A Monte Carlo approach for estimation of haplotype probabilities

Abstract

A number of studies have tested for effects associated with different alleles of the milk protein genes on production traits in dairy cattle (e.g. Cowan et al., 1992). Because numerous protein loci exist and several are closely linked in the genome, one may also be interested in estimating the effects of combinations of alleles at more than one loci. One approach for this type of analysis is multiple regression on all loci and their interactions. An alternative strategy is an analysis of haplotype effects. Advantages of this strategy are that 1) the number of haplotypes in a population is typically much lower than the number of allelic combinations among loci, thus resulting in more precision and statistical power; and 2) haplotype effects can account for effects of unknown loci flanked by the genotyped loci. Ikonen et al. (2001) applied such an approach to test for effects of β and κ casein on production traits of Finnish Ayrshire cattle. Effects were observed for protein content in particular. One obstacle to haplotype analysis is that genotyping procedures generally do not provide information about the recombination phase between loci, necessary for obtaining haplotypes.