CoMM-Ssup4 sup

CoMM-Ssup4 sup integrates GWAS summary statistics and eQTL summary-level data to assess expression–trait associations and infer regulatory mechanisms by which non-coding genetic variants influence complex traits.

Key Features:

• Integration of GWAS and eQTL data: Integrates summary-level eQTL data with GWAS summary statistics to evaluate expression–trait associations without requiring individual-level eQTL genotypes.
• Probabilistic modeling: Implements an efficient probabilistic model using a variational Bayesian Expectation–Maximization (EM) algorithm for estimation and a likelihood ratio test for association significance.
• Regulatory focus: Targets regulatory effects of non-coding variants on gene expression to link genetic variation to complex trait biology.
• Performance comparison: Demonstrates comparable performance to CoMM-S2 and S-PrediXcan on simulated and real-world datasets.

Scientific Applications:

• Complex disease locus discovery: Applied to identify novel susceptibility loci for complex diseases such as cardiovascular diseases and osteoporosis.
• Cross-dataset integrative analysis: Analyzes GWAS summary statistics from Biobank Japan alongside eQTL summary statistics from eQTLGen and GTEx to discover regulatory variants affecting traits.

Methodology:

Constructs a probabilistic model for summary-level data and uses a variational Bayesian EM algorithm for parameter estimation with a likelihood ratio test for association testing.