Large-Scale Prospective Virtual Screening Using a Co-Folding-Rescoring Workflow for NPM1-TYK2 Inhibitor Discovery
4 affiliations
- 1School of Pharmacy, Zhejiang University, Hangzhou 310058, China
- 2Department of Computer Science and Engineering, The Chinese University of Hong Kong, Shatin, Hong Kong
- 3Zhongguancun Academy, Beijing 100094, China
- 4Zhejiang Provincial Key Laboratory for Intelligent Drug Discovery and Development, Jinhua Institute of Zhejiang University, Jinhua 321099, China
Abstract
Conventional large-scale structure-based virtual screening typically relies on fixed-receptor docking, in which ligand poses are sampled against a predefined receptor conformation. Protein-ligand co-folding models can instead generate candidate complex conformations by jointly modeling ligand placement and local receptor-ligand structural adaptation. Although recent benchmark studies have demonstrated the promise of AlphaFold3-like approaches for retrospective virtual screening, their practical application to target-specific, large-scale prospective screening with experimental validation remains limited. Here, we developed and prospectively applied a co-folding-rescoring workflow for NPM1-TYK2 inhibitor discovery. Using an in-house set of 4608 compounds experimentally annotated in Ba/F3-NPM1-TYK2 cells, we evaluated 13 pipelines combining AlphaFold 3, Protenix, Glide, or AutoDock Vina pose generation with scoring functions or confidence metrics. AlphaFold 3 followed by AutoDock Vina rescoring achieved the numerically highest BEDROC value, closely comparable to Protenix + Vina, and was selected for screening a commercial library of more than 300,000 compounds. AF3-generated complex conformations were subsequently rescored with Vina and prioritized by physicochemical assessment, clustering, and binding-mode inspection. Of 30 compounds tested experimentally, FV13 was identified as a cell-active primary hit, with a TYK2 kinase IC50 of 21.70 μM and a Ba/F3-NPM1-TYK2 cellular IC50 of 17.07 μM. Subsequent analog exploration yielded TX14, exhibiting an approximately 9.4-fold improvement in biochemical TYK2 inhibitory potency (IC50 = 2.31 μM). Additional selectivity profiling and Western blot analyses further characterized these compounds and indicated attenuation of downstream STAT5 phosphorylation at higher concentrations. Together, these findings demonstrate the feasibility of integrating co-folding, empirical rescoring, and target-specific experimental data for large-scale prospective hit discovery.
Related
- Discovery of a molecular glue inhibitor that stabilises a non-productive Kalirin-Rac1 complex
- An inspectable Grover algorithm-based amplitude-amplification workflow for k-mer discovery and motif scanning with genomic coordinate traceability
- Characterization of the N7 RNA cap methyltransferase from Trichomonas vaginalis and inhibitor discovery
- Accelerated discovery and optimization of a potent andselective MMP12 inhibitor with the machine learningplatform CAMOLE-1
- A Proximity-Driven Functional Screening Platform for the Discovery of Noninhibitory USP7 Ligand Enabling Targeted Protein Stabilization
BibTeX
@article{tian2026largescale,
title = {{Large-Scale Prospective Virtual Screening Using a Co-Folding-Rescoring Workflow for NPM1-TYK2 Inhibitor Discovery}},
author = {Weixin Tian and Xujun Zhang and Odin Zhang and Rubi Hu and Yuanyi Ye and Shicheng Chen and Jiale Zheng and Hui Zhang and Xin Chai and Wanglin Qu and Huating Wang and Dan Li and Yu Kang and Tingjun Hou and Peichen Pan},
journal = {International Journal of Molecular Sciences},
year = {2026},
month = nov,
doi = {10.3390/ijms27219864},
url = {https://www.mdpi.com/1422-0067/27/21/9864},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Large-Scale Prospective Virtual Screening Using a Co-Folding-Rescoring Workflow for NPM1-TYK2 Inhibitor Discovery
AU - Weixin Tian
AU - Xujun Zhang
AU - Odin Zhang
AU - Rubi Hu
AU - Yuanyi Ye
AU - Shicheng Chen
AU - Jiale Zheng
AU - Hui Zhang
AU - Xin Chai
AU - Wanglin Qu
AU - Huating Wang
AU - Dan Li
AU - Yu Kang
AU - Tingjun Hou
AU - Peichen Pan
PY - 2026
DA - 2026/11/03
JO - International Journal of Molecular Sciences
DO - 10.3390/ijms27219864
UR - https://www.mdpi.com/1422-0067/27/21/9864
AB - Conventional large-scale structure-based virtual screening typically relies on fixed-receptor docking, in which ligand poses are sampled against a predefined receptor conformation. Protein-ligand co-folding models can instead generate candidate complex conformations by jointly modeling ligand placement and local receptor-ligand structural adaptation. Although recent benchmark studies have demonstrated the promise of AlphaFold3-like approaches for retrospective virtual screening, their practical application to target-specific, large-scale prospective screening with experimental validation remains limited. Here, we developed and prospectively applied a co-folding-rescoring workflow for NPM1-TYK2 inhibitor discovery. Using an in-house set of 4608 compounds experimentally annotated in Ba/F3-NPM1-TYK2 cells, we evaluated 13 pipelines combining AlphaFold 3, Protenix, Glide, or AutoDock Vina pose generation with scoring functions or confidence metrics. AlphaFold 3 followed by AutoDock Vina rescoring achieved the numerically highest BEDROC value, closely comparable to Protenix + Vina, and was selected for screening a commercial library of more than 300,000 compounds. AF3-generated complex conformations were subsequently rescored with Vina and prioritized by physicochemical assessment, clustering, and binding-mode inspection. Of 30 compounds tested experimentally, FV13 was identified as a cell-active primary hit, with a TYK2 kinase IC50 of 21.70 μM and a Ba/F3-NPM1-TYK2 cellular IC50 of 17.07 μM. Subsequent analog exploration yielded TX14, exhibiting an approximately 9.4-fold improvement in biochemical TYK2 inhibitory potency (IC50 = 2.31 μM). Additional selectivity profiling and Western blot analyses further characterized these compounds and indicated attenuation of downstream STAT5 phosphorylation at higher concentrations. Together, these findings demonstrate the feasibility of integrating co-folding, empirical rescoring, and target-specific experimental data for large-scale prospective hit discovery.
SN - 1422-0067
ER -