CFE 2020: Start Registration
View Submission - CMStatistics
Title: A novel Bayesian framework for the analysis of GWAS data Authors:  Jacob Williams - Virginia Tech (United States)
Marco Ferreira - Virginia Tech (United States) [presenting]
Tieming Ji - University of Missouri at Columbia (United States)
Abstract: A Bayesian analysis is proposed for genome-wide association studies (GWAS) that selects significant single nucleotide polymorphisms (SNP) in two stages. The first stage fits as many linear mixed models as the number of SNPs, with each model containing one SNP as well as random effects to take into account kinship correlation. The result of the first stage is a set of candidate significant genes. The second stage performs a stochastic search through model space with a genetic algorithm, where each model is a linear mixed model with principal components for population structure as well as kinship random effects and may include multiple SNPs from the candidate set obtained from the first stage. The result of the second stage is a list of models with their respective posterior probabilities. We illustrate our proposed Bayesian GWAS framework with an analysis of publicly available experimental data on root architecture remodeling of a model plant in response to salt stress.