Identifying recombination events and haplotypes in beef cattle from half‐sib families

Abstract

Identifying the recombination events and haplotype segments that have been passed to offspring is useful for several applications in genomic analysis. In livestock populations half-sib families are common, therefore development of methods for identifying recombination events and phasing in half-sib families will allow us to phase most individuals in the population. The generated haplotypes make the phasing of the remained individuals straight forward. Subsequently haplotypes can be used to build the genomic relationship matrix. In this study we propose a likelihood based method that is more flexible than the rule based methods for development and to handle incorrect genotypes. This algorithm first imputes sire haplotypes from offspring genotypes and then creates a descent graph from parent to offspring that shows the recombination events in the sire and origin of the paternal haplotype in offspring. With knowledge of recombination events and haplotype segments passed to the offering, phasing the haplotypes of offspring become feasible and simple. In addition to the phasing and imputation, this method had several benefits such as identifying pedigree errors and mapping errors. The method was tested on real genotypes of 14,600 beef cattle. The accuracy of sire imputation was = 95% with the average of 98% for the half sib family size of 7 to 79. The average number of recombination events in paternal haplotype strands transferred to the offspring was 26.5 recombinations with a standard deviation of 6.3. In conclusion, the high sire imputation accuracy indicated a high accuracy in haplotype inference. This method can remove the need for sire genotyping and allow us to identify unusual recombination events that may indicate mapping or pedigree errors.