aceHarnessing machine learning, docking and molecular dynamics for the virtual screening of compounds as CDK4/6 dual inhibitors
Abstract: Cyclin-dependent kinases 4 and 6 (CDK4/6) are pivotal regulators of the G1-to-S phase transition, and their dysregulation is a hallmark of numerous malignancies. Despite the clinical success of existing CDK4/6 inhibitors, there remains a persistent need for chemically diverse scaffolds with potent dual-target affinity. In this study, we developed and implemented a virtual screening workflow that synergistically integrates ligand-based machine learning with structure-based molecular docking. By benchmarking multi…
Identification of AtHsp90.6 involved in early embryogenesis and its structure prediction by molecular dynamics simulations by Luo, An; Li, Xinbo; Zhang, Xuecheng; Zhan, Huadong; Du, Hewei; Zhang, Yubo; Peng, Xiongbo. CC BY 4.0 · https://creativecommons.org/licenses/by/4.0
Researchers built a virtual screening funnel that combined a Bayesian Ridge regressor trained on ChEMBL CDK4/6 inhibitor data with structure-based docking and molecular dynamics. The ML model using ECFP4 fingerprints achieved cross-validated R2 around 0.73 for CDK4 and 0.72 for CDK6 and was used to prioritize a 22,823-compound library down to three hits.
Biochemical testing by the August 2026 publication date showed one hit, HY-18,623, with IC50s of 3.5 nM for CDK4 and 17.4 nM for CDK6, stabilized by hydrogen bonds to Val96 and Val101. The work demonstrates computational efficiency for kinase inhibitor discovery, but remains at the lead-candidate stage without clinical efficacy, safety, or patient outcome data.
- Benchmarked ML models on ChEMBL datasets of 265 CDK4 and 402 CDK6 inhibitors; Bayesian Ridge with ECFP4 fingerprints was most predictive.
- ML filter prioritized a 22,823-compound library followed by dual-target docking refinement to select three hits for biochemical testing.
- 200-ns molecular dynamics and binding free energy analyses showed persistent hydrogen bonds with hinge residues Val96 (CDK4) and Val101 (CDK6).
An integrated ligand-based machine learning and structure-based docking funnel screened 22,823 compounds and identified HY-18,623 as a potent dual CDK4/6 inhibitor with low-nanomolar IC50s.
The rundown
The workflow combined curated ChEMBL data for CDK4 and CDK6, ECFP4 fingerprinting, and cross-validated regression to create an ML filter, then applied rigorous dual-target docking to narrow 22,823 compounds to three candidates for testing.
Biochemical evaluation confirmed HY-18,623 potency, and 200-ns simulations characterized its binding mode via hinge interactions, supporting its status as a lead for further oncology development rather than a clinically deployed therapy.
Sources
- Peer-reviewedMolecular Diversity2026-08-29
How should this claim be treated?
The debate