一种人工智能驱动的蛋白受体-小分子互作结构预测及筛选方法
By constructing a virtual drug library XH-GPR18-1.0 for the GPR18 receptor using a deep learning-based all-atomic-level protein-ligand complex structure prediction model, the problem of low drug screening efficiency in the prevention and treatment of mammary fibrosis in dairy cows was solved, and effective candidate drugs were screened out and their therapeutic effects were verified.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- XIANGHU LABORATORY
- Filing Date
- 2026-04-14
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies lack effective drugs for the prevention and treatment of mammary fibrosis in dairy cows, and existing virtual screening methods have insufficient structural prediction accuracy and low batch screening efficiency, making it difficult to discover effective drug candidate molecules.
A deep learning-based, all-atom-level protein-ligand complex structure prediction model was used to construct a virtual drug library, XH-GPR18-1.0, by batch screening of ligands for the GPR18 receptor. Candidate drugs were screened by combining binding energy, conformational stability, and pocket coverage indicators, and then validated by wet experiments.
It significantly improved the accuracy and efficiency of drug screening, identified a variety of effective candidate drugs, enriched the prevention and treatment strategies for breast fibrosis, improved the treatment success rate, and verified the actual efficacy of the drugs through wet experiments.
Smart Images
Figure CN122417136A_ABST