In this study, our researchers combined an automatic model generation process for building QSAR models with the Gaussian Processes machine learning method. The article outlines the process for model construction and validation, and applies the automatic process to  blood-brain barrier penetration and aqueous solubility data sets. Results are compared with ‘manually’ built models using external test sets, and show the automatic process to be highly effective for both in vivo and physico-chemical data.

Citation details

O. Obrezanova, J. M. R. Gola, E. Champness, M. D. Segall, J. Comp. Aided Mol. Design, 2008, 22(6-7) pp. 431-440.
DOI: 10.1007/s10822-008-9193-8

More QSAR and ADME resources