Quantitative Structure Activity Relationship (QSAR) Approach to Multiple Drug Resistance (MDR) Modulators Based on Combined Hybrid System
Jianhua Wu, Xin Yan Li, Weidong Cheng, Qinjian Xie, Yin-Qian Liu, Chunyan Zhao · QSAR & Combinatorial Science · 2009
Abstract The aim of this paper was to build hybrid QSAR models for predicting multiple drug resistance (MDR) modulators modulating activity of 70 compounds based on single multiple linear regression (MLR) models and support vector machine (SVM) models. All models were validated using more strict criteria to make sure that the results are reliable and robust. For the best hybrid model, it gave the best performance, with corresponding correlation coefficientsR2of 0.85, 0.81, and 0.84 for training, test, and whole data set, respectively. The hybrid method was proved to be a very promising tool in the prediction of MDR‐modulating activity. Additionally, the application of strict validation criteria permits the high quality of the hybrid model. Thus, these in silico methods can be applied to predict this property in the early stage of drug development and some of them will be important tools to select new drug candidates.