blitzGSEA: efficient computation of gene set enrichment analysis through gamma distribution approximation

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Interpreting transcriptomics data relies on identifying biological pathways from differential gene expression.
• Gene Set Enrichment Analysis (GSEA) is a common method but is computationally intensive due to its reliance on permutation tests.

Purpose of the Study:

• To introduce blitzGSEA, a novel algorithm designed to enhance the speed and accuracy of pathway analysis in transcriptomics.
• To provide a computationally efficient alternative to traditional GSEA for analyzing differential gene expression signatures.

Main Methods:

• Implemented blitzGSEA, an algorithm utilizing the same running sum statistic as GSEA.
• Replaced permutation tests with approximations based on Gamma distributions to calculate enrichment score probabilities.

Main Results:

• Achieved significant performance improvements compared to existing GSEA implementations.
• Demonstrated more accurate approximation of small P-values.

Conclusions:

• blitzGSEA offers a substantially faster and more accurate approach for pathway identification from transcriptomics data.
• The algorithm provides a valuable tool for researchers analyzing large-scale gene expression datasets.