Comparative Evaluation of in Silico Systems for Ames Test Mutagenicity Predict.pdf
The current step 2 document of the International Conference on Harmonisation (ICH) M7 guidance recommends the use of two (Q)SAR prediction methodologies to qualify genotoxic impurities during the drug safety evaluation and approval process. One methodology should be expert rule-based and the second methodology should be statistical-based. Derek Nexus is a widely accepted and used expert rule-based system. The predictive power of two statistical tools for mutagenicity prediction (Leadscope Model Applier and Sarah Engine Prototype) was evaluated in a comparative manner using a proprietary dataset of 190 intermediates tested in the Ames assay (14,2% positives). In addition the performance of both statistical systems combined with Derek Nexus was assessed.