BG2

BG2 performs Bayesian genome-wide association studies (GWAS) analysis for non-Gaussian data using penalized quasi-likelihood with nonlocal priors to identify single nucleotide polymorphisms (SNPs).

Key Features:

• Non-Gaussian GWAS Analysis: BG2 handles non-Gaussian phenotypic distributions that violate standard linear model assumptions in traditional GWAS.
• Penalized Quasi-Likelihood: BG2 employs penalized quasi-likelihood estimation to fit generalized linear mixed models for non-normal data.
• Nonlocal Priors: BG2 incorporates nonlocal priors for Bayesian variable selection to identify associated SNPs with improved discriminatory power.
• Two-Step Identification: BG2 implements a two-step procedure to identify significant SNPs in genome-wide association analysis.

Scientific Applications:

• Genome-Wide Association Studies: BG2 identifies SNPs associated with non-Gaussian traits such as count data, binary outcomes, or skewed continuous phenotypes.

Methodology:

BG2 uses penalized quasi-likelihood along with nonlocal priors in a two-step manner to identify SNPs in GWAS analysis.