Prediction of bioactivity of HIV-1 integrase ST inhibitors by multilinear regression analysis and support vector machine
Xuan, S.Y.; Wu, Y.B.; Chen, X.F.; Liu, J.; Yan, A.X.*
Bioorganic & Medicinal Chemistry Letters, 2013,23(6), 1648-1655.
In this study, four computational quantitative structure–activity relationship models were built to predict the biological activity of HIV-1 integrase strand transfer (ST) inhibitors. 551 Inhibitors whose bioactivities were detected by radiolabeling method were collected. The molecules were represented with 20 selected MOE descriptors. All inhibitors were divided into a training set and a test set with two methods: (1) by a Kohonen’s self-organizing map (SOM); (2) by a random selection. For every training set and test set, a multilinear regression (MLR) analysis and a support vector machine (SVM) were used to establish models, respectively. For the test set divided by SOM, the correlation coefficients (rs) were over 0.91, and for the test set split randomly, the rs were over 0.86.
QSAR Models performance: Dataset (551 HIV-1 Integrase ST inhibitors)
|Model Name||Algorithm||Descriptors||Spliting method||Training set numbers||Training set r||Training set RMSE||Training set MAE||Test set numbers||Test set r||Test set RMSE||Test set MAE|
|Model 1A||MLR||20 MOE descriptors||Kohonen’s self-organizing map (SOM)||355||0.89||0.58||0.48||196||0.91||0.41||0.44|
|Model 2A||MLR||20 MOE descriptors||Random||368||0.91||0.54||0.44||183||0.86||0.47||0.52|
|Model 1B||SVM||20 MOE descriptors||Kohonen’s self-organizing map (SOM)||355||0.97||0.21||0.13||196||0.93||0.36||0.39|
|Model 2B||SVM||20 MOE descriptors||Random||368||0.99||0.21||0.13||183||0.90||0.41||0.44|