Computational Network Analysis for Drug Toxicity Prediction.

Area of Science:

• Toxicogenomics
• Computational Biology
• Network Pharmacology

Background:

• Compound safety assessment relies on predicting molecular effects.
• Adverse Outcome Pathways (AOPs) integrate mechanistic information for toxicity prediction.
• Transcriptome analysis reveals gene expression changes indicative of toxicity.

Purpose of the Study:

• To develop a computational approach for identifying network modules from transcriptomics data.
• To predict compound effects at the molecular interaction level.
• To aid in the construction of AOPs for hazard and risk assessment.

Main Methods:

• Biostatistical quantification of gene expression changes.
• Functional annotation and gene prioritization using literature mining.
• Construction of high-confidence interaction networks and identification of predictive modules.

Main Results:

• The developed computational approach successfully identifies network modules from transcriptomics data.
• Demonstrated performance using public data for drugs inducing hepatic and cardiac toxicity.
• The method integrates molecular interactions to pinpoint key events in compound toxicity.

Conclusions:

• The computational approach is effective for identifying molecular network modules related to compound toxicity.
• This method supports the construction of AOPs and enhances toxicogenomics studies.
• The approach provides a valuable tool for drug safety and chemical risk assessment.