Overview of the OECD (Q)SAR validation principles and the role of QMRF documentation in regulatory toxicology.
Regulatory use of (Q)SAR depends not only on a prediction but also on transparent model documentation. The OECD principles provide a framework for evaluating whether a model is scientifically defined and fit for interpretation.
Defined endpoint
The biological endpoint should be clearly specified.
Unambiguous algorithm
The method used to generate the prediction should be sufficiently defined.
Applicability domain
The chemical and response space should be characterized.
Performance
Goodness-of-fit, robustness and predictivity should be evaluated with appropriate data.
Mechanistic interpretation
When possible, the model should support a scientifically meaningful interpretation of the structure-effect relationship.
Frequently asked questions
What is a QMRF?
A structured reporting format for describing a (Q)SAR model and its scientific properties.
Does a QMRF prove regulatory acceptance?
No. It supports transparency and evaluation.
Are proprietary models compatible with QMRF?
Yes, provided sufficient model-level information is documented without requiring disclosure of protected data.
Why is mechanistic interpretation optional?
Some endpoints or models do not permit a unique mechanistic explanation, but interpretation should be provided when scientifically possible.
