ZJU-ENS Joint Laboratory of Medicinal Chemistry, College of Pharmaceutical Sciences, Zhejiang University, Hangzhou 310058, China.
J Mol Graph Model. 2012 Jul;37:77-86. doi: 10.1016/j.jmgm.2012.04.003. Epub 2012 Apr 23.
Pharmacophore models of cyclin-dependent kinase-1 (CDK1) inhibitors were established by using the Catalyst/HypoGen. The best pharmacophore model, Hypo1, consists of one hydrogen bond acceptor (HBA), one hydrogen bond donor (HBD), one hydrophobic (HY) and one ring aromatic (RA) feature. The validation results of Hypo1 through cost analysis, test set prediction, Fisher's cross method and receiver operating characteristic (ROC) study indicated that the Hypo1 was statistically valuable and reliable in identifying structural diverse CDK1 inhibitors. It is further supported by the consistent results from molecular docking studies. Finally, the Hypo1 was used to "in silico" screen the NCI and MayBridge database. The preferable hits obtained were further docked into ATP binding site of CDK1, and nine promising compounds were retrieved as novel potential CDK1 inhibitors for further studies.
通过 Catalyst/HypoGen 建立了细胞周期蛋白依赖性激酶-1（CDK1）抑制剂的药效团模型。最佳药效团模型 Hypo1 由一个氢键受体（HBA）、一个氢键供体（HBD）、一个疏水（HY）和一个环芳基（RA）特征组成。通过成本分析、测试集预测、Fisher 交叉法和接收器工作特性（ROC）研究对 Hypo1 的验证结果表明，Hypo1 在识别结构多样的 CDK1 抑制剂方面具有统计学意义上的价值和可靠性。分子对接研究的一致结果进一步支持了这一点。最后，使用 Hypo1 对 NCI 和 MayBridge 数据库进行“计算筛选”。获得的较好命中物进一步对接至 CDK1 的 ATP 结合位点，检索到 9 个有前途的化合物作为新型潜在 CDK1 抑制剂，供进一步研究。
