|KACHMAN, STEPHEN - University Of Nebraska|
|SPANGLER, MATTHEW - University Of Nebraska|
|Thallman, Richard - Mark|
Submitted to: Journal of Animal Science Supplement
Publication Type: Abstract Only
Publication Acceptance Date: 5/22/2017
Publication Date: 7/8/2017
Citation: Snelling, W.M., Kachman, S.D., Bennett, G.L., Spangler, M.L., Kuehn, L.A., Thallman, R.M. 2017. Functional SNP associated with birth weight in independent populations identified with a permutation step added to GBLUP-GWAS [abstract]. Journal of Animal Science Supplement. 95(Supplement 4):97-98. https://doi.org/10.2527/asasann.2017.197.
Technical Abstract: This study was conducted as an initial assessment of a newly available genotyping assay containing about 34,000 common SNP included on previous SNP chips, and 199,000 sequence variants predicted to affect gene function. Objectives were to identify functional variants associated with birth weight in the Germplasm Evaluation (GPE) project, a crossbred population representing 18 breeds having U.S. national cattle evaluations; and to evaluate variants identified from GPE in other populations also genotyped with the functional variant assay. Genotypes available for GPE included 1,109 sires, 506 dams, and 237 non-parents, and were imputed to 13,286 animals having genotypes from the moderate-density assays. Independent populations with functional genotypes included 2,500 animals from 4 selection populations at the U.S. Meat Animal Research Center, and 1,500 animals from two breeds in seedstock herds cooperating in a project to evaluate genomic predictions. Genome-wide association studies (GWAS) of birth weight were conducted in GPE using genomic BLUP (GBLUP) then solving variant effect estimates from the GBLUP animal solutions. A permutation step was implemented to determine if variant effect estimates might be systematically biased. Animal solutions were randomly reordered 10,000 times, and variant effects were solved for each permutation. The reordered vectors of animal effects were not correlated (mean r2<0.0001; max r2=0.002) but variant effects solved from the reordered vectors were correlated (mean r2=0.21; max r2=0.94), indicating a systematic influence which appears to be related to linkage disequilibrium (r2) among variants. Variants having at least moderate r2 (>0.25) with several other variants tended to have smaller effect estimates and permutation variances than variants having only weak r2 with others. Testing the estimated effect of each variant against its permutation variance revealed 293 variants with significantly (Bonferroni-corrected P<0.05) larger effects than expected from permutation. These variants explained 36% of phenotypic variation in GPE birth weights, and molecular breeding values trained using GPE effects had genetic correlations with birth weight in other populations ranging from 0.25 to 0.44. Similar correlations were obtained from an 11 variant panel identified by repeating permutation on two successively smaller sets. Genetic correlations between birth weight and genotypes for the single most significant variant in GPE, a non-synonymous SNP in NCAPG, were between 0.17 and 0.34 in the independent populations.