DiffSLC

DiffSLC identifies essential proteins in protein–protein interaction (PPI) networks by integrating network topology and gene coexpression data into a differential centrality weighting to prioritize essential genes and proteins.

Key Features:

• Centrality measure: Combines traditional network centrality metrics with gene coexpression values to score node importance in PPI networks.
• Weighting factor: Uses the number of interactions (node degree) and gene coexpression values as a weighting factor to adjust centrality scores.
• Eigenvector centrality enhancement: Incorporates eigenvector centrality to elevate nodes with influential neighbors, aiding detection of potentially essential proteins with low degree.
• Comprehensive analysis: Assesses the impact of including or excluding gene coexpression data, different coexpression measures, and various gene expression datasets on essentiality prediction.

Scientific Applications:

• Essential protein detection in Saccharomyces cerevisiae: Applied to Saccharomyces cerevisiae PPI networks to identify essential genes and proteins.
• Benchmarking centrality measures: Serves to compare DiffSLC performance against traditional centrality measures in essentiality prediction.
• Evaluation of coexpression effects: Quantifies how inclusion of gene coexpression and choice of coexpression measures affect essential protein identification.
• Predictive performance assessment: Enables assessment of predictive accuracy using ROC analysis and area under the ROC curve (AUC).

Methodology:

Computes network centrality metrics including eigenvector centrality, weights nodes by number of interactions and gene coexpression values, compares inclusion/exclusion of coexpression and different coexpression measures and gene expression datasets, and evaluates predictive performance using ROC curves and AUC; implementations provided in R (igraph) and Python (networkx).