Resource · Guide
ICH M7 Workflow Guide
From SMILES to a mutagenicity table ready for review by your experts: what ICH M7 (R2) requires and how CovaSyn makes the pre-work reproducible and traceable.
1. What ICH M7 (R2) Requires
ICH M7 (R2) is the international guideline for assessing and controlling DNA-reactive mutagenic impurities in pharmaceuticals. It defines five classes (1: known mutagenic carcinogens, 2: known mutagens with unknown carcinogenic potential, 3: alerting structures without mutagenicity data, 4: alerting structures shared with the drug substance or related compounds that tested negative in the Ames assay, 5: no structural alert, or an alert with sufficient data showing no mutagenicity). Each synthetic impurity is assessed using two complementary QSAR methods: an expert rule-based system (for example Derek Nexus) and a statistical system (for example Sarah Nexus, Leadscope or Bayesian models).
2. Classic Workflow (Excel and Manual Statistics)
Typical workflow today: copy SMILES by hand from synthesis notes, paste them into two separate web interfaces, export results as CSV, merge them in Excel, walk through the ICH M7 decision tree by hand, check the applicability domain and prepare the expert review table. It takes time, and reproducibility after months is weak: tool versions change, inputs drift, and reconstructing results becomes detective work.
3. Workflow with CovaSyn
You send us the list of your structures. CovaSyn runs two complementary assessments, derives a preliminary classification (without data at most class 3 or 5; classes 1, 2 and 4 require data or a comparison with the drug substance), documents the applicability domain per prediction and delivers a table your experts review and approve. The same inputs yield the same outputs.
4. Audit Trail
Every tool call is logged with timestamp, tool, version, result status and key identifier, and outputs carry a SHA-256 hash. Logs can be exported as CSV or JSON. During an inspection you can trace which assessment was produced with which tool version.
5. What Stays Human
Class 3 substances (alert present, no mutagenicity data) still require review by a qualified toxicologist. Classes 1 and 2 (known mutagens) are flagged but not blocked automatically; the decision stays with your toxicology, quality assurance and regulatory experts. Class 5 (no alert) is marked as unremarkable with an audit trail; release remains part of your process. The calculation takes the groundwork off your experts; the decision stays with them.
6. GAMP 5 Category 4 Classification
The ICH M7 endpoint can be classified as GAMP 5 Category 4 (configured product), not Category 5 (custom software). Validation therefore focuses on URS, FS, configuration, risk assessment and user acceptance testing, not on source-code reviews. For regulated environments a validation package is available on request (Enterprise) with templates for URS, risk assessment and IQ, OQ and PQ, which your QA adapts and signs. The validation itself takes place in your organisation.
7. Limits
The tool does not replace a qualified toxicologist. It makes the pre-work auditable and reduces manual effort. The final ICH M7 classification and submission strategy remain the responsibility of QA and regulatory affairs. Applicability domain flags must be taken seriously: structures outside the model space require explicit expert assessment.
Next step
Talk to us about your impurities.
Typical lab questions on publicly known compounds, every number computed and checked.