DIGEP-Pred 2.0: A web application for predicting drug-induced cell signaling and gene expression changes

Area of Science:

• Computational chemistry
• Pharmacogenomics
• Bioinformatics

Background:

• Drug-induced gene expression profiles (DIGEP) are crucial for assessing drug efficacy, toxicity, and mechanisms.
• Experimental DIGEP data is often unavailable for many compounds, hindering drug discovery and development.
• Predictive models are needed to bridge this data gap.

Purpose of the Study:

• To develop and present DIGEP-Pred 2.0, a web application for predicting DIGEP and drug targets using only a compound's structural formula.
• To integrate structure-activity relationships (SARs) and network analysis for accurate predictions.
• To provide insights into potential therapeutic and adverse effects of drug-like compounds.

Main Methods:

• Utilized PASS technology for SAR model creation with data from CTD, CMap, PubChem, and ChEMBL.
• Employed network analysis to identify potential drug targets and master regulators.
• Developed DIGEP prediction models for various cell lines (MCF7, PC3, HL60) using CMap data.
• Integrated pathway, Gene Ontology, and disease enrichment analyses.

Main Results:

• Achieved high prediction accuracy: 86.5% for genes, 94.8% for proteins (CTD data), and 97.9% for MoAs.
• Developed SAR models with 87.5% mean accuracy for CMap data across different fold-change thresholds.
• DIGEP-Pred 2.0 provides predicted gene expression changes, drug targets, and enriched biological information.
• The web application is freely accessible at https://www.way2drug.com/digep-pred.

Conclusions:

• DIGEP-Pred 2.0 effectively predicts drug-induced gene expression profiles and potential drug targets from chemical structures.
• The tool facilitates the exploration of drug mechanisms and effects, especially in the absence of experimental data.
• DIGEP-Pred 2.0 serves as a valuable resource for drug discovery and toxicological assessments.