EpiTracer - an algorithm for identifying epicenters in condition-specific biological networks.

Area of Science:

• Systems Biology
• Bioinformatics
• Network Biology

Background:

• Biological diseases arise from small disruptions in complex protein interaction networks.
• These perturbations often impact a few proteins, triggering wider network disturbances and cellular stress responses.
• Identifying key proteins in perturbation spread or response is crucial for understanding disease mechanisms.

Purpose of the Study:

• To develop an algorithm, EpiTracer, for identifying key proteins (epicenters) in condition-specific biological networks.
• To introduce a novel centrality measure, ripple centrality, for quantifying a node's influence across the protein-protein interaction (PPI) network.
• To demonstrate the algorithm's ability to find critical nodes beyond simple differential expression analysis.

Main Methods:

• Developed EpiTracer algorithm utilizing protein-protein interaction (PPI) networks and gene expression data.
• Introduced ripple centrality to measure how effectively a node's perturbation propagates through the network.
• Applied EpiTracer to an overexpression study (PARK2) and a knockdown study (SP1).

Main Results:

• EpiTracer successfully identified key epicenters in both overexpression and knockdown studies.
• In the PARK2 overexpression study, PARK2 was the top epicenter, with other identified nodes supporting or counteracting its activity.
• The method identified epicenters not detected by differential expression analysis, highlighting its unique capability.

Conclusions:

• EpiTracer effectively identifies condition-specific epicenters in biological networks using PPI and gene expression data.
• The algorithm provides tools for analyzing the influence zone of epicenters and summarizing dysregulated genes.
• EpiTracer demonstrates general applicability, robustness to minor network changes, and is publicly available.