Pooled Genotyping from Un-Pedigreed Beef Cows in Tropical Environments of Australia

Genotyping pools of DNA samples can be a cost-effective method to conduct genome-wide association studies (GWAS). In animal breeding and genetics, this approach is particularly suited for phenotypes that are difficult to measure or non-existent in the seedstock nucleus, and in situations when phenotypes are cheap, abundant and collected in un-pedigreed animals. Here we present our experience with DNA pooling in the context of pregnancy test outcome at the first re-breeding of tropical composite (Santa Gertrudis) cows raised extensively in the tropical environments of Australia. Blood samples from 959 age-matched heifers were pooled according to their pregnancy status: 644 pregnant and 315 non-pregnant. A total of 69 pools were created, DNA extracted and genotyped using the Bovine 50K SNP chip. The number of DNA samples per pool averaged 13.25 and ranged from 1 for 43 pools, to 25 for 6 pools. We examined the fluorescent intensity signals to explore the suitability of model-based clustering and regression approaches to estimate SNP allele frequencies and compared these with the B-Allele frequencies provided by the Illumina BeadStudio platform. Finally, for the GWAS we adopted numerical algorithms developed for case-control designs that accounts for the variance of the pool specific error in the allele frequency estimation. We found 145 SNP significantly associated with pregnancy status at P-value < 0.0001 (FDR = 3.29%). When the genomic regions of these SNP were surveyed we found that 34 of them were related to genes reported to influence fertility in beef cattle.  Prominent among these genes are PENK and IGF2.