应用简化分子线性输入系统和基于分子图的描述符进行胰腺脂肪酶抑制剂的预测和设计。
Use of Simplified Molecular Input Line Entry System and molecular graph based descriptors in prediction and design of pancreatic lipase inhibitors.
机构信息
Department of Pharmaceutical Sciences, Guru Jambheshwar University of Science & Technology, Hisar, India.
出版信息
Future Med Chem. 2018 Jul;10(13):1603-1622. doi: 10.4155/fmc-2018-0024.
AIM
The inhibition of pancreatic lipase (PL) enzyme is the most explored strategy for the treatment of obesity. The present study describes the development of quantitative structure-activity relationship (QSAR) models for a diverse set of 293 PL inhibitors by means of the Monte Carlo optimization technique. Methodology & results: The hybrid optimal descriptors were used to build QSAR models with three subsets of three splits. The developed QSAR models were further validated with corresponding external sets. The best QSAR model has the following statistical particulars: R = 0.752, Q LOO 2 = 0 . 736 for the test set and R = 0.768, Q F 1 2 = 0 . 628 , Q F 2 2 = 0 . 621 for the validation set.
CONCLUSION
The developed QSAR models were robust, stable and predictive and led to the design of novel PL inhibitors.
目的
抑制胰脂肪酶（PL）是治疗肥胖症最常用的方法。本研究采用蒙特卡罗优化技术，描述了对 293 种不同 PL 抑制剂的定量构效关系（QSAR）模型的开发。
方法与结果
采用混合最优描述符，通过三个三分集建立 QSAR 模型。所开发的 QSAR 模型进一步用相应的外部集进行验证。最佳 QSAR 模型具有以下统计特征：用于测试集的 R=0.752，Q LOO 2 =0.736，用于验证集的 R=0.768，Q F 1 2 =0.628，Q F 2 2 =0.621。
结论
所开发的 QSAR 模型具有稳健性、稳定性和预测性，并可用于设计新型 PL 抑制剂。
Zotero 插件