XL-mHG

XL-mHG assesses gene set enrichment in ranked lists by providing a semiparametric generalization of the nonparametric minimum hypergeometric (mHG) test for 0/1-valued ranked data.

Key Features:

• Semiparametric generalization: Extends the mHG test with semiparametric elements to control the type of enrichment being tested for Boolean (0/1) ranked entries.
• Enhanced power: Demonstrates greater sensitivity than one-sided Kolmogorov–Smirnov (KS)-type tests for detecting gene set enrichment in biological contexts.
• Efficient algorithms: Implements a quadratic-time algorithm for exact XL-mHG p-values and a linear-time algorithm for computing tighter upper bounds on those p-values.

Scientific Applications:

• Gene set enrichment analysis: Applied to ranked lists from gene expression studies and other biological data sets to identify enriched gene sets associated with biological processes or conditions.
• Statistical benchmarking: Used as an alternative to KS-type tests when increased power and a semiparametric testing framework are required.

Methodology:

Builds on the mHG test via a semiparametric generalization; computes exact p-values using a quadratic-time algorithm and tighter upper bounds using a linear-time algorithm; validation reported via simulation studies.