Deep learning integrated network pharmacology approach for deciphering the molecular mechanisms of Indian medicinal plant compounds against irritable bowel syndrome

Irritable bowel syndrome (IBS) is a chronic gastrointestinal disorder with limited effective treatments. This study investigates the therapeutic potential of bioactive compounds from Indian medicinal plants by prioritizing those with strong binding affinity to IBS-associated molecular targets. From an initial set of 790 phytochemicals, 270 drug-like compounds were selected via ADMET filtering. Target analysis identified 69 overlapping molecular targets between IBS and the compounds. Protein–protein interaction networks and enrichment analysis revealed key IBS-related targets including EGFR, AKT1, SRC, and JUN. Molecular docking identified quercetin, Butein, and Camptothecin as promising candidates, with binding affinities up to –8.0 kcal/mol. Deep learning models—DeepDTA and DeepDGC—trained on KIBA and DAVIS datasets further validated these interactions. Predicted binding scores for top compounds reached as high as 11.99, confirming strong binding potential. Notably, quercetin and Tylophorinidine consistently demonstrated high affinity for EGFR and JUN, respectively. By integrating network pharmacology, molecular docking, and AI-based prediction, this study offers a robust pipeline for identifying multi-target phytochemicals for IBS therapy. The findings highlight strong candidate compounds for further in vitro and in vivo evaluation.