A Novel ILP Framework to Identify Compensatory Pathways in Genetic Interaction Networks with GIDEON

Area of Science:

• Systems Biology
• Genomics
• Computational Biology

Background:

• Baker's yeast has extensive pairwise epistasis data for non-essential genes.
• This data forms a genetic interaction network, representing gene relationships.
• Identifying compensatory pathways is crucial for understanding genetic robustness.

Purpose of the Study:

• Introduce GIDEON, an Integer Linear Programming (ILP) method, to find diverse Between-Pathway Models (BPMs).
• Improve the identification of gene sets indicative of compensatory pathways.
• Enhance the discovery of novel biological insights from genetic interaction data.

Main Methods:

• Utilized pairwise epistasis experiments from Baker's yeast.
• Developed GIDEON, an ILP-based approach with improved edge weighting.
• Applied GIDEON to a weighted signed graph of the genetic interaction network.

Main Results:

• GIDEON identified substantially larger collections of BPMs compared to previous methods.
• The discovered BPMs showed improved functional enrichment.
• Identified novel gene sets, including those linking ergosterol and aromatic amino acid biosynthesis.

Conclusions:

• GIDEON is an effective method for discovering biologically relevant gene sets from genetic interaction networks.
• The findings suggest potential new targets for antifungal drugs by revealing connections in metabolic pathways.
• This work advances the analysis of complex genetic interactions for biological discovery.