4009.
Bumstead, N., Millard, B.M., Barrow, P.A. and Cook, J.K.A. (1991): Genetic basis of disease resistance in chickens. In: Owan, J.B. and Axford, R.F.E. (Eds.), Breeding for disease resistance in farm animals. CAB International, Wallingford, England, pp. 10-23.
Bumstead, N. and Palyga, J. (1992) A preliminary linkage map of the chicken genome. Genomics 13, 690-697.
Bumstead, N., Reece, R.L. and Cook, J.K.A. (1993) Genetic differences in susceptibility of chicken lines to infection with infectious bursal disease virus. Poultry Science 72, 403-410.
Bumstead, N., Wain, H., Salmon, N. and Sillibourne, J. (1994a): Genomic mapping of immunological genes. In: Davison, T., Bumstead, N. and Kaiser, P. (Eds.), Advances in Avian Immunology Research. Carfax, pp. 75-.
Bumstead, N., Young, J.R., Tregaskes, C., Palyga, J. and Dunn, P.P.J. (1994b) Linkage mapping and partial sequencing of 10 cDNA loci in the chicken. Animal Genetics 25, 337-341.
Bumstead, N., Sillibourne, J., Rennie, M., Ross, N. and Davison, F. (In Press) Quantification of Marek's disease virus in avian lymphocytes using the polymerase chain reaction with fluorescence detection.
Burch, J.B.E., Davis, D.L. and Haas, N.B. (1993) Chicken repeat 1 elements contain a pol-like open reading frame and belong to the non-long terminal repeat class of retrotransposons. Proc. Natl. Acad. Sci. USA.90, 8199-8203.
Burke, D.T., Carle, G.F. and Olson, M.V. (1987) Cloning of large segments of exogenous DNA into yeast by means of artificial chromosome vectors. Science236, 806-812.
Burt, D.W., Bumstead, N., Bitgood, J.J., Ponce de Leon, F.A. and Crittenden, L.B. (1995) Chicken genome mapping: A new era in avian genetics. T.I.G.11, 190-194.
Carefoot, W.C. (1987) Test for linkage between the eumelanin restrictor (Db) and the eumelanin extension (Ml) genes in the domestic fowl. Br. Poultry Sci. 28, 69-73.
Challoner, P., Smith, K., Parker, J., Macleod, D., Coulter, S., Rose, T., Schultz, E., Bennett, J., Garber, R., Chang, M., Schad, P., Stewart, P., Nowinski, R., Brown, J. and Burmer, G. (1995) Plaque-associated expression of human herpesvirus 6 in multiple sclerosis. PNAS 92, 7440-7444.
Chang, Y., Cesarman, E., Pessin, M.S., Lee, F., Culpepper, J., Knowles, D.M. and Moore, P.S. (1994) Identification of herpes-like DNA sequences in AIDS-associated kaposi's sarcoma. Science 266, 1865-1869.
Chen, Z., Ritzel, R., Lin, C. and Hodgetts, R. (1991) Sequence conservation in avian CR1: An interspersed repetitive DNA family evolving under functional constraints. Proceedings of the National Academic Society 88, 5814-5818.
Choi, T.K., Hollenbach, P.W., Pearson, B.E., Ueda, R.M., Weddell, G.N., Kurahara, C.G., Woodhouse, C.S., Kay, R.M. and Loring, J.F. (1993) Transgenic mice containing a human heavy chain immunoglobulin gene fragment cloned in a yeast artificial chromosome. Nature Genetics 4, 117-123.
Churchill, G. and Doerge, R. (1994) Empirical threshold values for quantitative trait mapping. Genetics 138, 963-971.
Collins, F.S. (1992) Positional cloning; Let's not call it reverse anymore. Nature Genetics 1, 3-6.
Copeland, N.G. and Jenkins, N.A. (1991) Development and applications of a molecular genetic linkage map of the mouse genome. T.I.G. 7, 113-119.
Cox, D.R., Burmeister, M., Price, E.R., Kim, S. and Myers, R.M. (1990) Radiation hybrid mapping: A somatic cell genetic method for constructing high-resolution maps of mammalian chromosomes. Science 250, 245-250.
Crain, W., Eden, F., Pearson, W., Davidson, E. and Britten, R. (1976) Absence of short period interspersion of repetitive and non-repetitive sequences in the DNA of Drosophila melanogaster. Chromosoma56, 309-326.
Crittenden, L.B., Provencher, L., Santangelo, L., Levin, I., Abplanalp, H., Briles, R.W., Briles, W.E. and Dodgson, J.B. (1993) Characterisation of a Red Jungle Fowl by White Leghorn backcross reference population for molecular mapping of the chicken genome. Poultry Science 72, 334-348.
Crooijmans, R.P.M.A., van Oers, P., Strijk, J., Van Der Poel, J. and Groenen, M. (1996) Preliminary linkage map of the chicken (Gallus domesticus) genome based on microsatellite markers: 77 new markers mapped. Poultry Science 75, 746-754.
Davidson, E., Hough, B., Amenson, C. and Britten, R. (1973) General interspersion of repetitive with non-repetitive sequence elements in the DNA of Xenopus. Journal of Molecular Biology 77, 1-23.
Dean, M., White, M.B., Amos, J., Gerrard, B., Stewart, C., Khaw, K.-T. and Leppert, M. (1990) Multiple mutations in highly conserved residues are found in mildly affected cystic fibrosis patients. Cell 61, 863-870.
Delaney, D.E., Friebe, B.R., Hatchett, J.H., Gill, B.S. and Hulbert, S.H. (1995) Targeted mapping of rye chromatin in wheat by representational difference analysis. Genome 38, 458-466.
Desjardins, P. and Morais, R. (1990) Sequence and gene organisation of the chicken mitochondrial genome. J. Mol. Biol. 212, 599-634.
Donis-Keller, H., Green, P., Heims, C., Cartinhour, S., Weiffenbach, B., Stephens, K., Keith, T.P., Bowden, D.W., Smith, D.R., Lander, E.S., Botstein, D., Akots, G., Rediker, K.S., Gravius, T., Brown, V.A., Rising, M.B., Parker, C., Powers, J.A., Watt, D.E., Kauffman, E.R., Bricker, A., Phipps, P., Muller-Kahle, H., Fulton, T.R., Ng, S., Schumm, J.W., Braman, J.C., Knowlton, R.G., Barker, D.F., Crooks, S.M., Lincoln, S.E., Daly, M.J. and Abrahamson, J. (1987) A genetic linkage map of the human genome. Cell51, 319-337.
Dunnington, E.A., Stallard, L.C., Siegel, P.B. and Hillel, J. (1993) Evaluating linkage between DNA fingerprint bands and quantitative traits in chickens: Interactions. Poultry Science 72, 1000-1004.
Eden, F.C. and Hendrick, J.P. (1978) Unusual organisation of DNA sequences in the chicken. Biochemistry 17, 5838-5844.
Elgar, G., Sandford, R., Aparico, S., Macrae, A., Venkatesh, B. and Brenner, S. (1996) Small is beautiful: Comparative genomics with the pufferfish (Fugu rubripes). Trends in Genetics 12, 145-150.
Epplen, J., Leipoldt, M., Engel, W. and Schmidtke, J. (1978) DNA sequence organisation in avian genomes. Chromosoma 69, 307-321.
Fecheimer, N.S.: Chromosomes of chickens. In: Mcfeely, R.A. (Ed.), Domestic Animal Cytogenetics. Advances in Veterinary science and comparative medicine. Academic Press Inc., (1990) pp. 169-207.
Fukomoto, Y., Kaibuchi, K., Hori, Y., Fujioka, H., Araki, S., Ueda, T., Kikuchi, A. and Takai, Y. (1990) Molecular cloning and characterisation of a novel type of regulatory protein (GDI) for the rho proteins, ras p21-like small GTP-binding proteins. Oncogene 5, 1321-1328.
Gilmour, D.G., Brand, A., Donnelly, N. and Stone, H.A. (1976) Bu-1 and Th-1, two loci determining surface antigens of B or T lymphocytes in the chicken. Immunogenetics 3, 549-563.
Giovannoni, J.J., Wing, R.A., Ganal, M.W. and Tanksley, S.D. (1991) Isolation of molecular markers from specific chromosomal intervals using DNA pools from existing mapping populations. N.A.R. 19, 6553-6558.
Girard-Santosuosso, O., Menanteau, P., Lantier, I., Mariani, P., Protias, J., Colin, P., Guillot, J., Bumstead, N., Pardon, P., Beaumont, C. and Lantier, F. (In Press) Genetic control of resistance to Salmonella enteritidis in chicken. Animal Genetics.
Goss, S.J. and Harris, H. (1975) New method for mapping genes in human chromosomes. Nature 255, 680-684.
Gray, J.W., Carrano, A.V., Steinmetz, L.L., Van Dilla, M.A., Moore II, D.H., Mayall, B.H. and Medelsohn, M.L. (1975) Chromosome measurement and sorting by flow systems. Proc. Natl. Acad. Sci. USA. 72, 1231-1234.
Greulich, K.O. (1992) Chromosome microtechnology: microdissection and microcloning. TIBTECH 10, 48-51.
Grosschedl, R., Geise, K. and Pagel, J. (1994) HMG domain proteins: architectural elements in the assembly of nucleoprotein structures. T.I.G.10, 94-99.
Guillemot, F., Billault, A., Pourquié, O., Béhar, G., Chaussé, A.-M., Zoorob, R., Kreibich, G. and Auffray, C. (1988) A molecular map of the chicken major histocompatibility complex: The class II [beta] genes are closely linked to the class I genes and the nucleolar organiser. The EMBO Journal 7, 2775-2785.
Guillemot, F. and Auffray, C. (1989) Molecular biology of the chicken major histocompatibility complex. Poultry Biology 2, 255-274.
Hadjiolov, A.A., Georgiev, O.I., Nosikov, V.V. and Yavachev, L.P. (1984) Primary and secondary structure of rat 28 S ribosomal RNA. N.A.R.12, 3677-3693.
Hála, K., Vilhelmová, M. and Hartmanová, J. (1976) Probable crossing-over in the B blood group system of chickens. Immunogenetics3, 97-103.
Haley, C.S. and Knott, S.A. (1992) A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. Heredity69, 315-324.
Hansen, M.P., Van Zandt, J.N. and Law, G.R.J. (1967) Differences in susceptibility to Marek's disease in chickens carrying two different B locus blood group alleles. Poultry Science 46, 1268.
Heard, E., Simmler, M.-C., Larin, Z., Rougeulle, C., Courtier, B., Lehrach, H. and Avner, P. (1993) Physical mapping and YAC contig analysis of the region surrounding Xist on the mouse X chromosome. Genomics15, 559-569.
Hedemand, J.E., Sorensen, P., Ducro, B.J. and Simonsen, M. (1993) Resistance to Marek's disease associated with B²¹-like haplotypes. Archiv. fur Geflugelkunde 57, 73-76.
Hedges, S.B., Moberg, K.D. and Maxson, L.R. (1990) Tetrapod phylogeny inferred from 18S and 28S ribosomal RNA sequences and a review of the evidence for amniote relationships. Molecular and BIological Evolution 7, 607-633.
Hino, O., Kobayashi, E., Nishizawa, M., Kubo, Y., Kobayashi, T., Hirayama, Y., Takai, S., Kikuchi, Y., Tsuchiya, H., Orimoto, K., Kajino, K., Takahara, T. and Mitani, H. (1995) Renal carcinogenesis in the Eker rat. Journal of Cancer Research and Clinical Oncology 121, 602-605.
Holmes, N.G., Mellersh, C.S., Humphreys, S.J., Binns, M.M., Holliman, A., Curtis, R. and Sampson, J. (1993) Isolation and characterisation of microsatellites from the canine genome. Animal Genetics 24, 289-292.
Horvat, S. and Medrano, J.F. (1996) Fine genetic mapping of the region surrounding the high growth (hg) locus in mouse chromosome 10: Targeting random amplified polymorphic DNA (RAPD) markers. Mammalian Genome 7, 312-314.
Hu, J., Bumstead, N., Gros, P. and Malo, D. (1996) Structural organization, sequence, and expression of the chicken NRAMP1 gene encoding the natural resistance-associated macrophage protein 1. DNA and Cell Biology15, 113-123.
Hubank, M. and Schatz, D.G. (1994) Identifying differences in mRNA expression by representational difference analysis of cDNA. N.A.R.22, 5640-5648.
Hughes, A. and Hughes, M. (1995) Small genomes for better flyers. Nature377, 391.
Hutt, F. (1936): Genetics of the fowl VI. A tentative chromosome map. "Neue forschungen tierzucht und abstammungslehre" (Duerst Festschrift). In: AG, V. (Ed.), Bern, pp. 105-112.
Hutt, F.B.: Genetics of the Fowl. McGraw-Hill Book Company Inc, New York, (1949).
Jakobovits, A., Moore, A.L., Green, L.L., Vergara, G.J., Maynard-Currie, C.E., Austin, H.A. and Klapholz, S. (1993) Germ-line transmission and expression of a human-derived yeast artificial chromosome. Nature 362, 255-258.
Jeffreys, A.J., Wilson, V. and Thein, S.L. (1985) Individual-specific "fingerprints" of human DNA. Nature 316, 76-79.
Juul-Madsen, H.R., Hedemand, J.E., Salomonsen, J. and Simonsen, M. (1993) Restriction fragment length polymorphism analysis of the chicken B-F and B-L genes and their association with serologically defined B haplotypes. Animal Genetics 24, 243-247.
Kaufman, J., Skjoedt, K. and Salomonsen, J. (1990) The MHC molecules of nonmammalian vertebrates. Immunological Reviews 113, 83-117.
Kaufman, J., Bumstead, N., Miller, M., Reigert, P. and Salomonsen, J.: The chicken class II A gene is located outside of the B complex. In: Davison, T., Bumstead, N. and Kaiser, P. (Eds.), Advances in Avian Immunology Research. Carfax, (1993) pp. 119-127.
Kaufman, J., Völk, H. and Wallny, H.-J. (1995) A "Minimal essential Mhc" and an "Unrecognised Mhc": Two extremes in selection for polymorphism. Immunological Reviews 143, 63-88.
Kemp, S.J. and Teale, A. (1994) Randomly primed PCR amplification of pooled DNA reveals polymorphism in a ruminant repetitive DNA sequence which differentiates Bos indicus and B. taurus. Animal Genetics25, 83-88.
Khatib, H., Genislav, E., Crittenden, L.B., Bumstead, N. and Soller, M. (1993) Sequence-tagged microsatellite sites as markers in chicken reference and resource populations. Animal Genetics 24, 355-362.
Korenberg, J.R., Yang-Feng, T., Schreck, R. and Chen, X.N. (1992) Using fluorescence in situ hybridization (FISH) in genome mapping. TIBTECH10, 27-32.
Kunkel, L.M., Monaco, A.P., Middlesworth, W., Ochs, H.D. and Latt, S.A. (1985) Specific cloning of DNA fragments absent from the DNA of a male patient with an X chromosome deletion. Proc. Natl. Acad. Sci. USA.82, 4778-4782.
Kwok, S.C.M., Ledley, F.D., DiLella, A.G., Robson, J.H. and Woo, S.L.C. (1985) Nucleotide sequence of a full-length complementary DNA clone and amino acid sequence of human phenylalanine hydroxylase. Biochemistry24, 556-561.
Lamont, S., Bolin, C. and Cheville, N. (1987) Genetic resistance to fowl cholera is linked to the major histocompatibility complex. Immunogenetics25, 284-289.
Lamont, S. (1993).: The major histocompatibility complex in chickens. In: Etches, R. and Verrinder Gibbons, A. (Eds.), Manipulation of the Avian Genome. CRC Press, Guelph.
Lander, E.S. and Kruglyak, L. (1995) Genetic dissection of complex traits: Guidelines for interpreting and reporting linkage results. Nature Genetics11, 241-247.
Law, W. and Payne, L.: The Poultry Industry. In: Jordan, F. and Pattison, M. (Eds.), Poultry Disease. W.B. Saunders Company Ltd, London, (1996) pp. 1-8.
Lee, S.W., Tomasetto, C. and Sager, R. (1991) Positive selection of candidate tumor-suppressor genes by subtractive hybridization. Proc. Natl. Acad. Sci. USA. 88, 2825-2829.
Levin, I., Crittenden, L.B. and Dodgson, J.B. (1993) Genetic map of the chicken Z chromosomes using random amplified polymorphic DNA (RAPD) markers. Genomics 16, 224-230.
Levin, I., Crittenden, L.B. and Dodgson, J.B. (1994a) Mapping DNA polymorphisms using PCR primers from the sequence of an avian CR1 element. The Journal of Heredity 85, 73-78.
Levin, I., Santangelo, L., Cheng, H., Crittenden, L.B. and Dodgson, J.B. (1994b) An autosomal genetic linkage map of the chicken. The Journal of Heredity 85, 79-85.
Liang, P. and Pardee, A.B. (1992) Differential display of eukaryotic messenger RNA by means of the polymerase chain reaction. Science257, 967-971.
Liang, P., Averboukh, L. and Pardee, A.B. (1993) Distribution and cloning of eukaryotic mRNAs by means of differential display: Refinements and optimisation. N.A.R.21, 3269-3275.
Lindahl, T. and Nyberg, B. (1972) Rate of depurination of native deoxyribonucleic acid. Biochemistry 11, 3610-3618.
Lisitsyn, N., Lisitsyn, N. and Wigler, M. (1993) Cloning the differences between two complex genomes. Science 259, 946-951.
Lisitsyn, N.A., Segre, J.A., Kusumi, K., Lisitsyn, N.M., Nadeau, J.H., Frankel, W.N., Wigler, M.H. and Lander, E.S. (1994) Direct isolation of polymorphic markers linked to trait by genetically directed representational analysis. Nature Genetics 6, 57-63.
Little, P.F.R. (1992) Generating ´cloned DNA maps'. TIBTECH10, 33-35.
Lognonne, J.L. (1993) Introduction to pulsed-field gel electrophoresis. Methods in Molecular and Cellular Biology 4, 49-55.
Longnecker, B., Pazderka, F., Gavora, J., Spencer, J. and Ruth, R. (1976) Lymphoma induced by herpesvirus: Resistance associated with a major histocompatibility gene. Immunogenetics 3, 401-407.
Lovett, M., Kere, J. and Hinton, L.M. (1991) Direct selection; A method for the isolation of cDNAs encoded by large genomic regions. Proc. Natl. Acad. Sci. USA. 88, 9628-9632.
MacPherson, J.M., Eckstein, P.E., Scoles, G.J. and Gajadhar, A.A. (1993) Variability of the random amplified polymorphic DNA assay among thermal cyclers, and effects of primer and DNA concentration. Molecular and Cellular Probes 7, 293-299.
Manly, K. and Cudmore, R.: MapManager Software ver 2.6.5, (1988).
Mariani, P., Crittenden, L., Cheng, H., Wain, H., Vignal, A. and Bumstead, N. (In press) Current state of the genetic map of the chicken. Animal Genetics.
Martinez, O. and Curnow, R. (1992) Estimating the locations and the sizes of the effects of quantitative trait loci using flanking markers. Theoretical and Applied Genetics 85, 480-488.
McQueen, H., Fantes, J., Cross, S., Clark, V., Archibald, A. and Bird, A. (1996) CpG islands of chicken are concentrated on microchromosomes. Nature Genetics 12, 321-324.
Merrill, G.F., Harland, R.M., Groudine, M. and McKnight, S.L. (1984) Genetic and physical analysis of the chicken tk gene. Molecular and Cellular Biology 4, 1769-1776.
Meunier, J.-R. and Grimont, P.A.D. (1993) Factors affecting reproducibility of random amplified polymorphic DNA fingerprinting. Res. Microbiol.144, 373-379.
Michelmore, R.W., Paran, I. and Kesseli, R.V. (1991) Identification of markers linked to disease-resistance genes by bulked segregant analysis: A rapid method to detect markers in specific genomic regions by using segregating populations. Proceedings of the National Academic Society 88, 9828-9832.
Mignery, G.A., Newton, C.L., Archer III, B.T. and Sudof, T.C. (1990) Structure and expression of the rat inositol 1,4,5-triphosphate receptor. The Journal Of Biological Chemistry 265, 12679-12685.
Miller, M.M., Goto, R., Bernot, A., Zoorob, R., Auffray, C., Bumstead, N. and Briles, W.E. (1994a) Two Mhc class I and two Mhc class II genes map to the chicken Rfp-Y system outside the B complex. Proc. Natl. Acad. Sci. USA. 91, 4397-4401.
Miller, M.M., Goto, R.M., Wakenell, P., Jarvi, S., Bernot, A., Auffray, C., Zoorob, R., Bumstead, N. and Briles, W.E. (1994b): The RFP-Y system - A second system of histocompatibility genes in the chicken. In: Davison, T., Bumstead, N. and Kaiser, P. (Eds.), Advances in Avian Immunology Research. Carfax, pp. 95-103.
Miller, M.M., Goto, R.M., Taylor, R.L., Zoorob, R., Auffray, C., Briles, R.W., Briles, W.E. and Bloom, S.E. (1996) Assignment of Rfp-Y to the chicken major histocompatibility complex/NOR microchromosome and evidence for high-frequency recombination associated with the nucleolar organizer region. Proceedings of the National Academic Society 93, 3958-3962.
Monaco, A.P., Lam, V.M.S., Zehetner, G., Lennon, G.G., Douglas, C., Nizetic, D., Goodfellow, P.N. and Lehrach, H. (1991) Mapping irradiation hybrids to cosmid and yeast artificial chromosome libraries by direct hybridisation of Alu-PCR products. N.A.R. 19, 3315-3318.
Nakabayashi, Y. and Nishigaki, K. (1996) Single-stranded conformation polymorphism (SSCP) can be explained by semistable conformation dynamics of single-stranded DNA. Journal of Biochemistry 120, 320-325.
Nakano, T. and Graf, T. (1992) Identification of genes differentially expressed in two types of v-myb-transformed avian myelmonocytic cells. Oncogene 7, 527-534.
Neibergs, H.L., Dietz, A.B. and Womack, J.E. (1993) Single-strand conformation polymorphisms (SSCPs) detected in five bovine genes. Animal Genetics24, 81-84.
Nelson, S.F., McCusker, J.H., Sander, M.A., Kee, Y., Modrich, P. and Brown, P.O. (1993) Genomic mismatch scanning: a new approach to genetic linkage mapping. Nature Genetics 4, 11-17.
O'Brien, S.J. and Graves, J.A.M. (1991) Report of the committee on comparative gene mapping. Cytogenet. Cell Genet. 58, 1124-1151.
Oh, J., Bailin, T., Fukai, K., Feng, G., Ho, L., Mao, J., Frenk, E., Tamura, N. and Spritz, R. (1996) Positional cloning of a gene for Hermansky-Pudlak syndrome, a disorder of cytoplasmic organelles. Nature Genetics14, 300-306.
Okado, I. and Yamamoto, Y. (1987) Immunocompetences and Marek's disease resistance in three pairs of chicken lines selected for different immunological characters. Poultry Science 66, 769-773.
Okubo, K., Hori, N., Matoba, R., Niiyama, T., Fukushima, A., Kojima, Y. and Matsubara, K. (1992) Large scale cDNA sequencing for analysis of quantitative and qualitative aspects of gene expression. Nature Genetics2, 173-179.
Olofsson, B. and Bernardi, G. (1983) Organisation of nucleotide sequences in the chicken genome. Eur. J. Biochem. 130, 241-245.
Olsen, R.H. and Shipley, P. (1973) Host range and properties of the Pseudomonas aeruginosa R Factor R1822. Journal of Bacteriology113, 772-780.
Olson, M., Hood, L., Cantor, C. and Botstein, D. (1989) A common language for physical mapping of the human genome. Science 245, 1434-1435.
Orita, M., Iwahana, H., Kanazawa, H., Hayashi, K. and Sekiya, T. (1989a) Detection of polymorphisms of human DNA by gel electrophoresis as single-strand conformation polymorphisms. Proc. Natl. Acad. Sci. USA. 86, 2766-2770.
Orita, M., Suzuki, Y., Sekiya, T. and Hayashi, K. (1989b) Rapid and sensitive detection of point mutations and DNA polymorphisms using the polymerase chain reaction. Genomics 5, 874-879.
Paonessa, G., Frank, R. and Cortese, R. (1987) Nucleotide sequence of rat liver HMG1 cDNA. N.A.R. 15, 9077.
Patton, J.G., Porro, E.B., Galceran, J., Tempst, P. and Nadal-Ginard, B. (1993) Cloning and characterisation of PSF, a novel pre-mRNA splicing factor. Genes and Development 7, 393-406.
Payne, L. (1991): Development of new vaccines and other methods for the control of Marek's disease: A plain man's guide, Avian Tumour Virus Symposium. AAAP/AVMA, Seattle, pp. 37-41.
Payne, L. (1996): Marek's Disease. In: Jordan, F. and Pattison, M. (Eds.), Poultry Disease. W.B. Saunders Company Ltd, London, pp. 112-122.
Persico, M., Viglietto, G., Martini, G., Toniolo, D., Paonessa, G., Moscatelli, C., Dono, R., Vulliamy, T., Luzzatto, L. and D'Urso, M. (1986) Isolation of human glucose-6-phosphate dehydrogenase (G6PD) cDNA clones: Primary structure of the protein and unusual 5' non-coding region. Nucleic Acids Research14, 2511-2522.
Pink, J.R.L., Droege, W., Hála, K., Miggiano, V.C. and Ziegler, A. (1977) A three locus model for the chicken major histocompatibility complex. Immunogenetics 5, 203-216.
Plotsky, Y., Cahaner, A., Haberfeld, A., Lavi, U., Lamont, S.J. and Hillel, J. (1993) DNA fingerprint bands applied to linkage analysis with quantitative trait loci in chickens. Animal Genetics 24, 105-110.
Powell, P., Lee, L., Mustill, B. and Rennie, M. (1982) The mechanism of genetic resistance to Marek's disease in chickens. International Journal of Cancer 29, 169-174.
Powell, P.C., Hartley, K.J., Mustill, B.M. and Rennie, M. (1983) Studies on the role of macrophages in Marek's disease of the chicken. Journal of the Reticuloendothelial society 34, 289-297.
Powell, P.C.: Host resistance factors: Immune responses - A review. In: Calnek, B. and Spence, J. (Eds.), Proceedings of the International Symposium on Marek's Disease. AA Avian Pathologists, (1984) pp. 238-261.
Rafalski, J.A. and Tingey, S.V. (1993) Genetic diagnostics in plant breeding: RAPDs, microsatellites and machines. T.I.G. 9, 275-280.
Rajavashisth, T.B., Taylor, A.K., Andalibi, A., Svenson, K.L. and Lusis, A.J. (1989) Identification of a Zinc finger protein that binds to the sterol regulatory element. Science 245, 640-643.
Reitman, M., Grasso, J.A., Blumenthal, R. and Lewit, P. (1993) Primary sequence, evolution, and repetitive elements of the Gallus gallus (Chicken) á-globin cluster. Genomics 18, 616-626.
Rich, B.E. and Steitz, J.A. (1987) Human acidic ribosomal phosphoproteins P0, P1 and P2: Analysis of cDNA clones, in vitro synthesis and assembly. Molecular and Cellular Biology 7, 4065-4074.
Rosenberg, M., Przybylska, M. and Straus, D. (1994) "RFLP subtraction": A method for making libraries of polymorphic markers. Proc. Natl. Acad. Sci. USA. 91, 6113-6117.
Rothuizen, J. and Van Wolferen, M. (1994) Randomly amplified DNA polymorphisms in dogs are reproducible and display Mendelian transmission. Animal Genetics 25, 13-18.
Royer-Pokora, B., Kunkel, L., Monaco, A., Goff, S., Newburger, P., Baehner, R., Cole, F., Curnutte, J. and Orkin, S. (1986) Cloning the gene for an inherited human disorder-chronic granulomatous disease-one the basis of its chromosomal location. Nature 322, 32-38.
Sander, J.E. (1993) The major histocompatibility complex and its role in poultry production. World's Poultry Science Journal 49, 132-138.
Schedl, A., Montoliu, L., Kelsey, G. and Schutz, G. (1993) A yeast artificial chromosome covering the tyrosinase gene confers copy number-dependent expression in transgenic mice. Nature 362, 258-261.
Schierman, L. and Collins, W. (1987) Influence of the major histocompatibility complex on tumour regression and immunity in chickens. Poultry Science66, 812-818.
Schierman, L.W. and Nordskog, A.W. (1961) Relationship of blood type histocompatibility in chickens. Science 134, 1008-1009.
Schierwater, B. and Ender, A. (1993) Different thermostable DNA polymerases may amplify different RAPD products. N.A.R. 21, 4647-4648.
Schutte, M., da Costa, L.T., Hahn, S.A., Moskaluk, C., Hoque, A.T.M.S., Rozenblum, E., Weinstein, C.L., Bittner, M., Meltzer, P.S., Trent, J.M., Yeo, C.J., Hruban, R.H. and Kern, S.E. (1995) Identification by representational difference analysis of a homozygous deletion in pancreatic carcinoma that lies within the BRCA2 region. PNAS 92, 5950-5954.
Schwartz, D.C. and Cantor, C. (1984) Separation of yeast chromosome-sized DNAs by pulsed field gradient gel electrophoresis. Cell 37, 67-75.
Schwartz, D.E., Tizard, R. and Gilbert, W. (1983) Nucleotide sequence of rous sarcoma virus. Cell 32, 853-869.
Simons, J.N., Pilot-Matias, T.J., Leary, T.P., Dawson, G.J., Desai, S.M., Schlauder, G.G., Muerhoff, A.S., Erker, J.C., Buijk, S.L., Chalmers, M.L., Van Sant, C.L. and Mushahwar, I.K. (1995) Identification of two flavivirus-like genomes in the GB hepatitis agent. Proc. Natl. Acad. Sci. USA.92, 3401-3405.
Smith, E., J., Jones, C., P., Bartlett, J. and Nestor, K. (1996) Use of randomly amplified polymorphic DNA markers for the genetic analysis of relatedness and diversity in chickens and turkeys. Poultry Science75, 579-584.
Soller, M. and Beckmann, J.S. (1986) Restriction fragment length polymorphisms in poultry breeding. Poultry Science 65, 1474-1488.
Southern, E.M. (1975) Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98, 503-517.
Spike, C.A. and Lamont, S.J. (1995) Genetic analysis of three loci homologous to human G9a: Evidence for linkage of a class III gene with chicken MHC. Animal Genetics 26, 185-187.
Stockbridge, R.: Microsoft Corporation, ENCARTA, (1995).
Stone, H.A. (1969) Investigations of the genetic control of Marek's disease. Poultry Science 48, 1879.
Straus, D. and Ausubel, F.M. (1990) Genomic subtraction for cloning DNA corresponding to deletion mutations. Proc. Natl. Acad. Sci. USA.87, 1889-1893.
Stumph, W.E., Kristo, P., Tsai, M.-J. and O'Malley, B.W. (1981) A chicken middle-repetitive DNA sequence which shares homology with mammalian ubiquitous repeats. Nucleic Acids Research 9, 5383-5397.
Stumph, W.E., Baez, M., Beattie, W.G., Tsai, M.-J. and O'Malley, B.W. (1983) Characterisation of deoxyribonucleic acid sequences at the 5' and 3' borders of the 100 kilobase pair ovalbumin gene domain. Biochemistry22, 306-315.
Stumph, W.E., Hodgson, C.P., Tsai, M.-J. and O'Malley, B.W. (1984) Genomic structure and possible retroviral origin of the chicken CR1 repetitive DNA sequence family. Proc. Natl. Acad. Sci. USA. 81, 6667-6671.
Sturtevant, A.H. (1913) The linear arrangement of six sex-linked factors in Drosophila, as shown by their mode of association. J. Exp. Zoology14, 43-59.
Tindall, K. and Kunkel, T. (1988) Fidelity of DNA synthesis by the Thermus aquaticus DNA polymerase. Biochemistry 27, 6008-6013.
Toye, A.: Gene mapping in the domestic chicken (Gallus gallus), Genetics. Sydney, Australia, (1993).
Toye, A., Schalkwyk, L., Lerach, H. and Bumstead, N. (In Press) A yeast artificial chromosome (YAC) library containing 10 haploid chicken genome equivalents. Mammalian Genome .
Tregaskes, C. and Young, J.: Cloning of chicken lymphocyte marker cDNAs from eukaryotic expression libraries, Immunology Methods Manual. Academic Press Ltd, (1997), pp. 2295-2314.
Tsuchiya, H., Tsuchiya, Y., Kobayashi, T., Kikuchi, Y. and Hino, O. (1994) Isolation of genes differentially expressed between the Yoshida Sarcoma and long-survival Yoshida Sarcoma variants: Origin of Yoshida Sarcoma revisited. Japanese Journal of Cancer Research 85, 1099-1104.
Vandergon, T.L. and Reitman, M. (1994) Evolution of chicken repeat 1 (CR1) elements; Evidence for ancient subfamilies and multiple progenitors. Molecular and BIological Evolution 11, 886-898.
Vidal, S.M., Malo, D., Vogan, K., Skamene, E. and Gros, P. (1993) Natural resistance to infection with intracellular parasites: Isolation of a candidate for Bcg. Cell 73, 469-485.
Vos, P., Hogers, R., Bleeker, M., Reijans, M., van de Lee, T., Hornes, M., Frijters, A., Pot, J., Peleman, J., Kuiper, M. and Zabeau, M. (1995) AFLP: A new technique for DNA fingerprinting. Nucleic Acids Research23, 4407-4414.
Walter, M.A., Spillett, D.J., Thomas, P., Weissenbach, J. and Goodfellow, P.N. (1994) A method for constructing radiation hybrid maps of whole genomes. Nature Genetics 7, 22-28.
Ward, T. and Davies, K.E. (1993) The leading role of STSs in genome mapping. Human Molecular Genetics 2, 1097-1098.
Warden, C.H., Mehrabian, M., He, K.-Y., Yoon, M.-Y., Diep, A., Xia, Y.-R., Wen, P.-Z., Svenson, K.L., Sparkes, R.S. and Lusis, A.J. (1993) Linkage mapping of 40 randomly isolated liver cDNA clones in the mouse. Genomics18, 295-307.
Weber, J.L. (1990) Informativeness of human (dC-dA)_n (dG-dT)_n polymorphisms. Genomics7, 524-530.
Weinstock, G., Rhys, C., Berman, M., Hampar, B., Jackson, D., Silhavy, T., Weisemann, J. and Zweig, M. (1983) Open reading frame expression vectors: A general method for antigen production in Eschericia coli using protein fusions to [beta]-galactosidase. Proceedings of the National Academic Society 80, 4432-4436.
Welsh, J. and McClelland, M. (1990) Fingerprinting genomes using PCR with arbitrary primers. N.A.R. 18, 7213-7218.
Wigger, G., Fries, R. and Stranzinger, G. (1996) Die Isolierung Y-spezifischer Sonden beim Rind mittels genetischer Repräsentations-Differenz-Analyse. Schweizer Archiv Fur Tierhelkunde 138, 167-171.
Williams, J.G.K., Kubelik, A.R., Livak, K.J., Rafalski, J.A. and Tingey, S.V. (1990) DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. N.A.R. 18, 6531-6535.
Woodward, S.R., Sudweeks, J. and Teuscher, C. (1992) Random sequence oligonucleotide primers detect polymorphic DNA products which segregate in inbred strains of mice. Mammalian Genome 3, 73-78.
Xie, Y.-G., Han, F.-Y., Peyrard, M., Ruttledge, M.H., Fransson, I., DeJong, P., Collins, J., Dunham, I., Nordenskj"ld, M. and Dumanski, J.P. (1993) Cloning of a novel, anonymous gene from a megabase-range YAC and cosmid contig in the neurofibromatosis type 2/meningioma region on human chromosome 22q12. Human Molecular Genetics 2, 1361-1368.
Yamashina, M.Y. (1944) Karyotype studies in birds. I. Comparative morphology of chromosomes in seventeen races of domestic fowl. Cytologia 13, 270-296.
Yotova, I. (1988) Thermoregulatory mechanisms in connection with hereditary resistibility of broiler and egg type hens to Marek's disease. Zhivotnov'Dni-Nauki25, 78-84.
Yotova, I., Marinov, B., Georgieva, V., Stoyanchev, T. and Semerjiev, V. (1990) Relationship between resistibility of different alkaline phosphatase genotype broiler type birds to Marek's disease and their meat performance. Zhivotnov'Dni-Nauki27, 60-65.
Young, J.A.T., Bates, P. and Varmus, H.E. (1993) Isolation of a chicken gene that confers susceptibility to infection by subgroup A avian leukosis and sarcoma viruses. J. Virol. 67, 1811-1816.
Zeng, Z.-B. (1993) Theoretical basis for separation of multiple linked gene effects in mapping quantitative trait loci. Proc. Natl. Acad. Sci. USA.90, 10972-10976.