wHC

wHC enhances gene set–based signal detection by applying tail-focused goodness-of-fit (GOF) statistics to compare distributions of gene-level test results against theoretical nulls and identify genes that deviate from the null.

Key Features:

• Tail-focused GOF statistics: Implements tail-focused goodness-of-fit statistics to prioritize extreme deviations in gene-level test results.
• Gene-set signal aggregation: Accumulates signals across individual genes within predefined gene sets to detect set-level departures from the null hypothesis.
• Null-versus-alternative comparison: Compares observed distributions of gene-level test results against theoretical null distributions using GOF tests.
• Power determinants: Accounts for the influence of the proportion of non-null genes and the separation between null and alternative distributions on test power.
• R implementation: Implemented as an R package.
• Statistic selection guidance: Provides guidance for selecting appropriate statistics based on study-specific parameters such as non-null proportion and distribution separation.

Scientific Applications:

• Gene set association detection: Detecting associations between traits and predefined gene sets by aggregating gene-level signals.
• Whole exome sequencing analysis: Applied to whole exome sequencing (WES) data, exemplified by an amyotrophic lateral sclerosis (ALS) study.
• Method selection optimization: Optimizing the choice of tail-focused statistics to maximize detection power under varying proportions of non-null genes and effect-size separations.

Methodology:

Uses tail-focused goodness-of-fit tests to compare observed distributions of gene-level test results to theoretical null distributions, aggregates signals across genes within sets, and evaluates power dependence on the proportion of non-null genes and the separation between null and alternative distributions; implemented in R.