Technical overview of complementary rule-based and statistical (Q)SAR methodologies for ICH M7 bacterial mutagenicity assessment.
ICH M7(R2) provides the current framework for assessment and control of DNA-reactive mutagenic impurities. When (Q)SAR is used for bacterial mutagenicity, complementary expert rule-based and statistical methodologies provide independent lines of structural evidence.
Rule-based methodology
Structural alerts link specific molecular features to known or plausible mutagenicity mechanisms and provide interpretable evidence.
Statistical methodology
Statistical models compare the query structure with patterns learned from experimental Ames data and can use complementary molecular representations.
Applicability domain
Each model must be interpreted within the chemical and response space for which it is sufficiently supported.
Expert review
Predictions should be integrated with structure quality, analogue evidence, experimental data and the regulatory context.
Frequently asked questions
Does ICH M7 use two complementary methodologies?
Yes, an expert rule-based and a statistical-based methodology are used as complementary approaches when (Q)SAR is applied.
Is a structural alert automatically a positive regulatory conclusion?
No. Expert review and complementary evidence matter.
What is applicability domain?
The chemical/response space in which the model can make sufficiently supported predictions.
Can experimental data override a model?
Relevant reliable experimental evidence is an important part of the weight of evidence.
