fgsea

fgsea performs fast gene set enrichment analysis (GSEA) to compute pathway enrichment scores and accurately estimate extremely low P-values for pathway-level inference and multiple hypothesis correction in transcriptional data.

Key Features:

• Speed and efficiency: Uses an advanced algorithm that accelerates computation, enabling a larger number of permutations and finer-grained P-values.
• High accuracy in P-value estimation: Routinely estimates GSEA P-values as small as 10^-100 for precise significance assessment.
• Validation through exact algorithms: Performance validated against an exact algorithm for calculating GSEA P-values using integer gene-level statistics, showing small and predictable estimation error.
• Extensive evaluation: Systematically evaluated across 605 datasets, demonstrating superior recovery of statistically significant pathways compared to other GSEA implementations.

Scientific Applications:

• Preranked GSEA after differential expression: Applied to preranked GSEA workflows following differential gene expression tests to identify enriched pathways.
• Detection of subtle pathway enrichments: Enhances sensitivity to uncover biologically meaningful pathway enrichment signals that may be missed by less precise methods.

Methodology:

Computational steps include accelerated permutation-based GSEA to refine P-value calculations, estimation of extremely low P-values (down to 10^-100), validation against an exact algorithm using integer gene-level statistics, and handling of experimental designs comparing two conditions with multiple replicates.