The selection of animals based on DNA has revolutionised animal breeding, but the associated high cost of obtaining genotypes has limited its uptake. The object of this thesis was to develop a low-cost, low-density, multi-purpose genotyping panel for the procurement of reliable genotype information, and to quantify the long-term consequences of using such low-density genotype panels in breeding programs. The objectives were achieved through a combination of real-time cattle genotype data and simulations to mimic a cattle population. The in-silico development of genotype panels was based on actual genotypes from up to 58,705 beef cattle. Alternative novel strategies were used to identify informative single nucleotide polymorphisms (SNPs) with the purpose of developing a low density genotype panel which, once imputed to higher density, could be used in genomic evaluations. A minimum of 3,000 carefully selected SNPs were required if selected for use in a single breed, but this recommendation increased to 6,000 SNPs if the panel was to be applicable for imputation across breeds. Analysis using simulations suggest that successive imputation across generations, using approaches currently adopted globally, was suboptjmal with an accumulation of errors over generations ; an approach based on stepwise generational imputation was recommended to reduce the erosion in accuracy of imputed genotypes over generations. Results also indicated that a minimum of 300 SNPs were required to accurately predict breed composition with a mean standard error of prediction of 0.036. The advent of genomic evaluations has reduced generation intervals intensifying the importance of screening for DNA variants contributing to congenital defects. Using high density genotype data from 45 phenotypically normal calves and 23 half-sib contemporaries with a missing cleft pallet phenotype, a region on chromosome 13 was identified to harbour the underlying mutation. Analysis of sequence data from a selection of these animals identified putative causal mutations. In conclusion, appropriately designed genotype panels, coupled with optimised imputations strategies, can facilitate the widespread adoption of low-cost accurate genome-based breed programs - (Abstract)

This Ph.D is a joint research project between Teagasc Moorepark and Cork Institute of Technology.

Thesis Cork Institute of Technology, 2017.

Bibliography: (pages 153-166)