F1000Research F1000Research 2046-1402 F1000 Research Limited London, UK 10.12688/f1000research.172881.2 Research Article Articles Methods2AOP: A Collaboration to Strengthen the Integration of Test Methods into the Adverse Outcome Pathway Framework

[version 2; peer review: 1 approved, 3 approved with reservations]

 Karmaus Agnes L. Conceptualization Investigation Methodology Project Administration Writing – Original Draft Preparation Writing – Review & Editing a 1 Bisson William Conceptualization Investigation Methodology Project Administration Writing – Original Draft Preparation Writing – Review & Editing b 1 Braeuning Albert Investigation Writing – Review & Editing 2 Chang Xiaoqing Investigation Writing – Review & Editing 1 Clerbaux Laure-Alix Investigation Writing – Review & Editing 3 Filipovska Julija Investigation Writing – Review & Editing 4 Fostel Jennifer Investigation Writing – Review & Editing 5 Groh Ksenia Investigation Writing – Review & Editing https://orcid.org/0000-0002-3778-4721 6 Hench Ginnie Investigation Software Writing – Review & Editing 7 Lekka Eftychia Investigation Writing – Review & Editing 8 Lynn Scott G. Investigation Writing – Review & Editing 9 Magurany Kelly A. Investigation Writing – Review & Editing 10 Markey Kristan Investigation Writing – Review & Editing https://orcid.org/0000-0003-3911-2969 9 Masci Anna Maria Investigation Writing – Review & Editing 11 Mortensen Holly Investigation Writing – Review & Editing https://orcid.org/0000-0002-3180-5229 9 O'Brien Jason M. Investigation Writing – Review & Editing https://orcid.org/0000-0002-2546-8940 12 Reinke Emily Investigation Writing – Review & Editing 1 Tucker Nyssa Investigation Writing – Review & Editing 13 Virvilis Vassilis Investigation Writing – Review & Editing 8 Viviani Barbara Investigation Writing – Review & Editing 14 Vliet Sara Investigation Writing – Review & Editing 9 Wittwehr Clemens Conceptualization Investigation Methodology Project Administration Supervision Writing – Original Draft Preparation Writing – Review & Editing https://orcid.org/0000-0003-2760-7702 c 15 T. Hogberg Helena Conceptualization Investigation Methodology Project Administration Writing – Original Draft Preparation Writing – Review & Editing d 5 Inotiv, Research Triangle Park, USA Bundesinstitut fur Risikobewertung, Berlin, Berlin, Germany Universite catholique de Louvain Institut de Recherche Experimentale et Clinique, Louvain, Belgium Independent Researcher, Ohrid, North Macedonia National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina, USA Eawag Swiss Federal Institute of Aquatic Science and Technology, Zurich, Switzerland OpenBioDataModeling, Research Triangle Park, USA Biovista, Athens, Greece US Environmental Protection Agency, Washington, USA Verto Solutions LLC, Washington DC, USA The University of Texas MD Anderson Cancer Center, Houston, Texas, USA Environment and Climate Change Canada National Wildlife Research Centre, Ottawa, Ontario, Canada The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA Universita degli Studi di Milano-Bicocca, Milan, Lombardy, Italy European Commission Joint Research Centre, Ispra, Italy agnes.karmaus@syngenta.com wbisson74@gmail.com clemens.wittwehr@ec.europa.eu helena.hogberg-durdock@nih.gov

No competing interests were disclosed.

 18 2 2026 2025 14 1375 30 1 2026 Copyright: © 2026 Karmaus AL et al. 2026 This is an open access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The Adverse Outcome Pathway (AOP) framework is a pivotal tool for organizing mechanistic knowledge and linking it to adverse outcomes of regulatory significance. However, the integration of test method information, particularly New Approach Methods (NAMs), within the central repository for AOP knowledge, (the AOP-Wiki), has been suboptimal, limiting the framework’s utility for regulatory decision-making. The Methods2AOP collaboration, comprised of various international stakeholders, was established to address this gap and enhance the role of test methods within the AOP framework. This paper reviews their work emphasizing the importance of linking detailed test method information and conceptually proposes how it may be included in the AOP knowledgebase in alignment with existing assay documentation standards and governance frameworks. The Methods2AOP collaboration proposes using ontologies to standardize and structure information, thereby facilitating interoperability, enabling reusability, and establishing clear connections between test methods and Key Events (KEs). A conceptual model is presented to demonstrate qualitative similarities between concepts in key event components and structured methods information. The implementation of Methods2AOP recommendations would increase the clarity and transparency of method descriptions, which could support regulatory acceptance and a wider adoption of NAMs. The broad community of stakeholders impacted by this work stands to benefit from the Methods2AOP recommendations through enhanced regulatory decisions, increased visibility and scientific impact, new market opportunities, and the accelerated adoption of NAMs in regulatory affairs. In summary, the Methods2AOP collaboration presents a comprehensive effort to formally standardize the integration of test methods into the AOP framework, thereby fostering a more robust, and transparent system that aligns with the goals of the scientific and regulatory communities.

 Plain language summary

New approach methods (NAMs) are state-of-the-science methods that can provide mechanistic insight into chemical safety instead of relying on traditional whole animal test systems alone. The adverse outcome pathway (AOP) is a framework that combines data from various sources including NAMs and traditional studies, and can therefore link mechanisms to adverse outcomes. This helps building confidence in using NAMs instead of animals in hazard and risk assessment. AOPs can be found in the AOP-Wiki, a web interface that includes information supporting the AOP. However, it contains limited information about the actual test methods. The Methods2AOP collaboration was initiated to identify, prioritize, organize and standardize methods information using a harmonized language. The implementation of the Methods2AOP collaboration is expected to provide a powerful and transparent system that can support the use of NAMs and the AOP framework in various contexts, including hazard and risk assessment.

 Adverse Outcome Pathways (AOPs) New Approach Methods (NAMs) regulatory framework ontologies National Institute of Environmental Health Sciences, National Institutes of Health HHSN273201500010C US EPA Endocrine Disruptor Screening Program (EDSP) 68HE0H18D0008 the Brains for Brussels program supported by Innoviris – Brussels Capital Region 2023-BFB-1 The author(s) declared that no grants were involved in supporting this work. Amendments from Version 1

We revised the manuscript in response to the two reviewers’ constructive comments, which helped us improve clarity, positioning, and practical relevance. In particular, we clarified the intended scope of the Methods2AOP framework, its relationship to existing assay description standards and governance structures, and its role as a complementary, interoperable approach rather than a new or competing standard. We strengthened the conceptual description of core entities (e.g. method, assay, observation) and their boundaries, clarified how structured method information can support confidence in AOP evidence without prescribing regulatory workflows, and added explicit discussion of governance, quality control, and long-term maintenance considerations within existing AOP-Wiki and OECD processes. We also addressed perceived in vitro bias by clarifying applicability across evidence streams and study types. Importantly, following reviewers’ requests for greater concreteness, we substantially expanded the supplementary material. The supplementary file now includes four illustrative, real-world worked examples (in vitro, in vivo, and in chemico) demonstrating how selected Methods2AOP fields can be populated in practice. These examples intentionally illustrate progressive disclosure, partial field population, and the balance between ontology-based annotation and expert free text, thereby making the proposed approach more tangible and easier to interpret. Overall, the revisions improve transparency, usability, and practical understanding of the Methods2AOP proposal while remaining aligned with existing standards and governance frameworks.

 1. The adverse outcome pathway framework: Integrating data sources for hazard and risk assessment 1.1 Background

The Adverse Outcome Pathway (AOP) conceptual framework was developed to address the need of the regulatory and scientific community to better organize and streamline existing toxicological mechanistic knowledge, and link to adverse outcomes (AO) of regulatory significance ( Ankley et al., 2010). Moreover, AOPs outline a specific cascade and assess the evidence related to the essentiality of particular biological perturbations (i.e., molecular initiating events (MIEs) and key events (KEs)) and causality of the linkages between them (key event relationships (KERs)), thus proposing a plausible pathway leading to an AO at the organ, organism and/or population level. The AOP framework focuses on the evidence for linkages between the critical/essential KEs in a stressor-agnostic approach ( Villeneuve et al., 2014; Clerbaux et al., 2024). This supports integration of various data sources (Integrated Approaches to Testing and Assessment - IATA) including new approach methods (NAMs) that can detect and measure a KE at lower organizational levels, i.e., at the molecular and cellular level leading to an AO at the organizational level of interest. A benefit in developing and applying AOPs is to build confidence in using NAMs data to characterize trajectory to adversity, specifically hazard for an endpoint, without using whole animals. Once the AOP is well-established, several NAMs that measure the various KEs can be combined into a defined approach (DA) or IATA to provide an actionable framework for the use of AOPs. The newly adopted OECD Guideline No. 497: Defined approaches on Skin Sensitization is an example where the AOP framework has contributed to the acceptance of NAMs in a regulatory application ( OECD, 2023b). However, more complex endpoints, e.g., carcinogenicity and developmental neurotoxicity (DNT), require networks of many AOPs that still need to be enhanced ( Pitzer et al., 2023; OECD, 2023a; Sanz-Serrano et al., 2024). Application of NAMs for hazard evaluation does not require AOPs. However, well-developed and endorsed AOPs can help increase confidence in NAMs and support regulatory use by the organisation of test batteries that address events critical to the progression to a regulatory relevant AO ( ICCVAM, 2024; Bajard et al., 2023; van der Zalm et al., 2022).

The scientific development of NAMs and their uptake for mechanistic investigation are steadily progressing ( Bisson et al., 2024), yet application in toxicological risk assessment is lagging. To align the progress of method development and mechanistic assay use with the acceptance for regulatory application, more interaction between the method developers, regulatory communities and end users is needed ( van der Zalm et al., 2022; ICCVAM, 2018). To support this communication and bolster the role of AOPs in regulatory application, the international Methods2AOP collaboration was established to develop a formal proposal for the integration of structured methods documentation for AOPs. Given the fundamental premise that the KEs in an AOP must be measurable, methods have a crucial role and therefore should be robustly captured in the AOP description, which in turn can help with transparency and bolster regulatory confidence ( Carusi et al., 2022).

 1.2 Current AOP infrastructure

Paramount to the undertaking of the Methods2AOP effort was a clear understanding of the current AOP infrastructure and uses. With a future goal of integrating additional information into the schema, a diverse group of stakeholders were engaged to leverage past experiences. It is important to distinguish the Methods2AOP collaboration’s method-centric scoping within the AOP infrastructure from the existing AOP-knowledgebase (AOP-KB) as the intention was to identify information that could add to and refine the existing infrastructure, not to re-build established interfaces.

The central location used for managing, and disseminating AOP knowledge is the AOP-KB ( Figure 1), 1 the database underlying the AOP-Wiki. 2 The AOP-KB and AOP-Wiki are administerd by the AOP-KB Coordination Group under the auspices of the AOP Program 3 of the Organisation for Economic Co-operation and Development (OECD). The AOP-Wiki, the official resource for creating and editing AOP content, has been available since 2012 and is currently on version 2.7 with roughly 1000 registered users working on more than 500 AOPs. Many more non-registered users access the AOP-Wiki in read-only mode to view its content. The AOP-KB content, exposed in XML, JSON, and TSV formats through the AOP-Wiki, serves as a resource used by third party tools, a collection of web-based resources, developed by AOP community partners that are constantly undergoing development and refinement.

Database connection of the AOP Knowledge Base, AOP Knowledge Base Tools, and the AOP-Wiki.

The AOP-KB (left side) is the primary central repository for all AOPs developed either as part of the OECD AOP Programme, or by the larger scientific community. The AOP-KB Tools (right side) are a collection of web-based resources that are constantly undergoing development and refinement.

The Methods2AOP collaboration suggest content and a conceptual documentation schema, presented herein for future consideration to integrate into the AOP-KB. Accordingly, any structured method information inspired by Methods2AOP would be subject to the same governance, review, and quality control process that already apply to AOP-Wiki content. It is important to consider that any future changes made to the AOP-KB and interface of the AOP-Wiki would require considerable effort and restructuring that will ultimately have an impact on a variety of third-party tools. There are other ongoing efforts aimed at improving AOP standards within the AOP-KB ( Hench et al., 2024b), such as the Environmental Health Language Collaborative (EHLC) AOP Standards Workshop, 4 the FAIR AOP Cluster working group ( Mortensen et al., 2025) and the ELIXIR Toxicology Community, 5 that all need to be aligned. Ultimately, if revisions to the AOP-KB structure and data model relationships are made to reflect integration of methods information, it will need to be conducted thoughtfully with all tools that draw from the centralized AOP-KB in mind. This way updates can be pushed to ensure all official tools and even third-party tools can keep up to date.

In the current AOP-Wiki interface (representing database fields accounted for in AOP Data and mapped in the Data Model), AOP authors can provide qualitative information on KEs and appraise the evidence and quantitative understanding (high, moderate, low) of their linkages (i.e., KER) according to the Guidance Document on Developing and Assessing AOPs, 6 and its practical supplement, the Developers’ Handbook. 7 However, test methods are currently not documented in the AOP-Wiki in a systematic way, rather, AOP authors can fill in an unstructured free text section “How It Is Measured or Detected” (“It” being the relevant KE) to provide any information they chose regarding test method(s). According to the Developers’ Handbook, 6 this section is not intended to provide detailed protocols but to capture the type of output that can be employed to evaluate the KE and the level of scientific confidence in the underlying measurements. This does not reflect the important role of the test methods, and does not allow compiling and sharing structured information. The lack of structured documentation for test methods limits transparency and creates inconsistency, which potentially makes it difficult for users (e.g., regulators) to understand the relevance and the reliability of the KE.

 2 Role Served by the Methods2AOP Collaboration 2.1 Establishing the Methods2AOP Collaboration

One of the fundamental principles within the AOP community is “crowdsourcing”: the voluntary and result-driven temporary collaboration of a group of individuals to reach a common goal. Crowdsourcing means tapping into a big, diverse group of people to gather their ideas, skills, or knowledge for a specific purpose. Instead of relying on just a few experts, numerous different people can be invited to contribute their thoughts or talents. This is institutionalized in the AOP framework which aims at getting a wide range of input (recorded in MIEs, KEs, AOs and KERs) that might not be available in any other fashion. By its nature, crowdsourcing needs participants who are willing to pool their expertise and actively collaborate with others to produce relevant AOPs ( Carusi et al., 2023). When the goal of addressing test method descriptions in the AOP-Wiki was identified, it was therefore natural that the same crowdsourcing concept be applied to building the Methods2AOP collaboration. Participants from interested organizations were assembled to formulate the problem, review possible solutions, and agree on a way forward.

The Methods2AOP collaboration sought to refine the current “How It Is Measured or Detected” free text field into more structured documentation of methods information. Methods2AOP collaborators identified and defined relevant fields (including both the test system and the assay endpoint), and proposed a conceptual framework that would allow the association of methods information with one or more MIE/KE(s)/AO ( Figure 2). More specifically, the proposal includes detailed, structured annotation of methods that could benefit AOP-Wiki users as well as the method developer community. Inclusion of method developers takes the burden off the AOP developers and opens the door for further broadening the AOP community and facilitating collaborations. The envisioned outcome of implementing the Methods2AOP collaboration’s proposal would allow for interoperable methods information across the AOP-Wiki reinforcing the biological relevance of a NAM, thus helping build confidence in NAMs and AOPs for regulatory application.

Methods information incorporated to AOPs.

The Methods2AOP collaboration proposes the integration of discrete methods information into the Adverse Outcome Pathway (AOP) framework where the AOP construct portrays existing knowledge concerning the linkage between a molecular initiating event (MIE) and an adverse outcome (AO), via measurable key events (KE) at a biological level of organization relevant to risk assessment. The proposal allows for methods (visualized as A to D), including test system and assay endpoint, to be annotated to one or more KEs.

The Methods2AOP collaboration started as a discussion group of a few like-minded individuals which quickly attracted international interest. The original crowdsourcing approach that drew in stakeholders from the existing AOP community quickly evolved into a collaboration of people who represent diverse viewpoints and backgrounds, all committed to integrate the latest state of the science and policy into the AOP community. Acknowledging existing related activities focused on assay description and documentation ( Krebs et al., 2020; Krebs et al., 2019; Bal-Price et al., 2018; Leite et al., 2024), the Methods2AOP collaboration established strong connections to parties that had embarked on similar efforts, including some joining the collaboration outright. It was important that the Methods2AOP collaboration engaged government, academia, industry, and contract research organizations who all have experience with assay development and use from different perspectives. Regulators, researchers, commercial users, and industry end users with familiarity of using AOPs or NAMs were engaged to create the collaboration. This diversity was paramount to work toward attracting a new demographic into the AOP community: test method developers.

 2.2 Identifying and Prioritizing Relevant Methods Information

The Methods2AOP collaboration defined scope, confirmed relevance, and found balance in identifying and prioritizing sufficient and appropriate fields to document methods information. The initial scoping process identified several potential fields and revealed initial major challenges such as determining if a field should be required, assessing its usage potential (e.g., would AOP authors or regulators find this field informative and use it), and linking fields with an appropriate structured controlled vocabulary/same language. To exercise restraint and not overcomplicate, discussions were focused on the minimum critical information for AOP use and relevance to AOP adoption for regulatory applications. With this in mind, the collaboration prioritized key information by leveraging group polling to integrate ranking preference across the Method2AOP members.

Votes of variable weight were awarded such that each voter was required to award points to each field/contestant (8 points for the most favorite field, 1 point for the least favorite). In the end, fields with the highest number of points ranked highest. The outcome of this attempt identified 24 fields relevant to methods information that then were grouped based on their “scope”: (1) scientific/technical, (2) analysis/interpretation and (3) validation/regulatory ( Table 1). This exercise showed that more defined, discrete fields were needed. Interestingly, the highest ranked fields were the most defined while the open-ended information fields were among the lowest-ranked ones. Thus, by polling collaboration members, important information regarding not only scientific content but also what makes a field informative was gathered. These fields were then evaluated by populating them with real methods information from specific KE chosen by various members of the Methods2AOP collaboration.

Information fields initially scoped and ranked by the Methods2Aop collaboration.
Rank Field scope Information documented Input type
1 Scientific/Technical Targeted biological endpoints (process or pathway/gene or protein/metabolites) ConVocab
2 Analysis/Interpretation Relationship between detection technology and measured endpoint ConVocab
3 Validation/Regulatory Measured observation (if relationship is indirect) ConVocab
4 Scientific/Technical Method type ConVocab
5 Validation/Regulatory Standard protocol availability ConVocab
6 Scientific/Technical Experimental setting ConVocab
7 Scientific/Technical Detection technology ConVocab
8 Scientific/Technical Species specificity ConVocab
9 Validation/Regulatory Method trade name and provider ConVocab
10 Scientific/Technical References ConVocab
11 Scientific/Technical Tissue or cell type specificity ConVocab
12 Validation/Regulatory Status of the method (validated?) ConVocab
13 Analysis/Interpretation Relevance to the AOP KE (one per linked KE) ConVocab
14 Analysis/Interpretation Other KE this method has been applied to ConVocab
15 Scientific/Technical Interferences or confounder Free Text
16 Scientific/Technical Limitation(s) Free Text
17 Scientific/Technical Summary Free Text
18 Scientific/Technical Duration of treatment ConVocab
19 Scientific/Technical Window of sensitivity ConVocab
20 Scientific/Technical Complexity level ConVocab
21 Analysis/Interpretation Preference ranking among methods for this KE ConVocab
22 Validation/Regulatory Popularity Free Text
23 Scientific/Technical Other factors to be considered Free Text
24 Scientific/Technical Factor to be considered for IVIVE applicability Free Text

Note: Fields were ranked by 14 members of the Methods 2AOP collaboration.

ConVocab: Controlled vocabulary/same language 8 (entry would be defined by a term list, taxonomy, or ontology).

 2.3 Organizing and Standardizing Relevant Methods Information

To elevate the description of test methods from free text to structured information that is more easily searchable and comparable, metadata field organization was refined and ontologies were introduced. Content fields were organized hierarchically into tiers to offer further context and structure ( Figure 3). Each of the content tiers contains numerous fields of detailed information for which ontological corresponding terms were identified, where available, from the Open Biological and Biomedical Ontologies (OBO) Foundry 9 ( Karmaus et al., 2025). Establishing some standardized fields that leverage controlled vocabularies to ensure alignment with the FAIR principles (findable, accessible, interoperable, and reusable). 10 A simple structure of tiers for method information was defined: (1) study type, (2) study type specific methods information groups, (3) detailed fields per group. The study types include in vitro, in silico, in vivo, human, ecological, and other/unclear (e.g., ex vivo where likely a complement of in vivo and in vitro fields could be required to document sample origin in addition to subsequent methodology). Each of these study types would have their own groups of methods information needs. In addition to listing the proposed metadata fields and ontology references, the supplementary dataset also includes four illustrative, non-exhaustive worked examples demonstrating how selected fields can be populated for in vitro, in vivo, and in chemico concrete test methods have been provided ( Karmaus et al., 2025). The worked examples intentionally do not populate all available fields, reflecting the principles of minimizing reporting burden while increasing transparency and easily interpretable information entry. Resources containing annotated methods using similar metadata fields that can support such data integration can include CAMERA (Collection of Alternative Methods for Regulatory Application; https://camera.niehs.nih.gov/) and CEBS (Chemical Effects in Biological Systems; https://cebs.niehs.nih.gov/cebs/).

Method information tiering strategy.

Discrete Study Types are listed, each of which would have study type specific methods and detailed information. The study types include in vitro, in silico, in vivo, human, ecological, and other/unclear. Specific methods information groups for the in vitro study type are shown in blue (detailed fields are provided in Figure 4). In addition to the study type information, stressor information would be independently defined (green), as would assay information pertaining to technical aspects of quantification/readouts (orange). In sum, all this information would be combined in a comprehensive document relevant to interpreting methods data in the AOP context (these relationships are defined conceptually in Figure 5).

As an example, the details relevant to the in vitro study type are presented in Figure 4 including documentation, species, subject type, care, applicability domain, and input sample/assay. To conceptually structure test method information within the context of the existing AOP-KB Data Model, the Methods2AOP collaboration proposes an interactive progressive disclosure interface that would prompt appropriate content input presenting only the necessary or requested information to users at any given time, thereby keeping the interface simple and clean. This approach would gradually reveal additional options, features, or content as needed (e.g., select a study type and then see the fields relevant to that study type in the next step). Employing a progressive disclosure approach for information submission reduces the cognitive burden placed on content submitters, while facilitating an opportunity for deeper engagement and comprehension of details for AOP-Wiki information consumers.

Example of the detailed fields within the in vitro Study type for methods integration to the AOP infrastructure.

Stressor detailed fields and Assay detailed fields are also provided to offer a more complete representation of information that would be documented from the new tiered structured data approach.

The pivotal role of ontologies in this refined approach cannot be overstated. Showcasing real-world application of ontologies to replace formerly free-text fields underscores how data standardization can be leveraged to update the AOP framework, facilitating easier navigation, input, review, and interpretation for ultimate use. The selection of ontologies was guided by their consistency and the ease with which they facilitate the integration of terms ( Figure 4). The use of ontologies provides a structured framework that ensures terms are aligned and interoperable within the broader data ecosystem. By establishing clear relationships with the annotation terms used, we elevated the description of test methods from basic textual representations to comprehensive and meaningful structured sets of methods information. Furthermore, this can support identifying overlaps where methods can inform on more than one KE, drive the adoption of robust methods to assess KEs, and build trust and confidence from the regulatory community toward accepting data from methods selected to inform on steps in any given AOP.

 3. Prospective Integration and Implementation

The Methods2AOP collaboration further discussed how methods information prospectively could be integrated into the current AOP-KB framework. In the current AOP-Wiki infrastructure KEs that are AOs have an input field for entering the “Regulatory Significance of the Adverse Outcome”, thereby providing a way to document linkages between regulatory endpoints and AOPs. Moreover, KEs contain key event components, comprised of a biological object and/or process, along with an action term that describes a direction of perturbation ( Ives et al., 2017). We propose a conceptual model that highlights the qualitative similarities between concepts used in key event components and the structured and detailed methods information suggested by the Methods2AOP (illustrated by the grey box in Figure 5).

Methods2AOP test method and assay information conceptual model.

Schematic diagram demonstrating the contributions outlined by the Methods2AOP collaboration (within the large dark grey box) and its connection points to existing AOP-KB components in light blue (e.g., Regulatory Endpoint, Key Event, and Key Event Component. Terms are linked by defined relationships that each serve a distinct purpose within the schema determined by their ontological definitions and the nature of the entities they connect. Relationship definitions: “is proxy for” (ontological reference AFX: 0002821) representing a surrogate entity used for practical measurement or observation when the actual target entity cannot be directly referenced. This is particularly important in our schema where measurements serve as proxies for underlying biological phenomena. “Has evidence” / “is evidence for” (RO: 0002558/RO: 0002472) captures the support relationship between an observation (or material/process) and the assertion it substantiates. These relations articulate how test method components provide grounding for Key Events or Regulatory Endpoints. “Has part” (BFO: 0000051) reflects the compositional relationship, correctly applied where an entity (e.g., Observation) structurally includes subordinate components (e.g., Study Design, Effect). “Is specified by” (SIO: 000339) relates an entity to an information content entity that describes or specifies it. “Has output” (RO: 0002234) relates a process and an entity (material or information) generated as a result of that process. “Is about” (IAO: 0000136) a relationship between an information content entity and the entity it is about.

KEs do not currently have a field for “Observations” in the AOP-KB, rather there is a “methods of measurement” free text box. The proposed conceptual model, includes an “Observation” field defined as comprising information on study design and the measured effects (for which detailed inputs are summarized by the tiered data and detailed fields in Figures 3 & 4). Preliminary integration of “Observation” is already under development in the AOP-KB and has been presented as part of previous KER-centric Evidence Modeling prototyping efforts ( Hench et al., 2024a; O’Brien and Yauk, 2022). Challenges in integrating this field lie in the complexity of its parts (e.g., study design and effect which in turn also have their own ontological components) as “Observation” would include a diversity of evidence for any method that yields data to support a KE. By defining ontological terms appropriate per field, the Methods2AOP collaboration seeks to bound the diversity of metadata and simplify curation burden.

The Methods2AOP collaboration conceptual schema leverages structured relationships to ensure connections between information. The conceptual schema would be part of the Key Event as supporting evidence and is broken in to key elements wherein the “observation” has parts (is associated with information about) “study design” and the “effect” being evaluated. The “effect” would be directly about (and linked back to) the existing Key Event Component currently defined in the AOP-KB. Further attributes defined in the schema that are associated to these fields include “stressor” (e.g., test article), information about the “test system” and “assay endpoint”, and attributes providing “documentation” and the output “measurement” (fields that would comprise study design and assay details Figures 3 & 4). Figure 5 contains these terms and shows the relationship between the information. Type, details, and structure of information that could be added to the AOP-KB regarding study design and assay are described in more detail in Figure 4.

Integrating more detailed methods information could strengthen the confidence of the association between the method and the KE and consequently support the evidence for the regulatory endpoint. In fact, the proposed ontologies support interoperability with existing frameworks and complement established documentation standards (e.g., OECD Harmonized Templates) which could increase the utility of AOPs. The value of ontological approaches lies in its ability to decompose all the information associated with a method and then reconnect them by utilizing specific defined relationships. In this ontological representation, the term “method” is not synonymous with “assay”; rather, it encompasses a broader set of information, including the study design and technical assay itself as well as the additional context, as illustrated in Figure 5. This comprehensive approach ensures that all aspects of the method are documented and interconnected. The value of the ontological approaches to information organization is exemplified by the relationships/associations that are defined on each arrow. By establishing relationships to help make connections among the structured data, more information can be stored in the AOP-KB in a manner that is both computable and useful for humans. Consistent with the broader AOP framework, the proposed structuring of method information is intended to bolster confidence in the application of AOPs which can in turn support higher level regulatory applications such as Integrated Approaches to Testing and Assessment, Defined Approaches, and weight-of-evidence evaluations, without prescribing how such workflows should be constructed.

 4. Stakeholder Engagement

Subsequent to identifying, prioritizing, and organizing relevant methods information, the Methods2AOP collaboration actively engaged with, and sought feedback from, various stakeholders. Through presentating at various international conferences and hosting a focused workshop, broad feedback was collected. For example, stakeholders from the regulatory or government sectors were invited to participate in a virtual workshop entitled “Roles of AOP framework in incorporating NAMs into the regulatory process” in April 2024. Presentations at this workshop described how AOPs could support the use of NAMs in regulatory decisions and introduced the ongoing work by the Methods2AOP collaboration to improve the methods documentation in the AOP framework.

These presentations were followed by breakout group discussions which yielded feedback on how the various participants currently used the AOP framework and NAMs, how the suggested inclusion of the methods information could influence the use of AOPs and NAMs, and what their perceived future needs and wishes were. Three individual questions were posed to participants; the key themes discussed for each question are shown in Table 2. Some of the common themes proposed for facilitating the regulatory acceptance were already part of the Methods2AOP collaboration, e.g., establish more explicit linkages of AOPs to the test methods used in MIEs/KEs, make validation status of methods used to measure KEs and develop AOPs immediately visible and transparent, etc.

Charge questions and themes discussed in the workshop.
# Questions Themes discussed
1 Describe scenarios where use of the AOP framework (assays connected to key events and key events to AOs) could make your workflow faster, easier, more transparent, or better in some other way

Better understanding which specific methods were used to support MIEs and KEs

Linking methods to KEs could bring confidence to both the method and the AOP

Help identify opportunities for read across, improve read across or to propose data that can support a waiver for in vivo studies

Support the design of NAMs batteries or IATAs

Identify orthogonal methods and help to build confidence that a KE is really part of the mechanism that causes the AO

OECD endorsed AOPs implies that the mechanistic links are more reliable than for non-endorsed

2 What information or resources would you need to feel confident in the methods/assays and their connections to KEs to be able to incorporate the AOP framework into your existing workflows?

Documentation for both the method and method linkage to the KE are critical in establishing trust in the AOP and the relationship between Method/KE and KE/AO

Information on method uncertainty to evaluate the evidence for KEs

Validation or readiness status of a method is important to build confidence in that the data is useful for regulatory applications

Quantitative aspects, e.g., what does a 10% alteration in a KE mean

Inclusion of KERs in the AOP framework is crucial

Need agreement on what we are validating the AOP against since human data is limited

3 Based on the proposal from the Methods2AOP, what information is needed to best support workflow needs to better use NAMs? In an ideal world, what would you like to see in the AOP-Wiki?

Development of AOP networks to support regulatory applications

Early engagements by regulators in AOP development, to enhance their use in regulatory applications

Evaluate the relevance of KEs withing the AOP, e.g., are there a “point of no return”, are some KEs more relevant than others for the AO?

Document the timeline of AOP endorsement, progress and when it is used

Ability to link to open databases, possible literatures to help gain information of the AOP and methods

Revise the review of AOPs to make the process easier and faster

Three newly proposed common themes were identified by workshop participants representing potential opportunities for future initiatives, implementations, and pilot application cases on specific disease adverse outcomes (e.g., liver injury, neurotoxicity, carcinogenicity). These three common themes were:

Identification of reference chemicals/causal agents;

Identification of inter-related AOPs (AOP networks), and inclusion of more information within AOPs for the relationship between KEs;

Evaluation of the usefulness of quantifiable data/points of departure (POD)/no observed effect concentrations (NOECs) in concentration-responses.

Based on the number of registrants and the captivating and constructive discussions achieved during this virtual workshop, similar events will be planned in the future. This will ensure stakeholder engagement with an open and active channel of communication between all parties involved.

 5. Key Takeaways and Impact

The main purpose of the Methods2AOP collaboration is to increase confidence in AOP knowledge and boost its uptake in being leveraged for data generation by better structuring the information of the methods used to characterise KEs and KERs. Ultimately this could enhance the acceptance of AOPs and NAMs for regulatory applications. There are regulatory processes (e.g., European Food Safety Authority (EFSA) Pesticides Peer Review) where academic publications are reviewed and taken into consideration; however, before data are accepted, they need to be considered reliable and relevant. This includes a step of assessing the risk of bias of the study, which is only possible when it is described with enough detail regarding the method, data interpretation and conclusions ( EFSA, 2010). Unfortunately, this aspect is often overlooked by the authors of such studies, limiting the impact of the work. To overcome such potential limitations within the AOP context, further efforts need to focus on developing standardized methods reporting and provide consistency in data interpretation, as this would bring more visibility to how approaches are performed and how they can inform on KEs.

The Methods2AOP collaboration worked to find balance in identifying and prioritizing sufficient and appropriate fields to document methods information. Focused on end-users in the AOP community, the Methods2AOP approach not only enhanced the clarity of method descriptions and proposed a tiered data structure; the Methods2AOP collaboration also engaged stakeholders to ensure needs could be met and the solutions proposed are of relevance. To overcome ambiguity of free-text fields and support standardized information structure to increase functionality and interpretability, we propose the establishment of standardized fields that leverage controlled vocabularies/ontologies alongside existing free-text options. This hybrid model promotes enhanced interoperability across AOPs, enabling the seamless reuse of methods and the identification of overlaps where a single method may inform on multiple KEs and AOPs. Such standardization could not only foster the adoption of robust methodologies for assessing KEs but also cultivate trust and confidence within the regulatory community and be applied across the entire AOP-Wiki (i.e., not only for methods information refinement). By ensuring that data are generated with fit-for-purpose design to support investigation of the AOP, integration of method documentation makes it easier to input, review, and search methods information as well as assess data generated by them.

The application of robust methods documentation not only fulfills the role of addressing regulatory requirements but also empowers industry scientists and AOP users seeking to leverage AOP-based lines of evidence to assist mechanistic investigation. AOP authors, and reviewers, must provide sufficient information to meet AOP users’ needs for AOP knowledge to be applicable in practice, regulatory or research. The academic research community may use AOPs in a different context than risk assessors and regulatory scientists who are seeking established mechanistic lines of evidence. For example, detailed methods information can serve as a resource beyond the AOP context to assist scientists in designing their experiments or interpet their data.

Furthermore, having methods information structure that separates assay details from study design also opens the door to increase AOP community membership and encourage method developers to engage and contribute appropriate details for robust documentation. Engaging test method developers will provide an opportunity to promote relevant test methods and/or to scan for “underserved” KEs (i.e., KEs that have been defined but for which a suitable test method has not yet been sufficiently developed or validated). Such a win-win scenario would both highlight new opportunities for test method developers and help fill knowledge gaps in the AOP-Wiki.

 6. Challenges and Future Needs

Several key challenges were identified by the Methods2AOP collaboration and will likely be encountered if changes are made to the existing AOP-Wiki content to integrate the data structure and detailed documentation proposed herein. The Methods2AOP collaboration focused on engaging a diverse consortium, identifying relevant metadata, organizing, annotating, ontologizing fields, and devising a proposed vision for implementation. Outstanding themes discussed by the collaboration but not yet addressed include addressing how ease of use would be paramount to the adoption of any redefined methods details. Easing accessibility (e.g., via the simple tiered data input proposal) and committing to education for both input and retrieval of methods information, including revison of the AOPs Developer’s handbook, will be critical to the successful implementation, impact, sustainability, and growth of the AOP community. AOP authors and methods developers will need training to fill out the right information as the success of developing a new infrastructure will be an iterative learning process. Only by engaging and amassing repeated real-world use will we be able to identify ambiguities in field interpretation, inconsistencies in ontology term selection or expose differences in contributor practices. Lastly, the benefit of embracing this new documentation will need to be well communicated to ensure participation and engagement.

This is similar to the continuing need for guidance for method developers embarking on validation studies and Test Guideline Development of new assays for use in the international regulatory context (OECD Webinar Series). 11 Although information fields for methods have been identified and suggested here, the Methods2AOP collaboration members recognize the need to develop more case studies to evaluate and likely refine required information to ensure an easy and user-friendly approach and establish applicability to various contexts of use. Furthermore, with the evergreen continuous evolution of ontologies it will be important to maintain versioning to support traceability over time in case of updated or deprecated terms. In this context, taking a cue from the standard fields identified in OECD Test Guidelines (e.g., OECD GD211-describing non-guideline in vitro test methods), 12 and OECD Harmonized Templates ( Carnesecchi et al., 2023) (e.g., OHT201), 13 may provide additional information and consistency on the way forward in the future progress of Methods2AOP.

In addition, prospective changes made to the AOP-KB and interface of the AOP-Wiki would require considerable effort and restructuring that will ultimately impact a variety of third-party tools. Thus, Methods2AOP seeks to suggest content, but not to conduct the implementation of this information at this stage. While the proposed revision of the current AOP-KB structure and Data Model relationships could be substantial upon integration of methods information, it will be a benefit that all tools draw from this centralized resource. If designed carefully, third-party tools could retain backward compatibility and be updated to include the new information in their own schedule. Future efforts should include alignment of AOP mechanistic data in collaboration with other programs that aim to improve FAIR standards such as the FAIR AOP Cluster working group ( Mortensen et al., 2025), the Environmental Health Language Collaborative (EHLC), 14 and the ELIXIR Toxicology Community workgroup. 15 Finally, continued engagement with stakeholders to provide resources and education about any changes and updates would be critical for seamless uptake. To summarize, the work executed by the Methods2AOP collaboration could benefit a diverse group of stakeholders in different ways with tangible foreseeable impacts ( Table 3).

Stakeholder groups that can benefit from the inclusion of structrued methods information in the AOP framework.
Stakeholder group Next steps Potential benefits
AOP Framework Management at OECD

Understand the benefits of a methods centered approach and appreciate the relevant Methods2AOP suggestions

Facilitate, support, and promote the adoption of Methods2AOP insights in the AOP Framework

Assist in the identification of possible funding sources to implement the Methods2AOP recommendations in the AOP-KB/Wiki

A more widely adopted and powerful AOP concept
AOP-KB/Wiki Funding Bodies

Understand the impact that Methods2AOP recommendations can have on the regulatory landscape

Acknowledge how a methods-centered AOP framework gains visibility and boosts market opportunities for an adequate ICT solution

Actively encourage ICT developers to submit proposals for a Methods2AOP-powered next version of the AOP-KB/Wiki

Increased return on investment in an improved AOP-Wiki
AOP-KB/Wiki ICT Developers

Understand how the Methods2AOP recommendations would translate into ICT requirements

Identify gaps between the current AOP-KB/Wiki situation and these requirements

Help other stakeholders (e.g. funders) to understand potential costs and prioritize various options and make a convincing case for implementation

Enhanced technical knowledge to support ICT development
Regulators (not directly involved in AOP application)

Facilitate the integration of AOPs for regulatory decisions

Bring case studies for a more methods-oriented AOP approach to bodies like OECD, etc.

Support AOP users within their jurisdictions in the phase-in of the new approach by empowering them to take decisions based on Methods2AOP-powered AOP knowledge

Benefit from a pioneering/early adopter role in the introduction of a methods-centered paradigm More tools to support regulatory decisions
AOP Users (including industry, academic, and regulatory stakeholders directly involved in AOP application)

Understand the way more tangible methods-AOP connections can increase their decision quality

Support, from a hands-on perspective, the introduction of a methods-oriented approach among peers

Have more confidence leveraging AOP knowledge

More robust data and higher confidence in using AOPs and mechanistic assay based lines of evidence to support decisions
AOP Developers/Authors

Understand the way minor extra effort (i.e. add methods information to their AOPs) can boost their AOP’s usefulness in regulatory circles

Show support for the paradigm change in gatekeeper circles

Adapt own AOP development strategy accordingly

Increased visibility and scientific/regulatory impact of their work
Method Developers/Vendors

Familiarize and have more insight to regulatory applications

Promote the method-centered AOP approach in the method developer community

Identify gaps for future assay development opportunities

Increased visibility and broader use of products
Animal Advocacy Groups

Understand the important role an increased focus on methods in the AOP framework can have on regulatory adoption of NAMs

Informed to discuss the Methods2AOP suggestions with policy makers

Support the paradigm change in replacing animal testing

Faster and more-widespread adoption of NAMs to replace animal use

ICT: Information, and Communication Technology.

 Disclaimer

This manuscript does not necessarily represent US Environmental Protection Agency (US EPA) policy or any other US agency. This manuscript has been subjected to review and approved for publication by the US EPA’s Office of Chemical Safety and Pollution Prevention. Mention of trade names or commercial products does not indicate endorsement by the US EPA. This manuscript was not reviewed by and does not reflect the view of 3M.

 Data Availability

The original contributions presented in the study are all contained in the article Figures and Tables.

In addition, the study refers to the following underlying data, which is a table providing the extended data used to support the analyses described in Section 2.3. The table is available here: https://doi.org/10.5281/zenodo.18400991, under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0). Further inquiries can be directed to the corresponding author ( Karmaus, A., Bisson, W., Fostel, J., Masci, A. M., Wittwehr, C., & Hogberg, H. (2026)).

 Acknowledgements

The authors would like to thank Shannon Bell, Steve Edwards, Stephen Ferguson, Alison Harrill, Birgitte Landesmann, Isabel Lea, Sofia Batista Leite, Miriam Midali, Natalia Reyero, Charles Schmitt, and Dan Villeneuve for their early contributions to this work and helpful discussions.

 References Ankley GT Bennett RS Erickson RJ : Adverse outcome pathways: A conceptual framework to support ecotoxicology research and risk assessment. Environ. Toxicol. Chem. 2010;29:730–741. 20821501 10.1002/etc.34 Bajard L Adamovsky O Audouze K : Application of aops to assist regulatory assessment of chemical risks - case studies, needs and recommendations. Environ. Res. 2023;217:114650. 36309218 10.1016/j.envres.2022.114650 PMC9850416 Bal-Price A Hogberg HT Crofton KM : Recommendation on test readiness criteria for new approach methods in toxicology: Exemplified for developmental neurotoxicity. ALTEX. 2018;35:306–352. 29485663 10.14573/altex.1712081 PMC6545888 Bisson W Arroyave W Atwood S : Framework for evaluating the level of evidence of carcinogenicity from mechanistic studies. The report on carcinogens handbook. The Toxicologist (Supplement to Toxicological Sciences);2024. Reference Source Carnesecchi E Langezaal I Browne P : Oecd harmonised template 201: Structuring and reporting mechanistic information to foster the integration of new approach methodologies for hazard and risk assessment of chemicals. Regul. Toxicol. Pharmacol. 2023;142:105426. 37277057 10.1016/j.yrtph.2023.105426 Carusi A Wittwehr C Whelan M : Addressing evidence needs in chemicals policy and regulation. Publications Office of the European Union;2022. 10.2760/9130 Carusi A Filipovska J Wittwehr C : Ciao: A living experiment in interdisciplinary large-scale collaboration facilitated by the adverse outcome pathway framework. Front. Public Health. 2023;11:1212544. 37637826 10.3389/fpubh.2023.1212544 PMC10449328 Clerbaux LA Filipovska J Nymark P : Beyond chemicals: Opportunities and challenges of integrating non-chemical stressors in adverse outcome pathways. ALTEX. 2024;41:233–247. 37980615 10.14573/altex.2307061 EFSA: Application of systematic review methodology to food and feed safety assessments to support decision making. EFSA J. 2010;8:1637. 10.2903/j.efsa.2010.1637 Hench G Edwards SW Lynn G : Integrated application and transformation of the adverse outcome pathway (aop) framework for modeling evidence and advancing new approach methodologies (nams). 2024a. 10.6084/m9.figshare.26390980.v1 Hench G Edwards SW Karshnik T : Enhancing the fairness (findability, accessibility, interoperability, and reusability) of the aop-wiki. 2024b. 10.6084/m9.figshare.26153326.v1 ICCVAM: A strategic roadmap for establishing new approaches to evaluate the safety of chemicals and medical products in the united states. 2018. ICCVAM: Validation, qualification, and regulatory acceptance of new approach methodologies. N. I. o. E. H. Sciences. 2024. 10.22427/NICEATM-2 Ives C Campia I Wang RL : Creating a structured aop knowledgebase via ontology-based annotations. Appl. In vitro Toxicol. 2017;3:298–311. 30057931 10.1089/aivt.2017.0017 PMC6060416 Karmaus A Bisson W Fostel J : Methods2AOP Extended Data: Supplementary Table of Structured Method Metadata and Ontology Terms.[Data Set]. Zenodo. 2026. 10.5281/zenodo.18400991 Krebs A Waldmann T Wilks MF : Template for the description of cell-based toxicological test methods to allow evaluation and regulatory use of the data. ALTEX. 2019;36:682–699. 31658359 10.14573/altex.1909271 Krebs A Vugt-Lussenburg BMA van Waldmann T : The eu-toxrisk method documentation, data processing and chemical testing pipeline for the regulatory use of new approach methods. Arch. Toxicol. 2020;94:2435–2461. 32632539 10.1007/s00204-020-02802-6 PMC7367925 Leite SB Brooke M Carusi A : Promoting reusable and open methods and protocols (pro-map): Draft recommendations to improve methodological clarity in life sciences publications. OSFPreprints. 2024. Reference Source Mortensen HM Gromelski M Hench G : The fair aop roadmap for 2025: Advancing findability, accessibility, interoperability, and re-usability of adverse outcome pathways. Comput. Toxicol. 2025;35:100368. 40950714 10.1016/j.comtox.2025.100368 PMC12425154 O’Brien JM Yauk CL : Introducing aop reports: Collaborative review and publication of adverse outcome pathways. Environ. Mol. Mutagen. 2022;63:116–117. 35435297 10.1002/em.22481 OECD Publishing: Initial recommendations on evaluation of data from the developmental neurotoxicity (dnt) in-vitro testing battery. E. DIRECTORATE and C. A. B. COMMITTEE;2023a. OECD Publishing: Defined approaches on skin sensitisation. OECD;2023b. Pitzer EM Shafer TJ Herr DW : Identification of neurotoxicology (nt)/developmental neurotoxicology (dnt) adverse outcome pathways and key event linkages with in vitro dnt screening assays. Neurotoxicology. 2023;99:184–194. 37866692 10.1016/j.neuro.2023.10.007 PMC12117389 Sanz-Serrano J Callewaert E De Boever S : Chemical-induced liver cancer: An adverse outcome pathway perspective. Expert Opin. Drug Saf. 2024;23:425–438. 38430529 10.1080/14740338.2024.2326479 Villeneuve DL Crump D Garcia-Reyero N : Adverse outcome pathway (aop) development i: Strategies and principles. Toxicol. Sci. 2014;142:312–320. 25466378 10.1093/toxsci/kfu199 PMC4318923 Zalm AJ van der Barroso J Browne P : A framework for establishing scientific confidence in new approach methodologies. Arch. Toxicol. 2022;96:2865–2879. 35987941 10.1007/s00204-022-03365-4 PMC9525335

https://aopwiki.org/info_pages/10 [accessed May 2, 2025]

https://aopwiki.org [accessed April 17, 2025]

https://aopkb.oecd.org/index.html [accessed April 17, 2025]

https://www.niehs.nih.gov/research/programs/ehlc/events [accessed May 5, 2025]

https://elixir-europe.org/communities/toxicology [accessed April 17, 2025]

https://one.oecd.org/document/ENV/JM/MONO(2013)6/en/pdf [accessed April 17, 2025]

https://aopwiki.org/handbooks/5 [accessed April 17, 2025]

https://www.niehs.nih.gov/research/programs/ehlc/shared-language [accessed April 18, 2025]

http://obofoundry.org [accessed April 18, 2025]

http://go-fair.org [accessed April 18, 2025]

https://www.oecd.org/en/topics/sub-issues/testing-of-chemicals/webinar-series-on-testing-and-assessment-methodologies.html [accessed April 17, 2025]

https://www.oecd.org/en/publications/guidance-document-for-describing-non-guideline-in-vitro-test-methods_9789264274730-en.html [accessed April 17, 2025]

https://www.oecd.org/en/topics/assessment-of-chemicals/harmonised-templates-intermediate-effects.html [accessed April 17, 2025]

https://www.niehs.nih.gov/research/programs/ehlc [accessed April 17, 2025]

https://elixir-europe.org/communities/toxicology [accessed April 17, 2025]

 10.5256/f1000research.195945.r462675 Reviewer response for version 2 Tcheremenskaia Olga 1 Referee https://orcid.org/0000-0002-5029-3484 Insituto Superiore di Sanità (ISS), Rome, Lazio, Italy

Competing interests: No competing interests were disclosed.

 24 3 2026 Copyright: © 2026 Tcheremenskaia O 2026 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve

General comments

The manuscript analyses how the current free-text field “How It Is Measured or Detected,” used to describe Key Events (KEs) in the AOP-Wiki, could be refined into a more structured documentation format for methods information. This is highly valuable work for both the scientific and regulatory communities, as it directly addresses current limitations in transparency, interoperability, and regulatory uptake of methods information within the AOP framework.

I fully agree with the comments made by the two previous reviewers and with the revisions implemented by the authors. The supplementary materials are particularly useful, as they provide detailed examples of how selected  in vitro,  in vivo, and  in chemico methods can be mapped to ontology fields. The authors have done an excellent job identifying and reusing 15 existing ontologies (e.g. OBI, GO, ChEBI), in line with established best practices for semantic interoperability and the FAIR principles.

Overall, the work is timely, well-motivated, and highly relevant to the further integration of New Approach Methodologies (NAMs) into regulatory decision-making via the AOP framework.

Major comments

1. Mapping to OHT (for instance hCLAT to OHT 66‑1)

While the supplementary material provides excellent practical examples, such as the human Cell Line Activation Test (hCLAT) annotated using externally identified ontology fields, I would ecommend that the authors additionally map this example (or other two skin sensitisation related examples) directly to OECD Harmonised Template (OHT) 66‑1 (Skin Sensitisation). OHT 66‑1 is an established endpoint‑oriented template that has been updated to accommodate non‑apical observations and to link these to Adverse Outcome Pathways (AOPs)

I consider it important to evaluate how the proposed Methods2AOP fields align with the structured fields of OHT 66‑1. Such a mapping would provide a clear regulatory pathway and facilitate the direct use of Methods2AOP‑annotated methods in regulatory dossiers and Defined Approaches on skin sensitisation.

2. OHT terminology versus external ontologies

The use of external ontologies is highly useful and clearly strengthens semantic interoperability. However, it may be helpful first to evaluate how existing endpoint‑oriented OHT templates (which are not exclusively mechanistic, unlike OHT 201) already accommodate test system, assay, and observation descriptions. A key advantage is that OHTs include suggested terminology and picklists that are already familiar to the regulatory community.

For example, for cell lines “Cell line name: THP‑1 (CLO:0009348)” in OHT 66‑1, field number 44, the picklist entries such as: and “Details of test system: THP‑1 cell line [442E]; keratinocytes transgenic [442D]; THP‑G8 [442E]; U‑937 [442E]; other.”

Leveraging OHT terminology and picklists could help make test system and assay descriptions more readily usable in Defined Approaches (DAs), Test Guidelines (TGs), and validation studies conducted in line with OECD Guidance Document 34 (GD34). Looking forward, as additional OHTs are updated to capture non‑apical endpoints, it would be beneficial if the Methods2AOP framework could always foresee mapping to OHT terminology where possible.

Minor comments and suggestions

1. Glossary of core terms

To improve clarity, the manuscript would benefit from a short glossary that explicitly defines the core terms used throughout the Methods2AOP framework. In particular, it would be helpful to clearly distinguish between:

Test System (the biological model),

Assay (the specific technical process), and

Observation (the composite entity encompassing study design and measured effects).

 Although these definitions are available in the supplemental file “Suggested Ontologies and Fields – Term Definition,” briefly summarizing them in the main text would enhance accessibility for readers less familiar with the terminology.​

2. Feasibility of revising existing AOPs

I agree with the authors that refining the current “How It Is Measured or Detected” free-text field into structured documentation will be a resource‑intensive process. This will likely require revising the AOP Developers’ Handbook and updating the guidance for AOP Coaches. It would be useful if the authors could explicitly discuss whether the proposed approach is intended only for newly developed AOPs or there is a strategy to address the existing AOPs in AOP‑Wiki, some of which remain incomplete, outdated, and not yet endorsed.

3. Potential use of large language models (LLMs)

Given the considerable manual curation effort required, it might be worth briefly exploring the potential use of large language models (LLMs) to support and speed up the standardization of the existing free‑text “How It Is Measured or Detected” fields, under appropriate human oversight. A short discussion if the Methods2AOP structure could be used as a target schema for semi‑automated text-to-structure mapping would be of interest to the community.

4. Field for the linked Key Events

Although Key Event Relationships (KERs) are not the primary focus of this work, the authors could consider introducing a dedicated field, where to indicate if the KE is linked to another KE.

5. Possibility of prioritization of methods approaching regulatory uptake

The manuscript rightly notes that any future changes to the AOP‑KB and the AOP‑Wiki interface would require considerable effort and would affect multiple third‑party tools. In the short term, it might therefore be more feasible to keep the current free-text field in the AOP‑Wiki but provide concise, practical guidance on how to standardize this free-text when a test is intended for regulatory purposes, such as Defined Approaches (DAs) or for evaluation under GD34 as a Test Guideline candidate. Once the ongoing update of GD34 is finalized, the Methods2AOP information could potentially be mapped directly to the readiness‑criteria templates in GD34, thereby helping to bridge the gap between AOP development, method documentation, and formal regulatory validation.

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

Not applicable

Are all the source data underlying the results available to ensure full reproducibility?

No source data required

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Yes

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

Computational toxicology, regulatory toxicology, NAMs, databases

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.

 10.5256/f1000research.195945.r464632 Reviewer response for version 2 Willie Peijnenburg 1 Referee National Institute for Public Health and Environment (RIVM), Bilthoven, The Netherlands

Competing interests: No competing interests were disclosed.

 10 3 2026 Copyright: © 2026 Willie P 2026 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve-with-reservations

General assessment

This manuscript presents the Methods2AOP initiative, a collaborative effort aimed at improving the integration of test method information (particularly New Approach Methods (NAMs)) within the Adverse Outcome Pathway (AOP) framework and its central knowledge repository, the AOP-Wiki. The authors argue that although AOPs provide a structured framework linking molecular initiating events and key biological processes to adverse outcomes, the documentation of experimental methods used to measure these events remains insufficiently structured. As currently implemented, AOP-Wiki entries often describe measurement approaches in a free-text field (“How it is measured or detected”), which limits interoperability, reproducibility, and regulatory usability of the underlying data. To address this limitation, the manuscript proposes a conceptual framework for structured documentation of test methods using standardized metadata fields and ontology-based annotations. The proposed Methods2AOP schema introduces structured entities such as “method,” “assay,” and “observation,” and organizes method information through hierarchical tiers and controlled vocabularies to facilitate FAIR (findable, accessible, interoperable, and reusable) data integration. The authors further outline how these structured descriptors could be linked to key events (KEs) within the AOP knowledgebase, thereby strengthening the connection between mechanistic toxicology data and regulatory decision-making frameworks.

The topic is timely and highly relevant to the ongoing transition toward mechanism-based toxicology and the increasing use of NAMs in regulatory science. The manuscript is clearly written and provides a thorough overview of the collaborative effort, including stakeholder engagement, methodological prioritization of metadata fields, and conceptual implementation within the AOP knowledge infrastructure. Overall, the work represents a valuable conceptual contribution to the development of the AOP ecosystem. However, some aspects of the proposed framework would benefit from further clarification, particularly regarding practical implementation, conceptual boundaries between core entities, and long-term governance of ontology-based metadata.

Major comments

1 - Conceptual scope and intended role of structured “Observation” entities

One of the central elements introduced in the proposed framework is the “Observation” entity, which is described as representing evidence derived from a test method and linked to key events within an AOP. The manuscript explains that this entity contains information about study design and measured effects, serving as supporting evidence for a KE. However, the practical scope of this entity remains somewhat ambiguous. It is not entirely clear whether the Observation entity is intended primarily as a metadata container describing how a KE is measured, or whether it is intended to represent potentially large numbers of individual study-level evidence instances generated across multiple studies and datasets. This distinction has important implications for database scale, curation burden, and long-term maintainability of the AOP knowledgebase.

Clarifying whether the Observation concept represents a bounded metadata structure or a potentially extensible repository of study-level evidence would significantly strengthen the conceptual clarity of the framework. Additionally, it would be useful to explicitly discuss whether the framework is designed to capture contradictory or null findings, which are important for avoiding confirmation bias in evidence synthesis.

2 - Conceptual boundaries between “method,” “assay,” and related entities

The manuscript appropriately distinguishes between the broader concept of a “method” and the more specific technical “assay.” However, the practical distinctions between the entities used in the conceptual model (method, assay, observation, study design, and effect) are sometimes difficult to interpret when reading the text. While the conceptual diagram provided in the manuscript illustrates the relationships among these elements, a concise glossary or table defining each entity would significantly improve readability and conceptual clarity. In particular, readers would benefit from a practical example illustrating how information currently provided in the free-text “How it is measured or detected” field would map onto the proposed structured fields. Providing a concrete example demonstrating how a specific assay measuring a particular KE would be represented within the Methods2AOP schema would make the proposed framework more accessible and easier to evaluate.

3 - Ontology use and long-term maintenance considerations

The use of ontologies to standardize metadata and ensure interoperability represents one of the major strengths of the proposed framework. The manuscript highlights how ontology-based annotations can replace unstructured free-text descriptions and facilitate more robust data integration across different evidence streams. However, ontologies evolve over time as terms are updated, replaced, or deprecated. The manuscript briefly acknowledges the need for maintaining ontology versioning, but this issue could be discussed more explicitly. In particular, long-term traceability of annotations and compatibility with future ontology updates represent important challenges for any ontology-driven information infrastructure. A brief discussion of potential governance strategies—such as ontology version tracking, term mapping, or update mechanisms within the AOP knowledgebase—would strengthen the proposed framework and demonstrate awareness of these practical considerations.

4 - Practical implementation and community adoption

The manuscript appropriately emphasizes that the Methods2AOP collaboration proposes a conceptual schema rather than an immediate implementation within the AOP knowledgebase. Nonetheless, several aspects of real-world adoption remain relatively abstract. For example, implementing structured metadata fields within the AOP-Wiki would require substantial changes to the current database schema, user interface, and curation workflows. The manuscript acknowledges these challenges but could provide a more detailed discussion of possible implementation pathways, including pilot case studies or phased integration strategies.

The authors note that early applications of the framework will likely represent an iterative learning process as contributors gain experience with the system. Explicitly framing the proposed approach as an evolving community-driven infrastructure could help set appropriate expectations for the adoption and refinement of the framework.

Minor comments

The manuscript is generally well written and logically structured. Nevertheless, several minor improvements could further enhance clarity.

In the introduction, the discussion of the current AOP infrastructure could be slightly streamlined to focus more directly on the specific limitations that motivated the Methods2AOP initiative. Some sections describing stakeholder engagement and workshop outcomes could also be condensed without losing important content.

Several figures provide useful conceptual illustrations of the proposed framework, particularly the diagrams illustrating the hierarchical structure of method metadata and the conceptual integration with the AOP knowledgebase. The captions could be slightly expanded to ensure that readers unfamiliar with the AOP framework can fully interpret the diagrams without consulting the main text.

Finally, a short summary table describing the main components of the Methods2AOP schema would help readers quickly understand the proposed structure.

Assessment of methodological transparency and reproducibility

The manuscript presents a conceptual framework rather than an experimental study. Consequently, traditional criteria such as statistical analysis or experimental reproducibility are not directly applicable. Nevertheless, the authors provide sufficient detail regarding the development of the framework, including stakeholder consultations, prioritization of metadata fields, and ontology selection. The supplementary materials and linked datasets further support transparency of the proposed metadata structure. Overall, the methodological description is adequate to allow other researchers or infrastructure developers to evaluate and potentially implement similar approaches.

Data availability and reproducibility

The manuscript provides access to supplementary datasets containing extended metadata tables and ontology references through a publicly accessible repository. This is consistent with the open science principles promoted by the journal. The availability of these resources should allow readers to explore the proposed metadata structure and evaluate its potential implementation in other knowledge management systems.

Conclusion

This manuscript presents a well-conceived and timely proposal for improving the integration of test method metadata within the Adverse Outcome Pathway framework. By introducing structured, ontology-based descriptions of methods and linking them explicitly to key events, the Methods2AOP collaboration aims to enhance the transparency, interoperability, and regulatory relevance of AOP knowledge. The proposed framework represents a valuable conceptual contribution to the development of the AOP ecosystem and the broader transition toward mechanistic toxicology and NAM-based risk assessment. However, the manuscript would benefit from additional clarification regarding the conceptual scope of the Observation entity, clearer definitions of core entities within the proposed schema, and a more detailed discussion of practical implementation challenges. Addressing these points would further strengthen the manuscript and help readers better understand the potential impact and feasibility of the proposed framework.

Final recommendation: approved with reservations

The manuscript presents a valuable conceptual framework and contributes meaningfully to ongoing efforts to improve the interoperability and regulatory applicability of AOP-based toxicology. However, several aspects of the proposed structure would benefit from additional clarification and discussion before the framework can be fully evaluated by the community. With minor revisions addressing the issues outlined above, the manuscript will represent a useful and informative contribution to the literature on AOP infrastructure development and NAM integration.

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

Partly

Are all the source data underlying the results available to ensure full reproducibility?

Yes

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Yes

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

AI, mechanistic toxicology, modelling

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

 10.5256/f1000research.195945.r459591 Reviewer response for version 2 Schaffert Alexandra 1 Referee https://orcid.org/0000-0001-8159-1292 Tampere University, Tampere, Finland

Competing interests: No competing interests were disclosed.

 5 3 2026 Copyright: © 2026 Schaffert A 2026 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve-with-reservations

Thank you for the careful revisions and for clarifying the previous comments. I appreciate the added worked examples in the supplementary material, the clearer discussion of ontology lifecycle/versioning, and the explicit acknowledgement that practical uptake of the framework will be an iterative learning process shaped by repeated real-world use. These additions make the proposal more concrete and strengthen the manuscript overall.

My remaining comment is mainly conceptual, and I raise it because I think the current wording may still be difficult for readers to understand in relation to the existing AOP-Wiki structure.

In the revised manuscript, the Methods2AOP proposal is still clearly framed as method-centric and KE-centric. The paper states that the intention is to “add to and refine the existing infrastructure, not to re-build established interfaces,” and it emphasizes that the practical problem being addressed is the current KE-level free-text “methods of measurement” field. The manuscript further proposes an “Observation” field linked to KEs, defined as comprising study design and measured effects, with additional linked attributes such as stressor, test system, assay endpoint, documentation, and measurement.

At the same time, the revised wording now describes Observation as a diversity of evidence for any method yielding data to support a KE, and Figure 5 uses “has evidence / is evidence for” language for the relationship between observation and the claim it supports. The manuscript also notes that preliminary integration of Observation has already been presented in previous KER-centric evidence modeling efforts.

This is where I still found the text somewhat difficult to understand. The use of the word “evidence” in the KE-linked Observation context naturally evokes the established AOP understanding of weight of evidence, which is usually centered on KERs rather than on KEs. The manuscript itself notes earlier that the AOP framework focuses on evidence for the linkages between critical KEs, while Methods2AOP is presented as a way to better structure documentation of methods used to measure KEs. Because of this, I think some readers may be uncertain whether Observation is intended primarily as: (1) a KE-linked empirical observation/measurement object that helps structure how a KE is assessed in experiments, or (2) a broader evidence object that is also meant to feed into KER-level evidence assembly and weight-of-evidence considerations.

For me, the additional context from prior EMOD work helps resolve part of this ambiguity. In the EMOD prototype, Assay structures methods-of-measurement information, Observation holds empirically obtained KE measurements, and Evidence is explicitly a KER-level object that interlinks upstream and downstream observations linked to the respective KEs of the KER. With that broader context, the Methods2AOP proposal becomes easier to understand as a KE-level structuring layer for methods, assay, study design, and observation information, which could later support more transparent evidence assembly at the KER level. However, in the manuscript itself, that bridge remains mostly implicit.

I therefore think the manuscript would benefit from a short paragraph making this boundary and relation to KERs more explicit. My understanding after reading the revision is that Methods2AOP is not proposing to replace the existing KER evidence evaluation structure of the AOP-Wiki, but rather to strengthen the structured documentation of KE-linked methods and observations so that these can better support interpretation, comparability, and potentially downstream evidence assembly. If that is indeed the intent, stating it directly would help substantially. It would also help distinguish between the current KE-page role (“how a KE is measured”) and the KER-page role (“how evidence for the linkage is evaluated”), which remain conceptually distinct in the current AOP-Wiki.

Related to this, I still found the mapping between the conceptual entities in Figure 5 and the operationalized field groups in Figures 3-4 / the supplement somewhat implicit. My current reading is that Observation is a composite entity rather than a simple standalone field, and that its parts are represented in the supplementary file through grouped fields related to study design (Study Documentation, Study Subject Type, Study Husbandry/Care), assay/test system (Assay Documentation, Assay Characteristic), measurement (Assay Measurement). However, this is something the reader still has to infer, and it is still not clear where “effect” is represented, and how these pieces relate to the current KE-level free-text “How it is measured or detected” (and if relevant KER-level evidence) field. The worked examples are useful, but they do not yet fully resolve this conceptual mapping.

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

Not applicable

Are all the source data underlying the results available to ensure full reproducibility?

No source data required

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Yes

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

Mechanistic toxicology, computational toxicology, AOP development

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

 10.5256/f1000research.190643.r440513 Reviewer response for version 1 Schaffert Alexandra 1 Referee https://orcid.org/0000-0001-8159-1292 Tampere University, Tampere, Finland

Competing interests: No competing interests were disclosed.

 6 1 2026 Copyright: © 2026 Schaffert A 2026 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve-with-reservations

The manuscript proposes a structured, ontology-supported framework (Methods2AOP) to improve how test method information is captured and linked to KEs in the AOP-Wiki, moving beyond the current free-text “How it is measured or detected” field. The approach is timely and well aligned with needs for interoperability, comparability, and reuse of method metadata across different evidence streams. A major potential strength of the proposed framework is its capacity to act as a machine-readable interface between toxicological data (e.g. NAM, high-throughput, regulatory, or academic datasets) and existing AOP content. By linking structured, ontology-backed method descriptors directly to KEs, the framework could enable more systematic identification of which experimental results inform which KEs and AOPs, thereby supporting AOP-guided interpretation, evidence assembly, and comparison across studies.

The conceptual model and stakeholder-informed prioritization are a strong foundation, the main opportunities for strengthening the manuscript relate to clarifying intended scope and practical feasibility.

Comments:

1. The manuscript introduces an “Observation” entity comprising study design and measured effects and describes this as supporting evidence for KEs. As written, it is not fully clear whether Observation is intended primarily as a bounded metadata structure to contextualize a method-KE linkage, or whether it is intended to capture potentially large volumes of study-level evidence instances over time. Clarifying this intent would help readers understand the anticipated scale, curation burden, and the role of external links versus internal storage. If evidence instances are in scope, it may be helpful to briefly acknowledge how volume and redundancy could be managed conceptually. Additionally, the framework uses “evidence for” language that may be interpreted as primarily confirmatory. In practice, capturing contradictory or null findings is important to avoid systematic positivity bias and to support weight-of-evidence thinking. The authors may wish to clarify whether contradictory/inconclusive observations are intended to be captured and, if so, indicate at a conceptual level how this would be represented.

2. While the manuscript distinguishes “method” from “assay” and provides a functional description of “observation,” the practical boundaries between “method,” “assay,” “observation,” “study design,” and “effect” remain somewhat difficult to interpret as a reader, particularly in terms of what would constitute a single record of each type and how they relate in typical use. A short set of definitions (or a small glossary/box) clarifying what each term is intended to represent in this framework would be very useful. In addition, consider including concrete examples that demonstrate how these entities are put into practice, e.g., for a single KE, showing a typical current free-text “How it is measured or detected” description, how this information (and any additional relevant context) would map onto the proposed structured fields, and how the different entities (method, assay, observation, study design, effect) relate to each other within that example.

3. The use of ontologies is a major strength for interoperability, however, ontologies evolve (terms can be updated, obsoleted, or replaced). It would strengthen the manuscript to briefly note this issue and potential traceability over time.

4. The manuscript notes that case studies will be important for evaluating and refining the proposed framework. In this context, it may be helpful to clarify that the main value of such exercises lies not only in illustrating applicability, but in revealing practical issues that emerge when the framework is used by different contributors and across different examples. Experience with similar community-driven, ontology-based frameworks suggests that aspects such as field interpretation, level of detail, ontology term selection, and internal consistency are often only fully understood through repeated, real-world use rather than through a single demonstrative case. Briefly acknowledging that early application of the framework is likely to be an iterative learning process, informing clarification of guidance and incremental refinement of the structure over time, could help set appropriate expectations for future development.

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

Not applicable

Are all the source data underlying the results available to ensure full reproducibility?

No source data required

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Yes

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

Mechanistic toxicology, computational toxicology, AOP development

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

Wittwehr Clemens Systems Toxicology, European Commission Joint Research Centre Ispra, Ispra, Italy

Competing interests: None.

 28 1 2026

General Response to Reviewer 2:

We thank the reviewer for their thorough, thoughtful, and constructive review. We particularly appreciate the reviewer’s recognition of the timeliness and potential value of the proposed Methods2AOP framework, especially its alignment with current needs for interoperability, comparability, and machine-readable linkage between test methods, evidence streams, and AOP content. The reviewer’s detailed comments and suggestions regarding scope clarification, conceptual boundaries, ontology use, and practical implementation have been very helpful and have guided us in strengthening the manuscript accordingly.

Main comments

1. Clarification of the intended scope and role of the “Observation” entity

Reviewer 2 notes that it is not fully clear whether the proposed Observation entity is intended to be:

a bounded metadata structure that contextualises a method–KE linkage, or

a mechanism to capture potentially large numbers of study-level evidence instances over time.

 They highlight that this distinction has implications for:

anticipated scale,

curation burden,

and the balance between internal representation versus external linking.

 They also raise an important conceptual point about evidence polarity. The current wording may be read as focusing on confirmatory evidence, whereas capturing contradictory, null, or inconclusive findings is essential to avoid positivity bias and to support weight-of-evidence thinking. They ask for clarification on whether such observations are in scope and, if so, how they would be represented at a conceptual level.

Comment 1 Response:

We thank the reviewer for highlighting the need to be clear regarding the intention for the “Observation” entity. This is a central concept and we have made efforts to refine our description. In brief, the observation is the evidence supporting key event as derived from a test method. As the reviewer notes, this could be a large number of evidence and for any test method yielding data to support a KE, ample test method and assay data should be provided.

Additions to manuscript shown in italics below:

Section 3, Paragraph 2: “The proposed conceptual model, includes an “Observation” field defined as comprising information on study design and the measured effects (for which detailed inputs are summarized by the tiered data and detailed fields in Figures 3 & 4). Preliminary integration of “Observation” is already under development in the AOP-KB and has been presented as part of previous KER-centric Evidence Modeling prototyping efforts (Hench et al., 2024a; O'Brien and Yauk, 2022). Challenges in integrating this field lie in the complexity of its parts (e.g., study design and effect which in turn also have their own ontological components) as “Observation” would include a diversity of evidence for any method that yields data to support a KE. By defining ontological terms appropriate per field, the Methods2AOP collaboration seeks to bound the diversity of metadata and simplify curation burden.”

2. Need for clearer boundaries and definitions between core entities

While acknowledging that the manuscript distinguishes between “method,” “assay,” and “observation,” Reviewer 2 finds that the practical boundaries between: method, assay, observation, study design, and effect remain difficult to interpret for a reader. In particular, it is not always clear:

what constitutes a single record of each entity,

how these entities relate to one another in typical use.

 They suggest that this could be addressed through:

a short set of definitions or a small glossary or boxed text, and

a concrete example showing, for a single KE, how a current free-text “How it is measured or detected” description would map onto the proposed structured entities and fields, and how those entities interrelate

 Comment 2 Response:

Thank you for highlighting the need for a more concrete illustration of how each entity would be documented. We have revised the supplementary file where the granular definitions were provided to include some examples of documentation. Please refer to the newly updated supplementary file (Karmaus et al., 2025).

3. Ontology evolution and long-term traceability

Reviewer 2 strongly supports the use of ontologies for interoperability, but notes that ontologies are not static. Terms may be updated, deprecated, or replaced over time. They suggest that briefly acknowledging this reality and the need for traceability over time would strengthen the manuscript and demonstrate awareness of long-term maintenance considerations

Comment 3 Response:

This is a great point that we didn’t directly address but are in complete agreement. We have added a sentence to the future needs section to make this more explicit.

Additions to manuscript summarized in italics below:

Section 6, Paragraph 2: “Although information fields for methods have been identified and suggested here, the Methods2AOP collaboration members recognize the need to develop more case studies to evaluate and likely refine required information to ensure an easy and user-friendly approach and establish applicability to various contexts of use. Furthermore, with the evergreen continuous evolution of ontologies it will be important to maintain versioning to support traceability over time in case of updated or deprecated terms.”

4. Case studies and iterative refinement of the framework

Reviewer 2 agrees that case studies will be important, but suggests going one step further in how they are framed. Beyond illustration, their main value lies in:

revealing ambiguities in field interpretation,

identifying inconsistencies in ontology term selection,

exposing differences in contributor practices.

 Drawing on experience from other community-driven, ontology-based frameworks, they note that many practical issues only emerge through repeated real-world use, not from a single demonstrative example. They therefore suggest explicitly acknowledging that early application of Methods2AOP will be an iterative learning process, informing refinement of guidance and incremental evolution of the framework over time.

Comment 4 Response:

This is another point we absolutely agree with but didn’t explicitly state in the manuscript. We have made an addition in the manuscript, in italics below:

Section 6, Paragraph 1: “ Easing accessibility (e.g., via the simple tiered data input proposal) and committing to education for both input and retrieval of methods information, including revision of the AOPs Developer’s handbook, will be critical to the successful implementation, impact, sustainability, and growth of the AOP community. AOP authors and methods developers will need training to fill out the right information as the success of developing a new infrastructure will be an iterative learning process. Only by engaging and amassing repeated real-world use will we be able to identify ambiguities in field interpretation, inconsistencies in ontology term selection or expose differences in contributor practices. Lastly, and the benefit of embracing this new documentation will need to be well communicated to ensure participation and engagement.”

 10.5256/f1000research.190643.r440512 Reviewer response for version 1 Choi Jinhee 1 Referee Dufourcq Sekatcheff Elizabeth 2 Co-referee University of Seoul, Seoul, South Korea Department of Risk Assessment, Agence Nationale de Sécurité Sanitaire de l'Alimentation, de l'Environnement et du Travail, Maisons-Alfort, France

Competing interests: No competing interests were disclosed.

 31 12 2025 Copyright: © 2025 Choi J and Dufourcq Sekatcheff E 2025 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve-with-reservations

A) 

This manuscript, based on the activities of the Methods2AOP collaboration, proposes an approach to integrate and manage information on methods used to measure or detect Key Events in the AOP-Wiki by moving beyond the current free-text–driven format toward standardized metadata and a structured scheme supported by ontologies or controlled vocabularies. Overall, the attempt is timely and meaningful in that it aims to improve the applicability of NAMs and the reusability of data in AOP-based assessments; however, several core elements require further strengthening in terms of practical feasibility and regulatory and community use.

Major comments

1. The core of the manuscript is the proposal for structuring and standardizing method information, and readers will want to understand how this would function in practice within the AOP-KB. However, compared with the conceptual design and the presentation of candidate fields, the manuscript provides too few concrete application examples, which weakens the persuasiveness of the proposal. In particular, the manuscript would benefit from a more intuitive illustration of how the proposed metadata would be populated for a specific Key Event and how this would differ from, and improve upon, the current “How it is measured or detected” free-text field.

2. Standardization of method information is already supported by multiple existing frameworks, including OECD test guidelines, guidance documents, and harmonised templates. Therefore, the manuscript should more clearly explain how this proposal relates to those existing standards and what additional problems it is intended to solve. While related systems are mentioned, further clarity is needed regarding why existing standards are insufficient and what role Methods2AOP is expected to play within the AOP-KB.

3. Structured information systems often encounter challenges not at the design stage but during implementation and maintenance. Key issues include validation and quality control of entries, governance of terminology resources, and ensuring consistency across contributors, yet these aspects are not sufficiently addressed. Even if implementation is considered out of scope, the manuscript should provide at least a minimal operational and governance perspective, including a clearer view of which stakeholders might take on which roles in practice.

4. The manuscript claims that the proposed structure can cover multiple study types and evidence streams, but the narrative and examples read as being weighted toward specific contexts such as in vitro. In addition, if regulatory relevance is a central motivation, the manuscript should more clearly articulate how structured method information could connect to concrete regulatory workflows, such as IATAs, Defined Approaches, weight-of-evidence evaluations, read-across justifications, and NGRA. Because the nature of required metadata can differ substantially across evidence streams, further justification is needed to support both the breadth and regulatory applicability of the proposal.

Minor comments

1. There are instances where the collaboration name and related abbreviations are not fully consistent throughout the manuscript. For example, in Section 2.2 the text uses “Methods2AOP” but also contains a sentence using “Method2AOP,” and the Note under Table 1 uses a spaced form such as “Methods 2AOP.”

2. Several apparent spelling errors are present in the text. For example, “interpet,” “presentating,” “revison,” and “structrued” should be corrected.

B)

In this paper, the authors address a crucial topic on AOPs:  how to increase confidence in AOPs for regulatory purposes by integrating test methods in a standardized format in the AOP framework. The authors present here the result of a collaborative workshop and other initatives (EHLC, FAIR AOP cluster, ELIXIR), where they suggest an ontology based approach to adequately define method information needed.

I have a few comments/questions on the choice of identified criteria that define test methods and on the practical implementation/developer guidance

1.  In section 2.2, I was wondering whether it would be feasible to define information fields based on existing framework that establish test methods validation criteria (e.g., OECD GD 34) ? it would help to raise confidence in non standardized test methods in the AOP framework (e.g. literature) rather than having a separate section for non validated test methods, which I am concerned regulators will tend to overlook. 

2. in section 2.3, it was not clear to me what to categorize as 'study type' as for eg, ecological study type can be performed with in vivo/in silico methods. As they are not mutually exclusive, how do you expect developers to fill information in the right section ?

3. Overall, how do you envision the implementation of test methods in the AOPwiki database when there is no consensus yet on the terminology of KEs ? (numerous terminologies exist for e.g., oxidative stress, the term 'umbrella key event' was put forward to regroup several key events falling under the same scope...) I think this should be a problem to tackle first and the AOPwiki database would need cleaning to avoid duplication of KEs, accurate definitions of KE and associated test methods to enable correct linkage between AOPs and generation of AOP networks.

4. Overall, do you plan to engage with european regulatory agencies (e.g., EFSA, ECHA) or member states authorities in this project ? I think it would be a great opportunity to ensure all regulatory needs are taken into account

Overall, this is a great project and I'm looking forward to see the outcomes.

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

Not applicable

Are all the source data underlying the results available to ensure full reproducibility?

No source data required

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Yes

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

Environmental toxicology

We confirm that we have read this submission and believe that we have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however we have significant reservations, as outlined above.

Wittwehr Clemens Systems Toxicology, European Commission Joint Research Centre Ispra, Ispra, Italy

Competing interests: None.

 28 1 2026

General Response to Reviewer 1:

We thank the reviewer for their thoughtful and constructive comments regarding the impact of Methods2AOP relative to existing standards, the operational and governance perspective, and the applicability of the proposal across evidence streams and regulatory contexts. The Methods2AOP collaboration seeks to bolster confidence in the use of AOPs by elaborating upon the documentation of methods used.

With respect to existing standards, Methods2AOP does not seek to replace or duplicate established assay description frameworks or harmonized reporting templates developed in related initiatives. On the contrary, the collaboration explicitly considered these efforts and actively engaged with communities involved in assay documentation and standardization to ensure our proposal was complementary and interoperable. However, the current initiative described in our manuscript was focused on supporting knowledge integration across AOPs to allow for AOP-associated methods to be comparable, searchable, and reuseable within the AOP-Wiki and broader AOP-KB. Methods2AOP thus identified which method-related information should be captured in a structured, ontology-aware way to support AOP development and reuse, while remaining fully compatible with existing standards.

From an operational and governance perspective, Methods2AOP deliberately avoids proposing new review, validation, or terminology governance structures. Instead, it is designed to piggyback on the established AOP-Wiki and OECD governance mechanisms. In practice, method-related information would be subject to the same contributor access rules, expert review processes, and quality expectations that already apply to AOP content. As AOPs progress toward higher levels of maturity and, ultimately, OECD endorsement, the stringency of review and quality control naturally increases, and any Methods2AOP-inspired structured method information would be reviewed accordingly. In this way, governance and quality assurance are inherited from existing processes rather than redefined.

Regarding applicability across evidence streams and regulatory relevance, Methods2AOP is intentionally a bottom-up framework. It defines structured metadata elements that can capture method information across diverse study types (in vitro, in vivo, in silico, human, ecological), recognizing that the nature and granularity of relevant metadata will differ between these contexts. The framework does not attempt to define regulatory workflows such as IATAs, Defined Approaches, weight-of-evidence evaluations, read-across, or NGRA. Instead, it provides the structured building blocks that such workflows already rely on, enabling method information to be more transparent, comparable, and reusable when integrated into higher-level regulatory assessments. This is fully aligned with the broader AOP philosophy, in which modular, mechanistically grounded knowledge supports but does not prescribe regulatory decision-making.

Overall, Methods2AOP should be understood as a pragmatic, enabling contribution that strengthens the interface between methods and AOPs, while remaining firmly grounded in existing standards, governance structures, and the established scope of the AOP framework and ultimately the regulatory use of AOP-based data.

Main comments

1. Need for more concrete, worked examples

The reviewers emphasize that the core contribution of the manuscript is the proposal to move from free-text method descriptions toward structured, standardized metadata. While the conceptual design, candidate fields, and figures are clearly presented, the reviewers find the proposal somewhat abstract and would like to better understand how it would function in practice within the AOP-KB. In particular, they are looking for a more intuitive illustration of how the proposed metadata would actually be populated for a real Key Event, and how this would concretely improve on the current “How it is measured or detected” free-text field.

Comment 1 Response:

We thank the reviewers for highlighting the need for a more concrete illustration of how the proposed structured metadata could be populated in practice. We agree that such examples demonstrate feasibility and provide added value. We have expanded the supplementary material (Supplementary Table 1) providing multiple worked examples within a single, versioned Excel file hosted on Zenodo. Each worked example is provided on a separate tab, allowing different methods and evidence streams to be illustrated while using the same underlying Methods2AOP metadata structure. Each tab populates only a subset of relevant metadata fields, intentionally demonstrating the principles of progressive disclosure and burden minimization. The examples show where ontology-backed annotation can be applied and where free-text entries and expert judgement remain appropriate, reflecting the current state of ontology coverage in toxicology. Importantly, these worked examples are illustrative rather than prescriptive. They are not intended to define mandatory fields, data-entry workflows, or user-interface design for the AOP-Wiki, but rather to demonstrate the types of structured information in practice. Specifically, the supplementary dataset now includes:

two illustrative, non-exhaustive in vitro examples based on the human Cell Line Activation Test (hCLAT) and Epidermal Sensitization Assay (EpiSensA)

an illustrative, non-exhaustive in vivo example based on electrocardiograph in rats

an illustrative, non-exhaustive in chemico example based on Amino acid Derivative Reactivity Assay (ADRA)

 Additions to manuscript summarized in italics below:

Section 2.3 (after first paragraph): “…each of these study types would have their own groups of methods information needs. In addition to listing the proposed metadata fields and ontology references, four illustrative, non-exhaustive worked examples demonstrating how selected fields can be populated for in vitro, in vivo, and in chemico test methods have been provided (Karmaus et al. 2025 [this is the updated reference for the supplementary file]). The worked examples intentionally do not populate all available fields, reflecting the principles of minimizing reporting burden while increasing transparency and easily interpretable information entry. Resources containing annotated methods using similar metadata fields that can support such data integration can include CAMERA (Collection of Alternative Methods for Regulatory Application; https://camera.niehs.nih.gov/) and CEBS (Chemical Effects in Biological Systems; https://cebs.niehs.nih.gov/cebs/).”

2. Clearer positioning relative to existing standards and frameworks

The reviewers note that method standardization is already supported by several existing frameworks, including OECD Test Guidelines, guidance documents, and harmonized templates. They therefore ask for a clearer explanation of how the Methods2AOP proposal relates to these efforts, why they are not sufficient on their own in the AOP-Wiki context, and what specific additional value Methods2AOP brings. While related initiatives are mentioned in the manuscript, the reviewer feels the relationship between these efforts and Methods2AOP should be made more explicit.

Comment 2 Response:

The reviewers correctly note that multiple established frameworks already support assay description and standardization. Methods2AOP does not introduce a new or parallel standard but rather a proposed implementation of standardized vocabularies that is interoperable and complementary to existing frameworks. The added value of Methods2AOP lies in supporting comparability, searchability, and reuse of method information across AOPs, which is not the primary objective of existing assay-focused standards. The Methods2AOP suggested way of treating method information can serve as input for parties developing the AOP-Wiki 3.0 in the future, where relevant functionality can certainly be implemented.

Additions to manuscript summarized in italics below:

Abstract: “This paper reviews their work emphasizing the importance of linking detailed test method information and conceptually proposes how it may be included in the AOP knowledgebase in alignment with existing assay documentation standards and governance frameworks.”

Section 3 (fourth paragraph): “Integrating more detailed methods information could strengthen confidence of the association between the method and the KE and consequently support the evidence for the regulatory endpoint. In fact, the proposed ontologies support interoperability with existing frameworks and complement established documentation standards (e.g., OECD Harmonized Templates) which could increase the utility of AOPs.”

3. High-level operational and governance perspective

The reviewers highlight that structured information systems often encounter their main challenges during implementation and long-term maintenance rather than at the design stage. They specifically mention validation and quality control of entries, governance of terminology resources, and ensuring consistency across contributors. Even if implementation is considered out of scope, they ask for at least a minimal operational and governance perspective, including which stakeholders might realistically assume responsibility for these aspects.

Comment 3 Response:

The reviewers raise important considerations regarding validation, quality control, and governance of structured information systems. Methods2AOP intentionally does not define new operational or governance structures. While this consideration is out of scope for the current work, the assumption for the collaboration during this effort was that any proposal would have to operate under the existing AOP-Wiki and OECD governance mechanisms, including contributor access control, expert review, and increasing levels of scrutiny as AOPs mature toward OECD endorsement. Structured method information inspired by Methods2AOP would therefore be subject to the same review expectations and quality assurance processes as other AOP content.

Additions to manuscript summarized in italics below:

Section 1.2: “The Methods2AOP collaboration suggest content and a conceptual documentation schema, presented herein for future consideration to integrate into the AOP-KB. Accordingly, any structured method information inspired by Methods2AOP would be subject to the same governance, review, and quality control processes that already apply to AOP-Wiki content.”

4. Perceived in vitro bias and regulatory applicability

The reviewers observe that although the manuscript claims applicability across multiple study types and evidence streams, the narrative and examples appear weighted toward in vitro contexts. They ask for clearer justification that the proposed structure can cover diverse evidence streams (in vivo, in silico, human, ecological) and for a more explicit explanation of how structured method information would connect to concrete regulatory workflows such as IATAs, Defined Approaches, weight-of-evidence evaluations, read-across, and NGRA.

Comment 4 Response:

The reviewers note a perceived emphasis on in vitro contexts and ask for clearer links to regulatory workflows. The use of in vitro examples primarily reflects their suitability for illustration rather than a limitation of the framework. Methods2AOP is designed to accommodate method information across study types, recognizing that metadata needs differ between in vitro, in vivo, in silico, human, and ecological contexts. We have provided more detailed examples of other study types to these ends in the updated Supplementary File (Karmaus et al. 2025). The in vitro study type remains as the illustrative example in the manuscript given the overwhelming abundance of in vitro assays used as evidence for mechanistic KEs.

The Methods2AOP collaboration does not aim to define the impact of methods documentation within the AOP infrastructure on regulatory workflows such as IATAs, Defined Approaches, weight-of-evidence evaluations, read-across, or NGRA. Rather, it proposes structured and transparent method documentation which bolsters the use of AOPs in regulatory contexts (which would in turn include applying AOPs as lines of evidence in a multitude of contexts including the workflows listed. We have clarified this in the manuscript.

Additions to manuscript summarized in italics below:

End of Section 3: “ Consistent with the broader AOP framework, the proposed structuring of method information is intended to bolster confidence in the application of AOPs which can in turn support higher-level regulatory applications such as Integrated Approaches to Testing and Assessment, Defined Approaches, and weight-of-evidence evaluations, without prescribing how such workflows should be constructed.”