GSAn: an alternative to enrichment analysis for annotating gene sets

Area of Science:

• Genomics
• Bioinformatics
• Systems Biology

Background:

• Advancements in sequencing technologies enhance understanding of genotype-phenotype relationships.
• Statistical enrichment methods are standard for analyzing phenotype-grouped data but can overemphasize well-studied genes.
• Existing methods may limit the coverage of annotated genes within a gene set.

Purpose of the Study:

• To develop a novel gene set annotation method, GSAn, that leverages semantic similarity.
• To address limitations of traditional enrichment analyses by improving gene coverage and reducing term redundancy.
• To provide interactive visualization for multi-scale gene set annotation analysis.

Main Methods:

• Developed GSAn, a gene set annotation method utilizing semantic similarity measures.
• Employed Gene Ontology (GO) structure for pairwise gene comparisons.
• Focused on optimizing the trade-off between reducing annotation terms and increasing gene set coverage.
• Integrated interactive visualization tools for multi-scale analysis.

Main Results:

• GSAn effectively synthesizes Gene Ontology annotation terms using semantic similarity.
• The method achieves a superior balance between minimizing retained terms and maximizing related gene inclusion.
• GSAn demonstrates enhanced gene coverage compared to traditional enrichment analysis tools.
• Interactive visualization facilitates multi-scale analysis of gene set annotations.

Conclusions:

• GSAn offers a novel approach to gene set annotation, outperforming traditional enrichment methods.
• The method enhances biological insight by improving gene coverage and reducing term redundancy.
• GSAn provides valuable tools for comprehensive gene set analysis and interpretation.