SourceSet: A graphical model approach to identify primary genes in perturbed biological pathways

Area of Science:

• Genomics
• Systems Biology
• Bioinformatics

Background:

• Topological gene-set analysis is crucial for interpreting omics data.
• Existing methods often fail to differentiate between primary drivers and secondary responders of gene dysregulation.
• Distinguishing causal genes from reactive ones is essential for accurate biological network analysis.

Purpose of the Study:

• To introduce SourceSet, a novel method for distinguishing primary from secondary gene dysregulation.
• To apply SourceSet within a Gaussian graphical model framework.
• To enhance the interpretation of omics data by identifying the source of biological perturbations.

Main Methods:

• SourceSet compares gene expression profiles between control and perturbed conditions.
• It employs likelihood ratio tests to detect differences in mean and covariance parameters.
• Network propagation is used to infer primary and secondary gene sets.

Main Results:

• SourceSet demonstrates high specificity and sensitivity in simulations.
• The method performs effectively on real biological case studies.
• It successfully distinguishes genes responsible for primary dysregulation from those affected by it.

Conclusions:

• SourceSet provides a robust approach to differentiate primary and secondary gene dysregulation.
• The method enhances the understanding of molecular mechanisms underlying biological perturbations.
• The SourceSet R package facilitates pathway analysis and visualization, aiding biological discovery.