论文信息 - Prediction-Weighted Partial Least-Squares Regression Method (PWPLS) 2: Application to CoMFA

Prediction-Weighted Partial Least-Squares Regression Method (PWPLS) 2: Application to CoMFA

Comparative molecular field analysis (CoMFA) has been used in drug design and three-dimensional quantitative structure−activity relationships (3D-QSAR). In CoMFA analysis, Partial Least-Squares (PLS) is used to correlate a large number of variables with biological activity. However, PLS may not clearly indicate which variables affect the biological activity of compounds. We have developed PWPLS (Prediction-Weighted Partial Least-Squares) that can select good predictor variables and weight each predictor variable to improve the predictiveness of its model. In addition to PWPLS, we developed another method, Q2 oriented variable selection (QOVS), to select variables that also affect predictiveness of its model. In this paper, we applied PWPLS and QOVS to the CoMFA study reported by Dunne et al. for binding 21 steroids to corticosteroid-binding globulin (CBG).

Yukio Tominaga | Iwao Fujiwara | Y. Tominaga | Iwao Fujiwara

[1] Garland R. Marshall,et al. 3D-QSAR of angiotensin-converting enzyme and thermolysin inhibitors: A comparison of CoMFA models based on deduced and experimentally determined active site geometries , 1993 .

[2] G. Cruciani,et al. Predictive ability of regression models. Part II: Selection of the best predictive PLS model , 1992 .

[3] G. Cruciani,et al. Comparative molecular field analysis using GRID force-field and GOLPE variable selection methods in a study of inhibitors of glycogen phosphorylase b. , 1994, Journal of medicinal chemistry.

[4] S. Wold,et al. The Collinearity Problem in Linear Regression. The Partial Least Squares (PLS) Approach to Generalized Inverses , 1984 .

[5] J N Weinstein,et al. Three-dimensional quantitative structure-activity relationship (QSAR) of HIV integrase inhibitors: a comparative molecular field analysis (CoMFA) study. , 1995, Journal of medicinal chemistry.