Perpustakaan Digital ITB

Mutagenicity is one of the most popular end points to assess genotoxicity of chemicals. Mutagens are chemicals causing genetic mutations in humans. This can be experimentally examined by using Ames test. However, these days, there are numbers of mutagenicity models to predict the mutagenicity of the chemicals based on in silico/calculation method. To find out the best mutagenicity modelling among CONSENSUS model, CAESAR model, ISS model, SarPy model, KNN model, and CONCERT (aromatic amines) model, approximately 690 chemicals were collected in a database to be analyzed and calculated for its performance metrics. This prediction was validated based on various parameters including accuracy, balanced accuracy, Matthew’s Correlation Coefficient, coverage, sensitivity, selectivity, positive predictive value, and negative predictive value. Based on the results obtained in this study, it can be concluded that KNN model has the best prediction ability for mutagenicity, justified by the performance metrics conducted for each of the model.