ttgsea

ttgsea predicts enrichment scores for unique tokens or words derived from gene set names using deep learning to resolve redundancy from overlapping gene set terms in gene set enrichment analysis.

Key Features:

• Token-based enrichment prediction: Predicts enrichment scores for individual tokens or words extracted from text in names of gene sets rather than for predefined gene sets as a whole.
• Overlapping set resolution: Addresses redundancy within enriched terms caused by multiple terms existing for a single biological process and overlapping sets of related terms.
• Custom term generation: Enables coining of new terms by combining tokens and predicting enrichment scores for such combined tokens.

Scientific Applications:

• Gene expression data analysis: Infers results of gene expression data at the level of gene sets by calculating enrichment scores.
• Functional enrichment analysis: Provides an alternative approach to traditional gene set enrichment analysis (GSEA) when predefined gene sets have overlapping or redundant term annotations.

Methodology:

Uses deep learning to predict enrichment scores for unique tokens or words from text in names of gene sets.