Report of the Workshop on the Validation of QSARs and Other Computational Prediction Models
Andrew P. Worth and Mark T.D. Cronin
This workshop addressed current issues with regard to establishing a framework for the validation of quantitative structure–activity relationships (QSARs) and other computational prediction models. QSARs and related models attempt to associate physicochemical and structural properties of compounds to their biological activity. As such, they may permit the prediction of biological activity from chemical structure alone. As yet, no formal validation criteria have been agreed internationally for QSARs and related techniques. However, some general and preliminary criteria were discussed and agreed upon at a European Chemical Industry Council/International Council of Chemical Associations (CEFIC/ICCA) workshop on the regulatory acceptance of QSARs, held in Setubal, Portugal, in March 2002. These criteria were presented at a Fourth World Congress workshop, along with a proposal for the practical validation of computer models, such as QSARs.