Development of breeding values for susceptibility to virulent footrot in sheep: A strategy to accommodate variable disease progression at time of scoring.

Genetic evaluations utilising footrot scores from industry flocks in their essence, incorporate data from a wide range of challenge environments, resulting in potentially large differences in means, variances and distribution of scores across challenges. The date that commencement of infection occurs is generally unknown, and progression of the infection varies with the prevailing environmental and management conditions, virulence of the bacterium Dichelobacter nodosus, as well as the genetic potential and (permanent) environmental ability of animals to resist footrot. In practice, animals are unlikely to be repeatedly scored to identify the best time for comparison, or monitor development of disease progression. Furthermore, field challenges are limited by the need to treat animals before their welfare is compromised. Therefore, the duration and intensity of infection varies and this affects comparisons between animals for their susceptibility. Diseases such as footrot are characterised by multiple categorical scores reflecting clinical stages that describe the progression and relative impact of the disease. This provides the opportunity for the transformation of the data to a standardised prevalence. Scoring events from multiple footrot field challenges under a standardised protocol were used to establish a series of transition matrices to describe disease progression between scores over time. These transition matrices were used to standardise challenge events to the more severe scoring events, observed later in the challenge. The accuracy of the transition technique was tested by comparing the ranking of animals and sires against the observed scores. Transitioning the data from low disease prevalence to the higher prevalence at the subsequent scoring event improved the correlations between the scoring events, at the animal level, by upwards of 0.10 across challenges. The utilisation of a transition matrix to transform low prevalence disease challenges by taking into account the natural biological rate of progression through the clinical stages of the disease provides a more accurate technique to account for variation in disease prevalence. The transition technique increases the acceptable range of disease expression targeted by producers when scoring virulent footrot challenges reducing the need for repeat scoring and allowing earlier treatment and reducing the impact of the disease on the host animal.

Impact of genotypic errors with equal and unequal family contribution on accuracy of genomic prediction in aquaculture using simulation.

Genotypic errors, conflict between recorded genotype and the true genotype, can lead to false or biased population genetic parameters. Here, the effect of genotypic errors on accuracy of genomic predictions and genomic relationship matrix are investigated using a simulation study based on population and genomic structure comparable to black tiger prawn, Penaeus monodon. Fifty full-sib families across five generations with phenotypic and genotypic information on 53 K SNPs were simulated. Ten replicates of different scenarios with three heritability estimates, equal and unequal family contributions were generated. Within each scenario, four SNP densities and three genotypic error rates in each SNP density were implemented. Results showed that family contribution did not have a substantial impact on accuracy of predictions across different datasets. In the absence of genotypic errors, 3 K SNP density was found to be efficient in estimating the accuracy, whilst increasing the SNP density from 3 to 20 K resulted in a marginal increase in accuracy of genomic predictions using the current population and genomic parameters. In addition, results showed that the presence of even 10% errors in a 10 and 20 K SNP panel might not have a severe impact on accuracy of predictions. However, below 10 K marker density, even a 5% error can result in lower accuracy of predictions.

Use of diagnostic reports to estimate prevalence and distribution of skeletal lesions in young Thoroughbreds.

Diagnostic reports written to assist stud managers in the sale of young Thoroughbreds have not previously been used as a data source for the study of skeletal lesions. However, analyses of these reports may provide efficient and cost-effective insights into the prevalence and distribution of skeletal lesions within a population. Diagnostic reports written by veterinarians were acquired from Thoroughbred stud managers in Australia and New Zealand. The reports were based on approximately 1300 sets of weanling and yearling radiographs taken between 2002 and 2007. The prevalence and anatomical distribution of skeletal lesions in weanlings (299 horses) and yearlings (1004 horses) were determined from these reports. Overall, 69.9% of weanlings and 64.5% of yearlings were reported as having one or more skeletal lesions. Diagnostic reports in weanlings were a strong indication of what was likely to be seen in subsequent yearling reports. These diagnostic reports are typically used by stud managers in the sales process and the potential drawback is that some categories of skeletal lesions may be under-reported. However, there was substantial agreement between the prevalence and distribution of several skeletal lesions reported in this study and those previously reported from direct evaluation of radiographs for Australian and New Zealand Thoroughbred yearlings. Strong agreement was found for osteophytes, enthesiophytes and other modelling in the hocks, and for lesions in the hind fetlocks and stifles. This indicates that written diagnostic reports are a useful and a reliable source of data for the study of some skeletal lesions in young Thoroughbred horses.

Heritability of epistaxis in the Australian Thoroughbred racehorse population.

Post exercise epistaxis, the manifestation of a severe form of exercise-induced pulmonary haemorrhage (EIPH), has been observed in many equine racing populations. Although multiple analyses have suggested that non-genetic factors may lead to the development of this condition, relatively little consensus has been reached regarding its genetic aetiology. The objective of this study was to provide insight into both genetic and non-genetic factors that may contribute to the expression of epistaxis in the Australian Thoroughbred racing population. Racing records and reported epistaxis occurrences were acquired for 117,088 horses entered in races and official barrier trials from 1 August 2000 until 22 February 2011. Heritability was estimated using two different logistic generalised linear mixed models (lifetime epistaxis risk h(2) = 0.27 and individual race epistaxis risk h(2) = 0.50). Sex, age, and year of birth were shown to be significant; however, trainer, jockey, race distance, condition of the track (i.e. 'going'), racecourse, track surface, number of race starters, year and month of race were not significant. Evidence suggests genetic and non-genetic links to EIPH expressed as epistaxis.

QTL and association analysis for skin and fibre pigmentation in sheep provides evidence of a major causative mutation and epistatic effects.

The pursuits of white features and white fleeces free of pigmented fibre have been important selection objectives for many sheep breeds. The cause and inheritance of non-white colour patterns in sheep has been studied since the early 19th century. Discovery of genetic causes, especially those which predispose pigmentation in white sheep, may lead to more accurate selection tools for improved apparel wool. This article describes an extended QTL study for 13 skin and fibre pigmentation traits in sheep. A total of 19 highly significant, 10 significant and seven suggestive QTL were identified in a QTL mapping experiment using an Awassi × Merino × Merino backcross sheep population. All QTL on chromosome 2 exceeded a LOD score of greater than 4 (range 4.4-30.1), giving very strong support for a major gene for pigmentation on this chromosome. Evidence of epistatic interactions was found for QTL for four traits on chromosomes 2 and 19. The ovine TYRP1 gene on OAR 2 was sequenced as a strong positional candidate gene. A highly significant association (P < 0.01) of grandparental haplotypes across nine segregating SNP/microsatellite markers including one non-synonymous SNP with pigmentation traits could be shown. Up to 47% of the observed variation in pigmentation was accounted for by models using TYRP1 haplotypes and 83% for models with interactions between two QTL probabilities, offering scope for marker-assisted selection for these traits.

Resistance to liver fluke infection in the natural sheep host is correlated with a type-1 cytokine response.

Indonesian thin-tail (ITT) sheep can resist infection with Fasciola gigantica but not F. hepatica and presents an ideal model to investigate the mechanisms of liver fluke resistance in a natural host. This study examines the local and systemic immune responses of sheep during Fasciola infection and demonstrates that different anatomical tissues display distinct cytokine profiles consistent with liver fluke migration. The study also reveals a significant difference in the cytokine and antibody profiles of ITT sheep infected with F. gigantica compared with F. hepatica, with a higher ratio of IL-4/IFN-γ mRNA expression and specific IgG1/IgG2 antibodies strongly correlating with pathology. Interestingly, the significant type-1 cytokine profile occurred in the lymph node closest to the site of infection at a time when the effective immune response against F. gigantica liver flukes is thought to occur. When the same F. gigantica infection in the resistant ITT sheep was compared with the susceptible Merino breed, the resistant type-1 phenotype against liver fluke infection was only observed in the ITT sheep. These studies provide the first evidence to suggest that the induction of an early type-1 immune response in this natural sheep host may be responsible for the ability to resist liver fluke infection.

Symmetry of hip dysplasia traits in the German Shepherd Dog in Australia.

Canine hip dysplasia (CHD) is a common and debilitating developmental condition of the canine coxofemoral (hip) joint, exhibiting a multifactorial pattern of inheritance. British Veterinary Association hip traits (BVAHTs) are nine radiographic features of hips used in several countries to ordinally score both the right and left hip of potential breeding candidates to assess their suitability for breeding. The objective of this study was to examine some aspects of the relationship between contralateral scores for each BVAHT in a cohort of 13 124 Australian-registered German Shepherd Dogs. Goodman and Kruskal gamma coefficients of 0.48-0.95 and correlation coefficients of 0.50-0.74 demonstrate that the association between right and left hip scores varies between moderate and strong for BVAHTs. Principal component analysis of scores detected a sizeable left-versus-right effect, a finding supported by symmetry and quasi-symmetry analyses which found that seven of the nine BVAHTs display significant marginal asymmetry. Dogs showing asymmetry for one BVAHT are significantly more likely to display asymmetry at other BVAHTs. When asymmetry is expressed as a binary trait (either symmetrical or asymmetrical), it displays low to moderate heritability. Estimates of genetic correlations between right and left scores are very high for all BVAHTs (>0.945), suggesting right and left scores for each BVAHT are largely determined by the same set of genes. The marginal asymmetries are therefore more likely to be of environmental and non-additive genetic origin. In breeding programmes for CHD, we recommend that scores from both hips be used to estimate breeding values, with a term for side-of-hip included in the model to account for score variation owing to asymmetry.

Deterministic models of breeding scheme designs that incorporate genomic selection.

A deterministic model to calculate rates of genetic gain and inbreeding was used to compare a range of breeding scheme designs under genomic selection (GS) for a population of 140,000 cows. For most schemes it was assumed that the reliability of genomic breeding values (GEBV) was 0.6 across 4 pathways of selection. In addition, the effect of varying reliability on the ranking of schemes was also investigated. The schemes considered included intense selection in male pathways and genotyping of 1,000 young bulls (GS-Y). This scheme was extended to include selection in females and to include a "worldwide" scheme similar to GS-Y, but 6 times as large and assuming genotypes were freely exchanged between 6 countries. An additional worldwide scheme was modeled where GEBV were available through international genetic evaluations without exchange of genotypes. Finally, a closed nucleus herd that used juvenile in vitro embryo transfer in heifers was modeled so that the generation interval in female pathways was reduced to 1 or 2 yr. When the breeding schemes were compared using a GEBV reliability of 0.6, the rates of genetic gain were between 59 and 130% greater than the rate of genetic gain achieved in progeny testing. This was mainly through reducing the generation interval and increasing selection intensity. Genomic selection of females resulted in a 50% higher rate of genetic gain compared with restricting GS to young bulls only. The annual rates of inbreeding were, in general, 60% lower than with progeny testing, because more sires of bulls and sires of cows were selected, thus increasing the effective population size. The exception was in nucleus breeding schemes that had very short generation intervals, resulting in higher rates of both gain and inbreeding. It is likely that breeding companies will move rapidly to alter their breeding schemes to make use of genomic selection because benefits to the breeding companies and to the industry are considerable.

The sheep genome reference sequence: a work in progress.

Until recently, the construction of a reference genome was performed using Sanger sequencing alone. The emergence of next-generation sequencing platforms now means reference genomes may incorporate sequence data generated from a range of sequencing platforms, each of which have different read length, systematic biases and mate-pair characteristics. The objective of this review is to inform the mammalian genomics community about the experimental strategy being pursued by the International Sheep Genomics Consortium (ISGC) to construct the draft reference genome of sheep (Ovis aries). Component activities such as data generation, sequence assembly and annotation are described, along with information concerning the key researchers performing the work. This aims to foster future participation from across the research community through the coordinated activities of the consortium. The review also serves as a 'marker paper' by providing information concerning the pre-publication release of the reference genome. This ensures the ISGC adheres to the framework for data sharing established at the recent Toronto International Data Release Workshop and provides guidelines for data users.

Improving animal and human health through understanding liver fluke immunology.

Sheep, goats and cattle represent the most numerous and economically important agricultural species worldwide used as sources for milk, fibre and red meat. In addition, in the developing world, these species often represent the sole asset base for small-holder livestock farmers and cattle/buffaloes often provide the majority of draught power for crop production. Production losses caused by helminth diseases of these animals are a major factor in extending the cycle of poverty in developing countries and a major food security issue for developed economies. Fasciola spp. are one of the most important zoonotic diseases with a global economic impact in livestock production systems and a poorly defined but direct effect on human health. Improvements in human and animal health will require a concerted research effort into the development of new accurate and simple diagnostic tests and increased vaccine and drug development against Fasciola infections. Here, the use of definitive natural host breeds with contrasting resistance to Fasciola infections is discussed as a resource to contrast parasite-host interactions and identify parasite immune evasion strategies. Such studies are likely to boost the discovery of new vaccine, drug and diagnostic candidates and provide the foundation for future genetic selection of resistant animals.

The influence of extracellular matrix and prolactin on global gene expression profiles of primary bovine mammary epithelial cells in vitro.

An in vitro bovine mammosphere model was characterized for use in lactational biology studies using a functional genomics approach. Primary bovine mammary epithelial cells cultured on a basement membrane, Matrigel, formed three-dimensional alveoli-like structures or mammospheres. Gene expression profiling during mammosphere formation by high-density microarray analysis indicated that mammospheres underwent similar molecular and cellular processes to developing alveoli in the mammary gland. Gene expression profiles indicated that genes involved in milk protein and fat biosynthesis were expressed, however, lactose biosynthesis may have been compromised. Investigation of factors influencing mammosphere formation revealed that extracellular matrix (ECM) was responsible for the initiation of this process and that prolactin (Prl) was necessary for high levels of milk protein expression. CSN3 (encoding kappa-casein) was the most highly expressed casein gene, followed by CSN1S1 (encoding alphaS1-casein) and CSN2 (encoding beta-casein). Eighteen Prl-responsive genes were identified, including CSN1S1, SOCS2 and CSN2, however, expression of CSN3 was not significantly increased by Prl and CSN1S2 was not expressed at detectable levels in mammospheres. A number of novel Prl responsive genes were identified, including ECM components and genes involved in differentiation and apoptosis. This mammosphere model is a useful model system for functional genomics studies of certain aspects of dairy cattle lactation.

A functional genomics approach to evaluate candidate genes located in a QTL interval for milk production traits on BTA6.

The potential genetic and economic advantage of marker-assisted selection for enhanced production in dairy cattle has provided an impetus to conduct numerous genome scans in order to identify associations between DNA markers and future productive potential. One area of focus has been a quantitative trait locus on bovine chromosome 6 (BTA6) found to be associated with milk yield, milk protein and fat percentage, which has been subsequently fine-mapped to six positional candidate genes. Subsequent investigations have yet to resolve which of the potential positional candidate genes is responsible for the observed associations with productive performance. In this study, we analysed candidate gene expression and the effects of gene knockdown on expression of beta- and kappa-casein mRNA in a small interfering RNA transfected bovine in vitro mammosphere model. From our expression studies in vivo, we observed that four of the six candidates (ABCG2, SPP1, PKD2 and LAP3) exhibited differential expression in bovine mammary tissue over the lactation cycle, but in vitro functional studies indicate that inhibition of only one gene, SPP1, had a significant impact on milk protein gene expression. These data suggest that the gene product of SPP1 (also known as osteopontin) has a significant role in the modulation of milk protein gene expression. While these findings do not exclude other positional candidates from influencing lactation, they support the hypothesis that the gene product of SPP1 is a significant lactational regulatory molecule.

Predicting genetic merit for mastitis and fertility in dairy cattle using genome wide selection and high density SNP screens.

Two novel methods for genome wide selection (GWS) were examined for predicting the genetic merit of animals using SNP information alone. A panel of 1,546 dairy bulls with reliable EBVs was genotyped for 15,380 SNPs that spanned the whole bovine genome. Two complexity reduction methods were used, partial least squares (PLS) and regression using a genetic algorithm (GAR), to find optimal solutions of EBVs against SNP information. Extensive internal cross-validation was used tofind the best predictive models followed by external validation (without direct use of the pedigree or SNP location). Both PLS and GAR provided both accurate fit to the training data set for somatic cell count (SCC) (max r = 0.83) and fertility (max r = 0.88) and showed an accuracy of prediction of r = 0.47 for SCC, and r = 0.72 for fertility. This is the first empirical demonstration that genome wide selection can account for a very high proportion of additive genetic variation in fitness traits whilst exploiting only a small percentage of available SNP information, without use of pedigree or QTL mapping. PLS was computationally more efficient than GAR.

Host responses during experimental infection with Fasciola gigantica and Fasciola hepatica in Merino sheep II. Development of a predictive index for Fasciola gigantica worm burden.

This study reports on the predictive relationship between serological, immunological and pathological responses following experimental inoculation with incremental doses of Fasciola gigantica in sheep. Fifty, 6-month-old, naive Merino wethers were allocated to one of 5 experimental groups, four of which received 50, 125, 225 and 400 metacercariae, respectively, whilst a 5th group acted as non-inoculated control. Strong individual correlations were observed between liver score, GLDH (glutamate dehydrogenase), GGT (gamma glutamyl transferase), CatL5 (cathepsin L5) antibody titre (IgG1, IgA), eosinophilia, and the total worm count or worm biomass. A combination of multiple indicator traits performed significantly better than any single indicator trait alone. The best predictive index accounted for up to 88% of observed worm burden (Wb) if information on inoculation dose was available. Without knowledge of inoculation dose, such as under field conditions, up to 67% of variation in worm burden could be predicted. In contrast, the best single predictor variable (liver damage score) accounted for up to 50% of worm burden, and in the absence of post-slaughter information, serum levels of anti-cathepsin IgA antibody titres accounted for 35% of predicted variation in worm burden. The utility of a predictive index under both field and experimental inoculation conditions is discussed.

The effect of footrot on weight gain in sheep.

Footrot is a highly contagious bacterial disease of sheep affecting the interdigital skin and surrounding soft and hard horn of a hoof, often resulting in severe lameness. This study was aimed at estimating the effect of footrot on weight gain of affected animals, and characterising the variation between animals in terms of phenotypic, environmental and genetic components. A general approach was developed describing the relationship between the disease and weight gain, defining new traits such as the maximum weight loss as a result of disease and the time after infection that this occurs. In two trials, 1267 Merino sheep were artificially challenged with footrot when 10 months old and re-infected through exposure to footrot on pasture 33 weeks later. Their feet were scored for footrot and live weights were measured approximately every 3 weeks. From data on animals that were not affected by footrot throughout each trial, normal growth curves were calculated and applied to affected animals to predict their growth had they remained healthy, so that weight loss as a result of footrot could be predicted. Animals with average footrot severity in the two trials suffered weight losses of 0.5 to 2.5 kg live weight, but most animals regained lost live weight later in the trials as footrot healed following vaccination. The estimates of the heritabilities of weight loss, adjusted for the severity of footrot, were about 0.30 and 0.15 in the experimental and natural challenge groups, respectively. Animals with higher genotypic values for weights at the start of each trial appeared to cope better with infections, in terms of lower weight losses. The time of highest footrot score and the time of maximum weight loss after infection had only very small genetic components.

Effect of myostatin F94L on carcass yield in cattle.

In this study, a highly significant quantitative trait locus (QTL) for meat percentage, eye muscle area (EMA) and silverside percentage was found on cattle chromosome 2 at 0-15 cM, a region containing the positional candidate gene growth differentiation factor 8 (GDF8), which has the common alias myostatin (MSTN). Loss-of-function mutations in the MSTN gene are known to cause an extreme 'double muscling' phenotype in cattle. In this study, highly significant associations of MSTN with cattle carcass traits were found using maternally inherited MSTN haplotypes from outbred Limousin and Jersey cattle in a linkage disequilibrium analysis. A previously reported transversion in MSTN (AF320998.1:g.433C>A), resulting in the amino acid substitution of phenylalanine by leucine at position 94 of the protein sequence (F94L), was the only polymorphism consistently related to increased muscling. Overall, the size of the g.433C>A additive effect on carcass traits was moderately large, with the g.433A allele found to be associated with a 5.5% increase in silverside percentage and EMA and a 2.3% increase in total meat percentage relative to the g.433C allele. The phenotypic effects of the g.433A allele were partially recessive. This study provides strong evidence that a MSTN genotype can produce an intermediate, non-double muscling phenotype, which should be of significant value for beef cattle producers.

Genome-wide genetic diversity of Holstein Friesian cattle reveals new insights into Australian and global population variability, including impact of selection.

Past breeding strategies for dairy cattle have been very effective in producing rapid genetic gain to achieve industry targets and raise profitability. Such gains have been largely facilitated by intense selection of sires combined with the use of artificial insemination. However, this practice can potentially limit the level of genetic diversity through inbreeding and selection plateaus. The rate of inbreeding in Australia is increasing, primarily as a result of semen importation from a small number of prominent bulls from the USA. The effect of this genetic influx in the Australian dairy cattle population is poorly understood both in terms of diversity and local adaptation/divergence. This study uses 845 genome-wide SNP genetic markers and 431 bulls to characterize the level of genetic diversity and genetic divergence within the Australian and international Holstein Friesian dairy population. No significant differences in genetic diversity (as measured by heterozygosity [H(o)] and allelic richness [A]) were observed over the 25-year time period (1975-1999) for bulls used in Australia. The importation of foreign semen into Australia has increased the effective population size until it was in effect a sub-sample of the global population. Our data indicate that most individuals are equally closely related to one another, regardless of country of origin and year of birth. In effect, the global population can be considered as one single population unit. These results indicate that inbreeding, genetic drift and selection has had little effect at reducing genetic diversity and differentiating the Australian Holstein Friesian population at a genome-wide level.

Host responses during experimental infection with Fasciola gigantica or Fasciola hepatica in Merino sheep I. Comparative immunological and plasma biochemical changes during early infection.

This study reports the early biochemical changes in plasma, comparative host-immune responses and parasite recovery data in Merino sheep during the first 10 weeks of infection with Fasciola gigantica and Fasciola hepatica. One group of sheep were uninfected, four groups of sheep received incremental challenge doses of F. gigantica metacercariae (50, 125, 225 and 400, respectively) and the sixth group was challenged with 250 F. hepatica metacercariae. At 10 weeks post infection (wpi), sheep challenged with F. hepatica showed the greatest fluke recovery (mean 119, range 84-166); a significantly higher biomass of parasites recovered (2.5-fold greater than the highest dose of F. gigantica); and a greater mean % parasite recovery (39.3%, range 27-55%) than any group challenged with F. gigantica. Within the groups dosed with F. gigantica a strong dose-dependent response was observed in both fluke recovery and fluke biomass with increasing dose of metacercariae. The mean % parasite recovery of F. gigantica infected groups 1-5 were 26, 23, 26 and 25%, respectively, suggesting a uniform viability of parasite establishment independent of infection dose. At 6 wpi, elevated levels of plasma GLDH were observed in the F. gigantica infected groups compared to the uninfected sheep (p<0.005) whereas the F. hepatica challenged group had four-fold higher levels of GLDH compared to the F. gigantica infected group (p<0.001). Elevated levels of GGT as an indicator of epithelial damage in the bile duct was only seen in the group challenged with F. hepatica at 10 wpi when it rose from below 100 IU/l to approximately 250 IU/l (p<0.0001) whereas no detectable increase in GGT was observed in any of the groups challenged with F. gigantica. The white blood cell response to F. hepatica infection was biphasic with the initial peak at 4 wpi and a second peak at 9 wpi, corresponding to the period of migration of juvenile fluke in the liver and the time when adult flukes are migrating into the bile duct, respectively. This biphasic response was also evident in the changes in the eosinophil counts and serum haemoglobin levels. There was a trend toward higher parasite-specific IgG2 titres in sheep infected with lower worm burdens, suggesting that higher F. gigantica or F. hepatica burdens suppress IgG2 responses. The findings of this study suggest that, in early infection in a permissive host, F. hepatica appears to be more pathogenic than F. gigantica because of its rapid increase in size and the speed of its progression through the migratory phases of its life cycle.

Identification of two allelic forms of ovine CD4 exhibiting a Ser183/Pro183 polymorphism in the coding sequence of domain 3.

The ovine CD4 cDNA sequence from four sheep sources (Australian Merino, Indonesian Thin Tail, Canadian cross bred, Prealpes du sud) predicts a protein of 455 residues with position 130 in the V2 domain exhibiting a W instead of C suggesting that, like the white whale, dog and cat sequences, sheep CD4 contains only two disulphide bonds. The sequence shows 73% amino acid identity and 83% nucleotide identity to a CD4 sequence from the white whale and significant identity to a partial sequence (314 residues) of bovine CD4 (87% amino acid identity, 93% nucleotide identity). Phylogenetic analysis showed that the ovine CD4 sequence forms a clade with the pig, white whale, dolphin, dog and cat CD4. Two forms of ovine CD4 were identified which differ by a single base pair (T/C) in their cDNA sequence at position 622. This polymorphism is also present in sheep genomic DNA in Hardy-Weinberg equilibrium, suggesting that at least two alleles of CD4 exist in the ovine genome with no selection for a particular allele. This polymorphism changes the first codon position of amino acid 183 and results in a Pro/Ser substitution in the N-terminal region of domain 3 of the CD4 protein.