Cyclin-dependent kinase 2 (CDK2) is the family of Ser/Thr protein kinases that has emerged as a highly selective with low toxic cancer therapy target. A multistage virtual screening method combined by SVM, protein-ligand interaction fingerprints (PLIF) pharmacophore and docking was utilised for screening the CDK2 inhibitors. The evaluation of the validation set indicated that this method can be used to screen large chemical databases because it has a high hit-rate and enrichment factor (80.1% and 332.83 respectively). Six compounds were screened out from NCI, Enamine and Pubchem database. After molecular dynamics and binding free energy calculation, two compounds had great potential as novel CDK2 inhibitors and they also showed selective inhibition against CDK2 in the kinase activity assay.

中文翻译：

---

基于支持向量机，药效团和对接模型的多阶段虚拟筛选方法鉴定新型CDK2抑制剂。

细胞周期蛋白依赖性激酶2（CDK2）是Ser / Thr蛋白激酶家族，已以高选择性和低毒性癌症治疗靶点的形式出现。利用SVM，蛋白质-配体相互作用指纹（PLIF）药效团和对接相结合的多阶段虚拟筛选方法筛选CDK2抑制剂。对验证集的评估表明，该方法可用于筛选大型化学数据库，因为它具有很高的命中率和富集因子（分别为80.1％和332.83）。从NCI，Enamine和Pubchem数据库中筛选出六种化合物。经过分子动力学和结合自由能计算，两种化合物具有作为新型CDK2抑制剂的巨大潜力，并且在激酶活性测定中还显示出对CDK2的选择性抑制作用。