COMPADRE: an R and web resource for pathway activity analysis by component decompositions.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Biological network analysis is crucial for understanding functional genomic data.
• Existing tools for pathway and gene set analysis have limitations in accuracy and scope.

Purpose of the Study:

• To introduce Compadre, a novel computational tool for biological network analysis.
• To enhance the estimation of pathway/gene set activity indexes and detection of altered gene sets.
• To provide a more comprehensive biological picture through integrated decomposition and over-representation analyses.

Main Methods:

• Compadre utilizes sub-matrix decomposition methods (PCA, Isomaps, ICA, NMF) to estimate gene set activity.
• It incorporates statistical tests to detect differences between sample groups, with and without considering differentially expressed genes.
• An integrated over-representation test is performed during the analysis.

Main Results:

• Compadre demonstrates superior performance in pathway detection compared to tools like David, Babelomics, and Webgestalt.
• It exhibits a lower false positive rate than PLAGE.
• The tool provides combined results from decomposition and over-representation analyses for a more complete biological interpretation.

Conclusions:

• Compadre offers a versatile, simple, and effective approach for biological network and gene set activity analysis.
• Its ability to detect more pathways with fewer false positives makes it a valuable tool for genomic data interpretation.
• The integrated analysis approach provides deeper biological insights than traditional methods.