A multi-dimensional computational screening strategy for rapid identification of active components from traditional...

This study aims to develop and validate a Multi-Dimensional Computational Screening (MDCS) strategy to identify bioactive components by jointly considering target affinity, exposure potential, and safety, using Liangdi decoction (LDD) as a case study.

Key therapeutic targets for endometriosis were identified by analyzing Gene Expression Omnibus and MalaCards databases. The components in LDD were identified using HPLC-MS/MS combined with feature-based molecular networking. An MDCS strategy integrating molecular docking, VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) multi-criteria decision analysis, and machine learning was developed. Molecular docking was conducted to estimate affinity, and VIKOR multi-criteria decision analysis was applied to generate rankings by simultaneously considering docking performance, content, intestinal absorption, and hepatorenal toxicity. Machine learning was used to identify physicochemical features associated with target-specific binding, providing interpretable rules for component prioritization. Molecular dynamics simulations and surface plasmon resonance experiments were performed to validate.

Estrogen receptor, progesterone receptor, and gonadotropin-releasing hormone receptor were identified as major targets relevant to endometriosis. Among 131 characterized components, kukoamine A, ophiopogonanone E, and verbascoside were prioritized as optimal candidates targeting the three receptors, respectively. Machine learning revealed QED, Fsp³, and logD as critical determinants of binding affinity.

This study establishes a structured computational framework for discovering bioactive TCM constituents by integrating efficacy, druggability, and safety considerations, addressing key limitations of docking-based screening and providing a transferable strategy for target-directed herbal research.