oCEM: Automatic detection and analysis of overlapping co-expressed gene modules

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Gene module detection faces challenges like overlapping modules and localized co-expression.
• Unsupervised clustering methods are standard but require separate, complex selections for sample and gene clustering.

Purpose of the Study:

• To present an R-package, Overlapping CoExpressed gene Module (oCEM), designed to overcome common challenges in gene module detection.
• To provide a user-friendly tool for complex statistical analyses in gene co-expression studies.

Main Methods:

• Utilized decomposition methods within the oCEM R-package.
• Developed a novel permutation-based statistical approach for optimal principal component selection.

Main Results:

• oCEM effectively addresses module overlap and local co-expression issues.
• The novel statistical approach aids in selecting the optimal number of principal components.
• oCEM demonstrated superior performance in detecting biologically relevant gene modules compared to state-of-the-art techniques.

Conclusions:

• oCEM empowers non-technical users to conduct sophisticated statistical analyses.
• The package yields robust results for gene co-expression module detection.
• oCEM, its applications, and example data are available open-source.