|King, David - Andy|
|Smith, Timothy - Tim|
Submitted to: American Society of Animal Science Annual Meeting
Publication Type: Abstract Only
Publication Acceptance Date: 3/21/2011
Publication Date: 7/10/2011
Citation: Kuehn, L.A., Keele, J.W., Casas, E., Jones, S., King, D.A., McDaneld, T.G., Smith, T.P.L., Wheeler, T.L. 2011. Use of sample pooling in a genome-wide association study identifies chromosomal regions affecting incidence of bovine respiratory disease [abstract]. Journal of Animal Science. 89 (E-Supplement 1):422 (Abstract #336). Interpretive Summary:
Technical Abstract: We hypothesize that genome-wide association (GWA) based on high-density SNP arrays can be used to identify chromosomal regions affecting disease incidence using a case/control type approach. However, the large sample size required to map a lowly heritable trait like susceptibility to bovine respiratory disease complex (BRDC) makes the cost of such an effort prohibitive. We applied pooling of lung samples from slaughter animals with severe lung lesions (a proxy for incidence of respiratory disease) as the case group, and lungs from animals with no visible lesions as the control group (n = 1,000 for each group), to evaluate sample pooling as an approach to reduce the cost of GWA experiments on complex traits. Comparison of allele frequency estimates for each SNP was used to identify chromosomal regions influencing BRDC. We prepared 10 pools of equal volume lung tissue cores from 100 animals in each of the two groups (20 total pools), and DNA from each pool was genotyped in duplicate. Bead level data was used to estimate allele frequency for each pool and combined within group for comparison between case and control. Distances between pools and their replicates across 775,996 SNP were calculated and used to form a neighbor-joining phylogeny, such that individual SNP have miniscule effect on the resulting phylogeny. Allele frequency differences between case and control groups were conditionally compared using phylogenetic comparative methods, resulting in SNP with genome-wide significant associations (P < 0.05) in seven chromosomal regions on BTA 9, 10, 13, 18, 20, 21, and 26. Two additional regions on BTA 7 and 17 harbor SNP under marginally less stringent correction (false discovery rate, FDR, 1%), while 1 or 2 regions on all bovine chromosomes are significant with a relaxed FDR of 5%. We conclude that the presence of lung lesions at slaughter is influenced by multiple loci, consistent with expectations from the multifactorial nature of BRDC, and that tissue pooling represents an economical means to dissect the genetic influences on this trait.