Solutions · By job to be done
Clear mutagenic-impurity calls with the trail to back them.
ICH M7 asks you to assess and control DNA-reactive impurities, and the classification hinges on structural alerts and a defensible (Q)SAR read. ToxIndex runs that assessment as a reproducible workflow so every impurity call carries its evidence.
Updated July 18, 2026
- 1:1
- reproducible: same inputs, same call
- Two
- complementary (Q)SAR reads per impurity
- Versioned
- audit trail on every run
The situation
The M7 workflow, made reproducible
ICH M7 leans on two complementary (Q)SAR methodologies and expert review to place each impurity into a control class. Doing that by hand across a synthetic route — starting materials, intermediates, reagents, degradants — is slow and hard to reproduce when an auditor asks how you got there.
Flow runs the assessment as an ontology-driven workflow. Each impurity is profiled for structural alerts and scored for bacterial mutagenicity, the results land in a classification, and the whole run keeps a versioned trail you can hand to a reviewer or rerun a year later unchanged.
How ToxIndex helps
From synthetic route to control strategy
A workflow built for M7
Flow runs the assessment as a structured template, one of the custom workflows Insilica hand-builds around a customer's process.
Structural-alert profiling
The QSAR Toolbox flags DNA-reactive alerts and groups impurities for read-across against the regulatory reference.
Mutagenicity prediction
ToxJobs scores bacterial mutagenicity with models like AMESformer, giving the second (Q)SAR read M7 calls for.
Ask the evidence directly
Sage lets a reviewer interrogate any impurity call in plain language and read the citations behind it.
The tools behind it
What runs underneath
FAQ
Questions teams ask
Frequently asked questions
How does ToxIndex handle ICH M7?
Flow runs the assessment as an ontology-driven workflow: each impurity is profiled for structural alerts and scored for bacterial mutagenicity, results land in a control classification, and the whole run keeps a versioned audit trail.
Does it provide the two (Q)SAR methodologies M7 asks for?
Yes. Structural-alert profiling through the QSAR Toolbox and a statistical mutagenicity prediction through ToxJobs give the two complementary (Q)SAR reads M7 calls for, with expert review on top.
Can an auditor reproduce an impurity call?
The same inputs always produce the same call, and the versioned trail lets a reviewer rerun the assessment a year later unchanged.
See it on your own chemistry.
Try the free version now, or book a walkthrough of the full platform on your substances.