High-level evidence base for Genotox-iS, including QMRF documentation, external validation and complementary rule-based/statistical approaches.
Genotox-iS is documented through separate QMRFs for its rule-based and statistical components and has been evaluated using internal and external validation approaches, including an international Ames/(Q)SAR challenge.
Rule-based component
The public QMRF describes a structural-alert model for bacterial mutagenicity and its applicability concepts.
Statistical component
The statistical battery uses complementary machine-learning hypotheses to evaluate Ames mutagenicity.
Validation
Performance should be considered using appropriate internal and external datasets, not only training-set fit.
Interpretation
Applicability-domain information and expert review remain essential at the compound level.
Frequently asked questions
Is Genotox-iS only a statistical model?
No. It combines complementary rule-based and statistical evidence.
Are the training datasets public?
Detailed proprietary training data are not publicly released, but model-level documentation is available in QMRFs.
Does validation remove the need for expert review?
No.
Can Genotox-iS be used in ICH M7 workflows?
It was designed to support ICH M7-aligned mutagenicity assessment.
