Location: Reproduction ResearchTitle: Meta-analysis of genome-wide association from genomic prediction models Author
|Bernal Rubio, Yeni - Michigan State University|
|Gualdron Duarte, Jose - Universidad De Buenos Aires|
|Bates, Ronald - Michigan State University|
|Ernst, Catherine - Michigan State University|
|Nonneman, Danny - Dan|
|King, David - Andy|
|Cantet, Rodolfo - Universidad De Buenos Aires|
|Steibel, Juan - Michigan State University|
Submitted to: Animal Genetics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/11/2015
Publication Date: 1/14/2016
Publication URL: http://handle.nal.usda.gov/10113/61848
Citation: Bernal Rubio, Y.L., Gualdrón Duarte, J.L., Bates, R.O., Ernst, C.W., Nonneman, D., Rohrer, G.A., King, A., Shackelford, S.D., Wheeler, T.L., Cantet, R.J.C., Steibel, J.P. 2016. Meta-analysis of genome-wide association from genomic prediction models. Animal Genetics. 47(1):36-48.
Interpretive Summary: Genome-wide association studies (GWA) are a common practice in animal breeding; however, effect sizes of GWA tests are small, requiring larger sample sizes to enhance power of detection of rare variants. Because of difficulties of increasing sample size in animal populations, one alternative is to implement a meta-analysis (MA), combining information and results from independent GWA. Although this methodology has been used widely in human genetics, implementation in animal breeding is a recent approach. Thus, we propose a straightforward methodology for implementing meta-analysis of GWA, accounting for population structure and heterogeneity of variance components across populations. We present methods for implementation of meta-analysis of GWA derived from multiple genomic evaluations, accounting for strength of association as well as whether the SNP effect is positive or negative. Application to real datasets shows that MA produces similar results to joint analysis, but without involving data-sharing and simplifying modelling of heterogeneity across populations. Meta-analysis of GWA is an attractive alternative to summarize results from multiple genomic evaluations, avoiding problems present in joint analysis, such as data sharing, definition of fixed effects, heterogeneous variances, population structure and different scales of measurement of evaluated traits.
Technical Abstract: A limitation of many genome-wide association studies (GWA) in animal breeding is that there are many loci with small effect sizes; thus, larger sample sizes (N) are required to guarantee suitable power of detection. To increase sample size, results from different GWA can be combined in a meta-analysis (MA-GWA). The goal of this research was to describe methods for implementation of MA-GWA, combining results from multiple genomic evaluations, using two alternatives for weighting SNP effects. The methodology is exemplified with an application to real data, comparing results with those obtained from a joint-analysis. Measures of chop redness (CIE a*) from three populations were analyzed: 1) F2 generation (n = 928) of the Michigan State University Pig Resource Population (MSUPRP), 2) U. S. Meat Animal Research Center Swine Population (MARC, n = 1237), and 3) a commercial population (n = 2001). Animals were genotyped using the PorcineSNP60 BeadChip or the Neogen PorcineGGP-LD and imputed with high accuracy. In individual population GWA, associations were observed on chromosome 6 but none of them reached the genome-wide significance threshold. Joint-analysis detected one significant genomic region on chromosome 6, which can be explained by the high linkage disequilibrium observed on this chromosome in our populations. This association was confirmed by MA-GWA, which also identified a significant QTL on chromosome 5. Thus, MA-GWA increased power of detection, confirming and identifying associations for CIE a*. In contrast to joint-analysis, MA-GWA is an attractive alternative to summarize results from multiple genomic evaluations, avoiding problems of combining different datasets (definition of fixed effects, heterogeneous variances, population structure and different scales of measurement).