目的 采用计算机辅助药物设计的方法发现潜在的PTP1B抑制剂。方法 应用3D-QSAR药效团模型中的Hypogen模块构建药效团模型，成本分析、测试集预测和Fisher检验3种方法来验证该模型可用于预测化合物的生物活性的能力。运用该药效团模型对ZINC数据库进行虚拟筛选，得到Fit value值较高的先导化合物ZINC35671983。根据药效团的特征对ZINC35671983进行结构改造得到相应化合物。将化合物用ADMET进行成药预测。结果 ZINC35671983进行结构改造筛选获得92个化合物，筛出对接得分高于ZINC3567198的8个化合物。结论 发现8个潜在的PTP1B抑制剂，这有助于发现新的PTP1B先导化合物。

Objective To identify Potential PTP1B inhibitors by computer-aided drug design. Methods The Hypogen module of 3D-QSAR pharmacophore model was used to construct the pharmacophore model, cost analysis, test set prediction and Fisher test to verify the ability of the model to predict the biological activity of the compound. The lead compound ZINC35671983 with high Fit value was obtained by virtual screening of ZINC database with the pharmacophore model. The corresponding compounds were obtained by structural modification of ZINC35671983 according to the characteristics of pharmacophore. The compounds were predicted by ADMET. Results ZINC35671983 obtained 92 compounds through structural modification screening, and screened out 8 compounds with higher docking scores than ZINC3567198. Conclusion A total of 8 potential PTP1B inhibitors were found which is helpful to find new PTP1B lead compounds.

R96

天津市医学重点学科（专科）建设资助项目（TJYXZDXK-032A）