PhenomeXcan

PhenomeXcan integrates GWAS summary statistics with GTEx v8 transcriptomic data using PrediXcan and fastENLOC to generate gene-based associations that link genetic variants to complex traits for translational research and precision medicine.

Key Features:

• Data integration: Integrates genome-wide association study (GWAS) summary statistics from multiple publicly available sources with Genotype-Tissue Expression (GTEx) version 8 using PrediXcan.
• Dataset scope: Synthesizes information from 8.87 million variants associated with 4,091 traits across transcriptome regulation data from 49 GTEx v8 tissues covering 22,255 genes.
• Bayesian colocalization: Applies the novel Bayesian colocalization method fastENLOC to prioritize likely causal gene–trait associations.
• Significant associations: Identified 65,603 significant associations using a Bonferroni-corrected p-value threshold of 5.5 × 10^−10, of which 19,579 (29.8 percent) were colocalized with a locus regional colocalization probability greater than 0.1.
• Replication benchmarks: Replicates associations from the PheWAS Catalog and OMIM with area under the curve (AUC) values of 0.61 and 0.64, respectively.
• ClinVar integration: Integrates with ClinVar to enable analysis of phenome-to-phenome relationships between common and rare diseases.
• Gene-based resource: Produces a comprehensive gene-based resource that provides transcriptome-informed context for variant-level and gene-level associations.

Scientific Applications:

• Novel association discovery: Discovery of novel genome-to-phenome associations by mapping variant effects through predicted gene expression.
• Gene-trait cluster analysis: Exploration of complex gene–trait clusters across the genome and phenome.
• Phenome-to-phenome investigation: Investigation of relationships between common and rare diseases via integration with ClinVar.
• Therapeutic target evaluation: Prioritization and evaluation of potential therapeutic targets based on transcriptome-informed gene–trait associations.
• Validation of known associations: Replication and validation of established associations from resources such as the PheWAS Catalog and OMIM.

Methodology:

Integrates GWAS summary statistics with GTEx v8 using PrediXcan, applies Bayesian colocalization with fastENLOC, analyzes 8.87 million variants across 4,091 traits and 49 GTEx v8 tissues to produce associations for 22,255 genes, and assesses significance with a Bonferroni-corrected p-value of 5.5 × 10^−10 and locus regional colocalization probability (RCP) thresholds.