Multivariate linear QSPR/QSAR models: Rigorous evaluation of variable selection for PLS

Basic chemometric methods for making empirical regression models for QSPR/QSAR are briefly described from a user's point of view. Emphasis is given to PLS regression, simple variable selection and a careful and cautious evaluation of the performance of PLS models by repeated double cross validation (rdCV). A demonstration example is worked out for QSPR models that predict gas chromatographic retention indices (values between 197 and 504 units) of 209 polycyclic aromatic compounds (PAC) from molecular descriptors generated by Dragon software. Most favorable models were obtained from data sets containing also descriptors from 3D structures with all H-atoms (computed by Corina software), using stepwise variable selection (reducing 2688 descriptors to a subset of 22). The final QSPR model has typical prediction errors for the retention index of ±12 units (95% tolerance interval, for test set objects). Programs and data are provided as supplementary material for the open source R software environment.

[1]  E. Castro,et al.  Partial‐Order Ranking and Linear Modeling: Their Use in Predictive QSAR/QSPR Studies , 2012 .

[2]  Shushen Liu,et al.  Molecular structural vector description and retention index of polycyclic aromatic hydrocarbons , 2002 .

[3]  Matthias Dehmer,et al.  QuACN: an R package for analyzing complex biological networks quantitatively , 2011, Bioinform..

[4]  Ross Ihaka,et al.  Gentleman R: R: A language for data analysis and graphics , 1996 .

[5]  J. Gasteiger,et al.  Automatic generation of 3D-atomic coordinates for organic molecules , 1990 .

[6]  Matthias Dehmer,et al.  RMol: a toolset for transforming SD/Molfile structure information into R objects , 2012, Source Code for Biology and Medicine.

[7]  Curt M. White,et al.  Retention indices for programmed-temperature capillary-column gas chromatography of polycyclic aromatic hydrocarbons , 1979 .

[8]  Igor I. Baskin,et al.  Machine Learning Methods for Property Prediction in Chemoinformatics: Quo Vadis? , 2012, J. Chem. Inf. Model..

[9]  Riccardo Leardi,et al.  Genetic algorithms in chemistry. , 2007, Journal of chromatography. A.

[10]  P. Filzmoser,et al.  Repeated double cross validation , 2009 .

[11]  R. Tibshirani.  Regression Shrinkage and Selection via the Lasso , 1996 .

[12]  Johann Gasteiger,et al.  Impact of Conformational Flexibility on Three-Dimensional Similarity Searching Using Correlation Vectors , 2006, J. Chem. Inf. Model..

[13]  Wolfgang Marquardt,et al.  What is Wrong with Quantitative Structure-Property Relations Models Based on Three-Dimensional Descriptors? , 2012, J. Chem. Inf. Model..

[14]  Adam Hunter,et al.  Yabi: An online research environment for grid, high performance and cloud computing , 2012, Source Code for Biology and Medicine.

[15]  Discrimination between linear and non-linear models for retention indices of polycyclic aromatic hydrocarbons in the so-called Lee's scale , 2004 .

[16]  S. D. Jong.  SIMPLS: an alternative approach to partial least squares regression , 1993 .

[17]  Ron Wehrens,et al.  The pls Package: Principal Component and Partial Least Squares Regression in R , 2007 .

[18]  K. Varmuza,et al.  Multivariate models for the concentration of oxygen-18 in precipitation based on meteorological and geographical features , 2007 .

[19]  S. Wold,et al.  PLS-regression: a basic tool of chemometrics , 2001 .