MARSplines approach for quantitative relationships between structure and pharmacological activity of potential drug candidates

An approach of multivariate adaptive regression splines (MARSplines) was applied for the quantitative structure-activity relationship studies of antitumor activity against murine leukemia L1210 of anthrapyrazoles as well as activated coagulation factor X (FXa) inhibitory activity of isosteviol analogues. These two different sets of molecules in the first stage underwent molecular modelling studies, i.e. geometrical optimization by the MM+ and the AM1 method using the Polak-Ribiere algorithm and finally about 5’000 molecular descriptors encoding structural features were calculated. Afterwards, statistical analysis using MARSplines algorithm was performed, which led to an establishment of a portfolio of submodels. As a result, the statistically significant MARS model for each set of the studied compounds that best describes quantitative structure-activity relationships was chosen. Elaborated models reveal, which molecular properties affect the most the pharmacological activity of anthrapyrazole and isosteviol compounds. Among the independent variables appearing in the statistically significant MARS models, descriptors belonging 2D Atom Pairs, 2D autocorrelations, 3D-MoRSE, GETAWAY, burden eigenvalues, RDF and WHIM descriptors, may be distinguished. The studies confirmed the benefit from using MARSplines algorithm, since high predictive power of obtained models make them useful for the prediction of antitumor and FXa inhibitory activity and possibly this approach can be considered as a tool for searching new drug candidates.