Integrating data and knowledge to identify functional modules of genes: a multilayer approach

Area of Science:

• Systems biology
• Computational biology
• Genomics

Background:

• Identifying functional modules in cellular networks is crucial for discovering therapeutic targets.
• High-throughput technologies enable network characterization but face data quality challenges.
• Existing computational methods for module identification are limited by data quality issues.

Purpose of the Study:

• To improve the accuracy of functional module identification by integrating knowledge from scientific literature.
• To develop a novel computational model and algorithm for enhanced functional module discovery.

Main Methods:

• Developed a new algorithm integrating mRNA expression data and biomedical knowledge.
• Applied the model to yeast and human protein-protein interaction networks.
• Compared performance against methods using transcriptomic data, literature knowledge, or unweighted networks.

Main Results:

• The new algorithm achieved over 90% protein coverage in yeast and human interactomes.
• Integrating mRNA expression and biomedical knowledge significantly improved functional module identification performance.
• The algorithm demonstrated superior performance, particularly in positive predictive value (PPV), compared to existing methods.

Conclusions:

• The multiplex approach effectively integrates diverse data sources, reducing false positives and increasing confidence in novel discoveries.
• The algorithm's high protein coverage facilitates the generation of more reliable biological hypotheses.
• This method enhances the discovery of novel genes for targeted therapeutics.