Recently several ideas on QSAR modeling, based on about 20 years of experience in Insubria have been implemented in the QSARINS software for the development and validation of QSAR models, based on Genetic Algorithm for variable selection. Several chemometric tools (PCA, MCDM, ecc) are present for the analysis and selection of the best models. Several validation parameters and tools are available for the rigorous validation, mainly on external sets. Various graphs are available for the regression plot, residual analysis, applicability domain and outliers’ detection, etc. The consensus modeling by many models is also implemented . A module , named QSARINS-Chem, collects the chemical structures (3D) and the data modelled by the Insubria Group in these years. Several Insubria models, based on PaDEL descriptors, which are calculable directly from QSARINS also for new compounds, are available and applicable to new chemicals for prediction of physico-chemical properties, PBT behaviour or various toxicities. The QMRF are also reported for each model.
Gramatica et al. J. Comput Chem (Software News and Updates), 2013, 34, 2121; JCC, 2014, 35, 1036.