Evaluating haplotype diversity within and between Australian sheep breeds

Genomic data can be used to predict differences in phenotype or breeding values between individuals, using linkage disequilibrium (LD) between marker and putative quantitative trait loci (QTL). So far, most genome association studies are based on genotypes at individual loci, and although this has given reasonable accuracies of predicting breeding value (Roos et al. (2008)), there is limited evidence of underlying QTL effects being consistant at the basis of such predictions. One problem with single markers in dense genotypic data is that different loci can easily be in LD by random chance, and SNPs apparently linked to QTL effects may have limited predictive ability in data from individuals that are genetically less related. When using ordered genotypes and information on haplotype similarity, the power of predicting QTL effects can be increased. If two individuals share the same extended haplotype over the same genomic region, the chance that they carry the same marker-QTL allele relationship by descent is much higher. In this paper, we explore the extent by which various haplotype lengths are shared within and between 4 Australian sheep breeds. We count how many haplotypes are present in the breed for a given number of loci, and how many of these haplotypes are shared between breeds. Such statistics will become important when assessing identity by descent (IBD) probabilities and how well these can be separated from identity by state (IBS) probabilities.