Table 1.

Proposed criteria to assess whether an Adverse Outcome Pathway (AOP) and associated bioassays are fit for purpose

Criteria	Description of criteria	Categories in each criterion
Types of AOPs	The criteria used to describe the types of AOPs relate to the extent of data surrounding KEs and key event relationships KERs. The Putative, Qualitative, Quantitative categories are adapted from Villeneuve et al. (2014).	Putative AOP: Assembly of a hypothesized set of KEs and KERs supported primarily through biological plausibility and/or statistical inference. Assembly of a partial AOP with incomplete linkages between the MIE and AO as a result of known data gaps and uncertainties. Qualitative AOP: Assembly of KEs supported by descriptions of how the KEs can be measured and KERs supported by empirical evidence in addition to plausibility or statistical inference, along with qualitative evaluation of the overall WoE supporting the AOP. A qualitative AOP does not contain threshold information for transitioning from one KE to another. Quantitative AOP: Assembly of KEs supported by mathematical descriptions of the KERs and the accuracy and precision with which the measurements are made, supported by quantitative understanding of what magnitude and/or duration of change in the upstream (preceding) KE is needed to evoke some magnitude of change in the downstream (subsequent) KE. A quantitative AOP contains biological thresholds for KER, or KE weighted in terms of AO probability (OECD 2016a).
Level of complexity of AOPs	Although AOPs may be interconnected across biological systems, they are first developed as more simple, linear constructs. Over time, connections to other AOPs at shared KE nodes may be developed and/or elucidated, such that a network of related AOP emerges (Knapen et al. 2018).	Linear AOP: Exists as a progression from MIE through consequent KEs to an AO without interactions from other AOPs at shared KE nodes. Network AOP: An assembly of 2 or more AOPs that share one or more KEs.
Confidence in the AOP	The level of confidence in the overall AOP is assessed by a WoE approach considering the KEs and KERs and their associated biological plausibility, essentiality, and empirical evidence (in that weight order) as described by Becker et al. (2015) and OECD (2016a, 2016b).	Low confidence: A WoE assessment of the AOP indicates that there is overall low confidence in the AOP in consideration of biological plausibility, essentiality, and empirical evidence. Moderate confidence: A WoE assessment of the AOP indicates that there is overall moderate confidence in the AOP in consideration of biological plausibility, essentiality, and empirical evidence. High confidence: A WoE assessment of the AOP indicates that there is overall strong confidence in the AOP in consideration of biological plausibility, essentiality, and empirical evidence.
Level of external review of the AOP	The AOP status in the AOP‐Wiki refers to the review process established by the OECD and included in the AOP knowledgebase (OECD 2016a, 2016b, 2016c, 2016d; AOPWiki 2019a). For reference, see https://aopwiki.org/oecd_page	Under development: oOpen for citation and comment, or oUnder development; not open for comment; do not cite. Under review by the EAGMST: oOpen for citation and comment, or oOpen for comment; do not cite. Approved by the EAGMST. Endorsed by the WNT and the WPHA.
KE in the AOP is of regulatory interest	The AOP includes a MIE, KE, or AO that is used for regulatory decision making. There is either an unambiguous connection to a regulatory endpoint (e.g., skin or eye irritation, sensitization, genotoxicity or mutagenicity, parameters linked to reproductive and/or developmental toxicity, endocrine‐disrupting properties, population‐level adverse effects in environmental organisms) or the AOP is several steps removed from an endpoint of regulatory interest.	Ambiguous connection: There is a more tenuous connection with the measurement and/or assessment endpoint of regulatory interest. Unambiguous connection: There is a clear, empirical connection to the measurement parameters and an assessment endpoint of regulatory interest.
Assay methods	Bioassay methods for investigating KEs in AOPs may be extensively standardized at the international or national level or may have been more recently developed without undergoing formal validation. Sufficient confidence in assay methods may be achieved through use of well‐described methods and assay performance in the context for which it is to be applied (ICCVAM 2018).	Nonvalidated and not well‐described methods Nonvalidated but well‐described methods (e.g., according to OECD GD 211 [OECD 2014b]) Nationally validated: The methods for chemical safety regulation have undergone national validation (e.g., USEPA) Internationally validated: The methods for chemical safety regulation (e.g., OECD Test Guideline) have undergone international validation or meet explicit performance criteria.

AO = adverse outcome; AOP = adverse outcome pathway; EAGMST = Extended Advisory Group on Molecular Screening and Toxicogenomics; GD = guidance document; KE = key event; KER = key event relationship; MIE = molecular initiating event; OECD = Organisation for Economic Co‐operation and Development; USEPA = United States Environmental Protection Agency; WNT = Working Group of the National Coordinators of the Test Guideline Program; WoE = weight of evidence; WPHA = Working Party for Hazard Assessment.