For those looking to deepen their understanding of the use of tools such as ToxTree or the OECD QSAR ToolBox, I highly recommend checking out our course NAMs - Use and application of QSAR and read-across, which provides valuable insights that complement the discussion in this post.
The Cramer classification scheme is a cornerstone of predictive toxicology, widely used to classify chemicals into three categories of toxicological concern. Its applications range from risk assessment to supporting the Threshold of Toxicological Concern (TTC) approach, particularly in food safety. However, inconsistencies in its application—whether by experts or in silico tools—pose challenges to its reliability and broader acceptance.
A new study titled "Evaluating the consistency of judgments derived through both in silico and expert application of the Cramer classification scheme" sheds light on these inconsistencies. This paper is a must-read for anyone looking to deepen their understanding of the scheme's complexities and the implications for computational toxicology.
Key Takeaways
Expert vs. In Silico Agreement. The study reveals a high degree of consistency (≥97%) among human experts in applying the Cramer rules. However, concordance between expert and in silico classifications (e.g., Toxtree and OECD QSAR Toolbox) is notably lower (~70%). This discrepancy highlights the need for improved algorithms and clarity in implementing Cramer rules digitally.
Prominent Sources of Disagreement. The paper identifies 22 chemical groupings, such as α,β-unsaturated carbonyls and secondary alcohols, where disagreements commonly arise. The issues stem from ambiguous rule interpretations, errors in in silico coding, and differences in subjective judgment.
The Role of Revised Decision Trees. The "Revised Cramer Decision Tree," developed to eliminate subjectivity, offers some improvements but introduces additional discrepancies. This demonstrates the challenge of balancing precision with usability in computational models.
Why This Matters. For toxicologists using these tools, understanding their limitations is crucial. Misclassifications could lead to unnecessary animal testing or overlooking hazardous substances, undermining confidence in TTC approaches.
How This Study Helps
By dissecting the origins of inconsistencies, the authors propose actionable solutions, such as improving in silico algorithms and refining the rules’ language. These insights empower toxicologists to apply the Cramer scheme more effectively and advocate for advancements in computational toxicology.
For those exploring computational toxicology or implementing TTC-based assessments, this paper is a valuable resource. It bridges the gap between expert judgment and in silico predictions, fostering a more consistent and reliable framework for chemical safety evaluation.
James W. Firman, Alan Boobis, Heli M. Hollnagel, Stefan Kaiser, David P. Lovell, Angelo Moretto, Severin Mueller, Cynthia V. Rider, Florian Schmidt, Szabina Stice, Sanjeeva J. Wijeyesakere, Geraldine Borja, Grace Patlewicz,
Evaluating the consistency of judgments derived through both in silico and expert application of the Cramer classification scheme,
Food and Chemical Toxicology, Volume 194, 2024, 115070, ISSN 0278-6915,
Abstract: The Cramer classification scheme has emerged as one of the most extensively-adopted predictive toxicology tools, owing in part to its employment for chemical categorisation within threshold of toxicological concern evaluation. The characteristics of several of its rules have contributed to inconsistencies with respect to degree of hazard attributed to common (particularly food-relevant) substances. This investigation examines these discrepancies, and their origins, raising awareness of such issues amongst users seeking to apply and/or adapt the rule-set. A dataset of over 3000 compounds was assembled, each with Cramer class assignments issued by up to four groups of industry and academic experts. These were complemented by corresponding outputs from in silico implementations of the scheme present within Toxtree and OECD QSAR Toolbox software, including a working of a “Revised Cramer Decision Tree”. Consistency between judgments was assessed, revealing that although the extent of inter-expert agreement was very high (≥97%), general concordance between expert and in silico calls was more modest (∼70%). In particular, 22 chemical groupings were identified to serve as prominent sources of disagreement, the origins of which could be attributed either to differences in subjective interpretation, to software coding anomalies, or to reforms introduced by authors of the revised rules.
Keywords: Cramer classification scheme; Toxtree; OECD QSAR toolbox; In silico; Risk assessment
Kommentarer