Efficient Integrative Multi-SNP Association Analysis via Deterministic Approximation of Posteriors

Area of Science:

• Genomics
• Statistical Genetics
• Bioinformatics

Background:

• Functional genomic data is increasingly used in genetic association studies.
• Current methods for integrating annotations often lack rigor and computational efficiency.
• Maximizing the utility of functional annotations requires improved analytical approaches.

Purpose of the Study:

• To propose a rigorous inference procedure for integrative association analysis using genomic annotations.
• To develop an efficient algorithm (Deterministic Approximation of Posteriors - DAP) for joint enrichment analysis and causal variant identification.
• To enhance the analysis of both Genome-Wide Association Studies (GWAS) and molecular Quantitative Trait Loci (QTL) mapping.

Main Methods:

• Developed the Deterministic Approximation of Posteriors (DAP) algorithm.
• Applied DAP to perform joint enrichment analysis and identify multiple causal variants.
• Validated the approach using simulation studies.
• Analyzed cross-population eQTL data (GEUVADIS) and multi-tissue eQTL data (GTEx).

Main Results:

• The DAP algorithm demonstrates high computational efficiency and accuracy.
• Genetic variants impacting transcription factor binding sites are significantly enriched in cis-eQTLs across all analyzed tissues.
• Enrichment estimates across tissues correlate with the cell types of the annotations.

Conclusions:

• The proposed DAP method offers a rigorous and computationally efficient approach for integrative genomic annotation analysis.
• The findings highlight the importance of transcription factor binding site disruption in eQTLs.
• The study provides a framework for leveraging functional annotations to better understand genetic associations in diverse biological contexts.