The abstract is read at the publisher; the summary is JClub's.
A population-calibrated Variant Priority Score (VPS) framework in a South Asian whole-exome cohort (960 individuals) yields trait-driven capacity scores, not individual-level differences, with ClinVar AUC 0.74.
- Why it matters: Exercise genomics often overstates individual athletic prediction because total genotype scoring is biased by panel composition and struggles with unannotated variants. A trait-level, evidence-traceable approach is needed to avoid inflated claims.
- What they did: The method integrates six evidence streams—deep learning protein/genomic language models (ESM-2, Evo 2), missense/splice predictors (AlphaMissense, SpliceAI), and regional pathogenic-density metrics—to compute a 1–10 VPS, renormalizing missing evidence instead of zero-filling. It aggregates normalized trait burden across 344 genes and 17 physiological traits, calibrating expectations using gnomAD South Asian and an in-house reference (N > 13,000).
- The result: After normalization, expected trait burden no longer depended on panel size, and gene count explained 2.7% of variance in capacity scores; trait variation dominated (45.5%) over individual variation (11.1%, p = 0.386). ClinVar benchmarking gave AUC 0.74 (dropping to 0.62 without archive-dependent streams), and switching to gnomAD South Asian shifted scores (median fold change 0.73) while preserving trait ordering (Spearman 0.96).
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.