Submitted to: Journal of Animal Science Supplement
Publication Type: Abstract Only
Publication Acceptance Date: 4/21/2016
Publication Date: 7/19/2016
Citation: Freking, B.A., Keele, J.W., Rohrer, G.A. 2016. Scan for allele frequency differences from pooled samples in lines of pigs selected for components of litter size [abstract]. Journal of Animal Science Supplement. 94 (E-Supplement 5):161 (Abstract # 0344).
Technical Abstract: Direct single trait selection within two seasonal replicates for 11 generations resulted in a 1.6 pig advantage for uterine capacity (UC) and a 3.0 advantage for ovulation rate (OR) compared to an unselected control (CO) population. Our objective was to gain insight and identify genetic loci impacted by quantitative selection for these component traits. We utilized historical DNA samples from all three lines obtained five and six generations after selection had ceased. A total of 402 gilts contributed to pooled samples; 8 unique pools per line with an average of 16.6 gilts per pool with paternal-half-sibs kept in the same pool. These 24 DNA pools were applied to Neogen GGP Porcine HD BeadChips. Bead-level normalized X and Y values were analyzed rather than allele calls. Analyses to compare the populations were conducted on individual SNP frequency estimates using REML and also using a sliding 25-SNP window with Wright’s fixation index (FST). Pedigree relationships of all gilts back to the common base population 18 generations prior were utilized to compare to the genomic relationships. The overall relationship between pedigree and genomic information was highly predictive (r**2 = 97.5%); however, OR and UC selected lines differed from CO line with above average genomic relationships observed for the same degree of pedigree relationship. This would be consistent with the idea that selection increased the average within line relationships beyond what was accounted for by drift. Sixteen SNPs had allele frequency differences that exceeded a modest false discovery rate (P < 0.05) with none exceeding a genome-wide level of significance. However, three SNP were closely linked on SSC 13 near 203 Mb, while two SNP each were identified on SSC 5 at 63-65 Mb and on SSC 14 at 22-24 Mb offering additional support to those locations. FST values marked similar chromosomal regions but, again, did not by themselves identify regions that exceeded genome-wide significance. In conclusion, the pooling strategy reduced the cost to initially scan the genome but estimates of allele frequency differences needed to be extreme to exceed differences expected from modeling genetic drift in these populations. Additional samples need to be added to supplement this initial scan.