Integrating single cell expression quantitative trait loci summary statistics to understand complex trait risk genes.

Area of Science:

• Genomics
• Immunology
• Computational Biology

Background:

• Transcriptome-wide association studies (TWAS) traditionally use bulk tissues, limiting cell-type specific gene discovery.
• Emerging single-cell expression quantitative trait loci (sc-eQTL) data offer higher resolution but are underutilized in TWAS.
• Summary statistics from large eQTL datasets are not broadly leveraged for TWAS.

Purpose of the Study:

• To develop a novel method, EXPRESSO, for analyzing sc-eQTL summary statistics.
• To integrate 3D genomics and epigenomic data for prioritizing causal variants.
• To identify novel gene-trait associations and potential drug repurposing candidates for autoimmune diseases.

Main Methods:

• Developed EXPRESSO (EXpression PREdiction with Summary Statistics Only) for sc-eQTL summary statistics analysis.
• Integrated 3D genomic data and epigenomic annotations to prioritize causal variants.
• Applied EXPRESSO to multi-ancestry GWAS data for 14 autoimmune diseases and developed a cell-type aware drug repurposing pipeline.

Main Results:

• EXPRESSO significantly improves upon existing TWAS methods.
• Identified 958 novel gene x trait associations for autoimmune diseases, a 26% increase over the next best method.
• Discovered 492 cell-type specific associations missed by whole-blood TWAS.

Conclusions:

• EXPRESSO enables robust gene-trait association discovery using sc-eQTL summary statistics.
• The method highlights the importance of cell-type resolution in understanding disease genetics.
• Identified potential drug repurposing candidates, including metformin for type 1 diabetes and vitamin K for ulcerative colitis.