Search

Structure-activity relationships (SARs) in toxicology have enabled the formation of structural rules which, when coded as structural alerts, are essential tools in in silico toxicology. Whilst other in silico methods have approaches for their evaluation, there is no formal process to assess the confidence that may be associated with a structural alert. This investigation proposes twelve criteria to assess the uncertainty associated with structural alerts, allowing for an assessment of confidence. The criteria are based around the stated purpose, description of the chemistry, toxicology and mechanism, performance and coverage, as well as corroborating and supporting evidence of the alert. Alerts can be given a confidence assessment and score, enabling the identification of areas where more information may be beneficial. The scheme to evaluate structural alerts was placed in the context of various use cases for industrial and regulatory applications. The analysis of alerts, and consideration of the evaluation scheme, identifies the different characteristics an alert may have, such as being highly specific or generic. These characteristics may determine when an alert can be used for specific uses such as identification of analogues for read-across or hazard identification., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2022 The Authors. Published by Elsevier Inc. All rights reserved.)

A group of triazole compounds was selected to investigate the confidence that may be associated with read-across of a complex data gap: repeated dose toxicity. The read-across was evaluated using Assessment Elements (AEs) from the European Chemicals Agency's (ECHA's) Read-Across Assessment Framework (RAAF), alongside appraisal of associated uncertainties. Following an initial read-across based on chemical structure and properties, uncertainties were reduced by the integration of data streams such as those from New Approach Methodologies (NAM) and other existing data. In addition, addressing the findings of the ECHA RAAF framework, complemented with specific questions concerning uncertainties, increased the confidence that can be placed in read-across. Although a data rich group of compounds with a strong mechanistic basis was analysed, it was clearly demonstrated that NAM data available from publicly available resources could be applied to support read-across. It is acknowledged that most read-across studies will not be so data rich or mechanistically robust, therefore some targeted experimentation may be required to fill the data gaps. In this sense, NAMs should constitute new experimental tests performed with the specific goal of reducing the uncertainties and demonstrating the read-across hypothesis., (Copyright © 2020 Elsevier Inc. All rights reserved.)

In silico models are used to predict toxicity and molecular properties in chemical safety assessment, gaining widespread regulatory use under a number of legislations globally. This study has rationalised previously published criteria to evaluate quantitative structure-activity relationships (QSARs) in terms of their uncertainty, variability and potential areas of bias, into ten assessment components, or higher level groupings. The components have been mapped onto specific regulatory uses (i.e. data gap filling for risk assessment, classification and labelling, and screening and prioritisation) identifying different levels of uncertainty that may be acceptable for each. Twelve published QSARs were evaluated using the components, such that their potential use could be identified. High uncertainty was commonly observed with the presentation of data, mechanistic interpretability, incorporation of toxicokinetics and the relevance of the data for regulatory purposes. The assessment components help to guide strategies that can be implemented to improve acceptability of QSARs through the reduction of uncertainties. It is anticipated that model developers could apply the assessment components from the model design phase (e.g. through problem formulation) through to their documentation and use. The application of the components provides the possibility to assess QSARs in a meaningful manner and demonstrate their fitness-for-purpose against pre-defined criteria. The three phases of QSAR development with related components associated with uncertainty, variability and bias. [Display omitted] • Ten components, or groups of assessment criteria, of QSARs are defined. • The components were mapped onto three phases of QSAR development and use. • QSARs assessed using the components with strategies to reduce uncertainty proposed. • Different uses of QSARs require different types of models. • The assessment components demonstrate fitness-for-purpose of QSARs. [ABSTRACT FROM AUTHOR]