CoVar: A generalizable machine learning approach to identify the coordinated regulators driving variational gene...

Area of Science:

• Genomics
• Systems Biology
• Bioinformatics

Background:

• Network inference models biological interactions to identify disease pathways.
• Machine learning (ML) offers advanced predictive capabilities for genomic data analysis.
• ML-based models are emerging as alternatives to statistical models for complex disease factor identification.

Approach:

• CoVar leverages ML-based network inference to capture regulatory interaction strengths.
• Identifies 'variational' genes whose expression variability highlights network differences between states.
• Constructs subnetworks of variational and strongly connected genes to pinpoint core drivers.

Key Points:

• CoVar pinpoints central genes driving perturbed gene expression across biological states.
• Variational genes capture expression variability by being differentially expressed or having such neighbors.
• Identifies core genes influencing variational activity within subnetworks.

Conclusions:

• CoVar successfully identifies key genes in yeast expression data perturbed by mitochondrial genome deletion.
• These key genes were not detected by independent differential expression analysis.
• Highlights the utility of ML-based network inference for uncovering novel biological insights.