Utility of published DNA reactivity alerts
Alun Myden1, Sebastien J Guesne1, Alex Cayley1
1Lhasa Limited, Granary Wharf House, 2 Canal Wharf, Leeds LS11 5PS, UK.
Regulatory Toxicology and Pharmacology : RTP
|May 28, 2017
Summary
Literature-based structural alerts (LBSAs) increase false positives in mutagenicity assessments for pharmaceutical impurities. Overly sensitive LBSAs may lead to unnecessary Ames tests, impacting drug development efficiency.
Area of Science:
- Pharmaceutical Science
- Toxicology
- Computational Chemistry
Background:
- Regulatory guidelines (ICH M7) mandate the identification of mutagenic impurities in pharmaceuticals.
- In silico toxicity prediction tools are commonly used, requiring expert review for mutagenicity assessment.
- Literature-based structural alerts (LBSAs) have been proposed as aids for expert review of in silico predictions.
Purpose of the Study:
- To evaluate the impact of using a specific set of LBSAs in the expert review of mutagenicity predictions from two complementary in silico tools.
- To assess the performance of LBSAs compared to in silico systems in predicting mutagenic potential.
Main Methods:
- Analysis of an Ames test dataset comprising 2619 compounds.
- Application and interpretation of literature-based structural alerts (LBSAs).
- Comparison of LBSA performance against two complementary in silico toxicity prediction tools.
Main Results:
- Interpretation of LBSAs was found to be a subjective process.
- LBSAs generated significantly more false positives compared to the in silico systems.
- While some LBSAs showed comparable performance, most were overly sensitive, sacrificing accuracy.
Conclusions:
- The use of LBSAs without considering mitigating factors can lead to an increase in false positives during mutagenicity assessments.
- This increased false positive rate may necessitate additional, potentially unnecessary, Ames tests, impacting pharmaceutical development timelines and costs.
