DIGEP-Pred 2.0 is a web-service for in silico prediction of drug-induced changes of gene expression profiles based on structural formula of drug-like compounds.

DIGEP-Pred 2.0 provides qualitative prediction of differentially expressed genes for a query structure. Two sources of drug-induced gene expression are used for the analysis of structure-activity relationships and predictions: (1) the literature-derived data on gene expression changes from Comparative Toxicogenomics Database (CTD) (https://ctdbase.org) at the levels of mRNA and protein. The training sets with data on mRNA and protein expression changes contain 2620 and 2671 structures, correspondingly; (2) the microarray experimental data from Connectivity Map (cMAP) build02 (https://clue.io/data/CMB02) obtained on three cell lines: MCF7, PC3 and HL60. The five training sets were created for each cell line based on different log fold change thresholds: 0.5, 0.7, 1, 1.5 and 2. The average accuracy of prediction was 86.5% for CTD mRNA-related models and 94.8% for CTD protein-related models. The average accuracy of prediction exceeds 87% for cMAP-related models (see “Training set” tab for more detailed information). SDF files and files with SMILES of the datasets used for the creation of SAR models for drug-induced gene expression are available for download.

DIGEP-Pred 2.0 allows performing enrichment analysis to identify KEGG (https://www.genome.jp/kegg) and Reactome (https://reactome.org) pathways, Gene Ontology (https://geneontology.org ) biological processes and diseases from DisGeNET (https://www.disgenet.org) associated with predicted gene expression changes.

DIGEP-Pred 2.0 can be also used to identify compound’s direct targets which are probable master-regulators responsible for observed gene expression changes. To do it, transcription factor enrichment analysis and “upstream” analysis in signaling network are performed subsequently. The potential master regulators identified by network analysis are intersected with direct targets of compound predicted by PASS using the corresponding information from ChEMBL (https://www.ebi.ac.uk/chembl) and PubChem (https://pubchem.ncbi.nlm.nih.gov) databases.

The previous version of DIGEP-Pred is available at https://www.way2drug.com/ge

Please cite us: Ivanov SM, Rudik AV, Lagunin AA, Filimonov DA, Poroikov VV. DIGEP-Pred 2.0: A web application for predicting drug-induced cell signaling and gene expression changes. Mol Inform. 2024 Jul 9:e202400032. doi: 10.1002/minf.202400032. Epub ahead of print. PMID: 38979651