Comparative analysis of gene sets in the Gene Ontology space under the multiple hypothesis testing framework

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Gene Ontology (GO) is crucial for interpreting high-throughput functional genomics data.
• Comparative analysis of gene lists requires identifying enriched or depleted GO terms.
• Standard statistical tests face challenges with correlated multiple testing of GO terms.

Purpose of the Study:

• To develop a robust procedure for identifying significant GO terms in correlated multiple testing scenarios.
• To address the limitations of individual p-values and independence assumptions in GO term analysis.
• To provide a practical tool for researchers analyzing gene lists from functional genomics studies.

Main Methods:

• Implementation of a False Discovery Rate (FDR) based procedure.
• Calculation of a conserved estimator of q-values for GO terms.
• Identification of significant GO terms based on desired q-value thresholds.

Main Results:

• A novel FDR procedure effectively handles correlated multiple testing for GO terms.
• The procedure provides reliable q-values for GO term significance.
• The GoSurfer software tool implements this FDR procedure.

Conclusions:

• The developed FDR procedure offers a reliable method for GO term enrichment analysis.
• GoSurfer provides a user-friendly graphical interface for applying this method.
• This approach enhances the interpretation of functional genomics study results.