Gene-SCOUT

Gene-SCOUT identifies genes with similar continuous-trait fingerprints from phenome-wide association studies (PheWAS) to elucidate gene-phenotype relationships across multiple rare-variant genetic architectures.

Key Features:

• PheWAS continuous-trait fingerprints: Analyzes continuous trait fingerprints derived from phenome-wide association studies (PheWAS).
• Densely-phenotyped cohorts: Leverages clinical and continuous trait data from densely-phenotyped cohorts such as the UK Biobank (UKB).
• Gene-level rare variant collapsing: Performs gene-level rare variant collapsing analysis across over 1,500 continuous traits.
• Exome-scale data: Operates on exome sequencing data from 394,692 UKB participants.
• Metabolomic associations: Incorporates metabolomic trait associations from Nightingale Health based on 121,394 participants.
• Similarity metric: Calculates gene–gene similarity using the cosine similarity measure to capture concordant effect directionality in high-dimensional trait space.
• Gene clustering: Identifies clusters of genes that share similar phenotypic association profiles.
• Enrichment analyses: Evaluates Gene Ontology (GO) enrichment and UKB clinical trait enrichment statistics to assess biological relevance.
• Performance: Produces gene similarity estimates that show stronger enrichments for clinical traits compared to existing methods.

Scientific Applications:

• Gene identification: Identify genes with similar continuous-trait fingerprints for follow-up functional studies.
• Gene-phenotype relationship analysis: Characterize gene-phenotype relationships across multiple rare-variant genetic architectures.
• Gene clustering and pathway inference: Elucidate gene clusters with shared phenotypic associations to inform biological interpretation.
• Metabolomic and clinical integration: Integrate metabolomic trait associations with exome-based PheWAS results to connect molecular biomarkers to gene effects.
• Enrichment benchmarking: Compare and assess enrichment for clinical traits relative to existing methods.

Methodology:

Performs gene-level rare variant collapsing analysis across >1,500 continuous traits using exome data from 394,692 UKB participants; incorporates metabolomic trait associations from Nightingale Health (121,394 participants); computes gene similarities from continuous-trait fingerprints using cosine similarity to capture concordant effect directionality in high-dimensional space; conducts Gene Ontology (GO) enrichment and UKB clinical trait enrichment statistic analyses.