ELIXIR and Toxicology: a community in development

Toxicology has been an active research field for many decades, with academic, industrial and government involvement. Modern omics and computational approaches are changing the field, from merely disease-specific observational models into target-specific predictive models. Traditionally, toxicology has strong links with other fields such as biology, chemistry, pharmacology and medicine. With the rise of synthetic and new engineered materials, alongside ongoing prioritisation needs in chemical risk assessment for existing chemicals, early predictive evaluations are becoming of utmost importance to both scientific and regulatory purposes. ELIXIR is an intergovernmental organisation that brings together life science resources from across Europe. To coordinate the linkage of various life science efforts around modern predictive toxicology, the establishment of a new ELIXIR Community is seen as instrumental. In the past few years, joint efforts, building on incidental overlap, have been piloted in the context of ELIXIR. For example, the EU-ToxRisk, diXa, HeCaToS, transQST, and the nanotoxicology community have worked with the ELIXIR TeSS, Bioschemas, and Compute Platforms and activities. In 2018, a core group of interested parties wrote a proposal, outlining a sketch of what this new ELIXIR Toxicology Community would look like. A recent workshop (held September 30th to October 1st, 2020) extended this into an ELIXIR Toxicology roadmap and a shortlist of limited investment-high gain collaborations to give body to this new community. This Whitepaper outlines the results of these efforts and defines our vision of the ELIXIR Toxicology Community and how it complements other ELIXIR activities.

life science efforts around modern predictive toxicology, the establishment of a new ELIXIR Community is seen as instrumental.In the past few years, joint efforts, building on incidental overlap, have been piloted in the context of ELIXIR.For example, the EU-ToxRisk, diXa, HeCaToS, transQST, and the nanotoxicology community have worked with the ELIXIR TeSS, Bioschemas, and Compute Platforms and activities.In 2018, a core group of interested parties wrote a proposal, outlining a sketch of what this new ELIXIR Toxicology Community would look like.A recent workshop (held September 30th to October 1st, 2020) extended this into an ELIXIR Toxicology roadmap and a shortlist of limited investment-high gain collaborations to give body to this new community.This Whitepaper outlines the results of these efforts and defines our vision of the ELIXIR Toxicology Community and how it complements other ELIXIR activities.

Introduction of the ELIXIR Toxicology Community
Toxicology as a field tries to understand the negative consequences that may arise from the interactions of chemicals with living organisms.This ELIXIR (European life-sciences Infrastructure for biological Information) 1 Community will concentrate on the focus areas of ELIXIR, including the protection of human, animal, and environmental health.There are several chemical and biological "interoperability" issues key to the toxicology field that translate into data interoperability issues.These include the connection between the action and activity of a particular chemical compound to its effective amount available at a biological target (the link between toxicodynamics and toxicokinetics).Typically, this is also a link between biological data analysis (including large-scale multi-omics) and kinetic modelling.Other examples include interactions between a compound and its target (a protein, nucleic sequence, or membrane structure for instance).This is primarily based on the interplay between chemistry (the chemical structure and, for example, its related properties in terms of functional groups, charge, shape, and related binding affinity) and biochemistry (like biomolecular 3D structures).Also, mixture toxicity needs to be considered as combinations of chemicals with synergistic or antagonistic behaviour, or a combination thereof.While chemicals with similar modes of action may act in terms of concentration addition, those with different modes may rather act according to independent action. 2,3Substances with low toxicity may interact in concentration addition rather than as excess toxicity drivers of one compound.Often there is a need to translate the knowledge about one compound into knowledge about other compounds, where approaches like quantitative structure-activity relationships (QSARs) help to elucidate this knowledge and predict required property and toxicity, and in more general when "read-across approaches" come into play that are based on chemical data only, biological data only, or are hybrid. 4,5These again need detailed information about the relationships between related chemical compounds, specific properties, and toxicological endpoints and, when considering chemical-biological interactions, details regarding both the chemical structures and adequate descriptions of the biological targets.Since toxicological endpoints can be represented in a myriad of ways, the toxicological effect data are often scattered over multiple repositories and databases hosting different types of data; i.e., chemical structures, toxicity data (in vivo and in vitro), biological target details, and omics data.This is not a problem, per se, but the separation and segregation make the data difficult to find and connect.Currently, many of these deposition databases (where datasets can be archived) do not provide adequate descriptions regarding typical toxicological study designs and parameters, quality control, data acceptance criteria, or even clear identification of the compounds tested.This all adds to the need for an adequate FAIRification process, to make toxicology data more Findable, Accessible, Interoperable and Reusable (FAIR -see Go-FAIR). 6,7e field of toxicology would certainly benefit from clear, standardized guidelines for data capture, approaches to integrate and connect across multiple databases, clear data licensing for these data repositories, and tools that support accessing the data (as being developed by the ELIXIR Converge data brokerage work).Existing study capture tools can be extended with templates defined for toxicology, which end up in a central place (e.g., Biosamples 8 ) with links to omics and other data distributed over technology-specific deposition databases and BioStudies. 9Relevant portals (e.g., FAIRsharing.org 10,11) should then be able to identify these studies, also linking to existing but scattered toxicology databases.
Risk assessment, consisting of hazard identification, characterisation, and risk evaluation in relation to exposure, involves expert opinions based on discussions and data interpretation.Streamlining this process is important also because of the huge number of chemicals known and produced in chemistry, biotechnology, or food production, with over 350,000 chemicals documented on the global market. 12However, expert evaluation is and will remain crucial and calls for extensive data provenance both for the actual data (e.g., how it was measured, where it was published, and whom to give credit for it), and for the risk assessments themselves.Interestingly this process stumbles on problems and solutions that have much in common with other fields, where part of the data needs to be hidden and other parts can be publicly accessed, as observed in pilot approaches in genetics, or patient data repositories.Toxicology, while established as a discipline, is also rapidly developing in areas, where, for example, new molecular methods to describe the adverse outcome of exposure to a toxic substance are not yet fully established.This offers opportunities for integration in systems biology approaches building on molecular pathway descriptions that benefit from

REVISED Amendments from Version 1
We updated the manuscript based on reviewers' comments, and we made some minor textual changes and grammar corrections.A new section has been added on how to join the community, including a reference to the toxicology community webpage.Also, funding information was improved, putting grants from the same funders together instead of separate statements, and two references were added to the outputs of community joint activities.
modern network biology approaches.Compound profiling, where, for example, using omics methods to characterize the effect of compounds on cells, is generally useful for categorizing compounds, or for the development of predictive pathway models.For such studies, the availability of high-quality annotations for compounds is of paramount importance to enable the use of omics profiles in toxicology.
Toxicology is also an applied field.There are important applications of toxicity tests in the regulatory field and large amounts of data are collected for that purpose, often for different (governmental) agencies.Making this data more "Findable" and "Reusable" is often seen as an important way to reduce both animal testing and the cost of registration of new compounds.If some data is not available to the public due to ownership by companies or other constraints, indexes should be developed to enable the use of this data in an aggregated form (such as the SPIN index of the Swedish Chemicals Agency (KEMI) and other Nordic chemical agencies).Typically, regulatory use requires very precise and rigid descriptions of protocols, including reporting methods.On the other hand, better insights into the toxicological mechanisms, including interactions between chemicals, benefit from more innovative research methods (e.g., single-cell omics, induced stem cell applications, and imaging methods) and from the creative development of new analysis methods.While it is beneficial to combine results from both types of approaches (termed New Approach Methods (NAM) in the regulatory field), the corresponding interoperability issues are often quite different.

User community
Europe is steadily increasing demands on the risk assessment of chemicals, drugs, cosmetics ingredients and nanomaterials to lead to safer products, resulting in a strong toxicology research community with sub-communities in, for example, the drug development, environmental, nanomaterial and rare-disease areas.Recent changes in European law for animal testing and new demands for testing of lower-volume chemicals and nanomaterials have triggered large-scale research into alternative testing approaches.These activities not only produce new biological and mechanistic insights, but also large amounts of new data, which must be managed and shared for re-usage to avoid unnecessary duplication of experiments and, in this way, reduce animal testing.The goal is that the combination of data from integrated in vitro and in silico approaches will support (ultimately personalized) risk/benefit health analysis, safer drug innovation with fewer needs to withdraw after registration, a fact-based perception of chemical safety, safe-by-design nanomaterials, and sustainable and safe economies.
The European Union supports toxicological and risk assessment projects with various funding programs.Recently, large collections of data have been released, resulting from research clusters, such as SEURAT-1, 13 the EU NanoSafety Cluster (NSC) with NANoREG, EU-ToxRisk, 14 the NORMAN Network, and the EU Innovative Medicines Initiative (IMI) funded projects related to drug toxicology, including eTOX, 15 eTRANSAFE, 16 and Eurion. 17Data from these and other projects are becoming available, sometimes as Open Data (e.g., NANoREG) and sometimes as FAIR data.An example of the latter is European REACH data, which has recently been made FAIR by the Cefic-LRI-funded project AMBIT-LRI. 18urthermore, the FAIRplus project is collaborating with eTOX on making their data more FAIR, and the new Precision Toxicology will develop a data commons following the FAIR principles.
However, there are a few opportunities for data handling that need to be taken. 19,20Recent studies show how powerful the combination of toxicology information and omics data is, 21,22 but to be able to obtain the statistical significance to draw these conclusions, data from the US and Japan had to be combined.In contrast to large data sets like DrugMatrix, 23 ToxCast/Tox21, 24 and TG-GATEs 25 from these countries, data from European projects is often not sufficiently integrated.Luckily, there are signs that the community is going in the right direction, e.g., the aforementioned data integration by diXa, 26 NANoREG and REACH. 27With respect to the European Chemical Industry, various authors have been involved in other Cefic-LRI activities related to data management (AIMT-3, AIMT-4).
In addition, the eTOX project 28 has established data integration approaches e.g., to enable the development of QSARs relating chemical structures to in vivo toxicopathological outcomes.As such, the project also delivered databases and approaches to ontology development, text mining approaches, and approaches for the prediction of drug metabolism and pharmacokinetic features.Moreover, in 2017 the Organisation for Economic Co-operation and Development (OECD) performed an online survey underscoring the fact that data integration for safety is of global concern for ultimate risk assessment.The purpose of the resulting knowledge base is the integration of eCHEMportal (The Global Portal to Information on Chemical Substances 29 ), IUCLID (International Uniform ChemicaL Information Database 30 ), and OECD's QSAR Toolbox 31 supporting the development of Adverse Outcome Pathways and associated infrastructures (AOP-Wiki). 32Despite these positive developments, the 'data integration struggle' from various perspectives (omics, computational chemistry and more 'conventional' toxicological data within REACH and pharmaceutical industry setting) remains a challenge.

Roadmap
The above initiatives are mainly driven by user communities themselves: the chemical industry, funding agencies, pharmaceutical companies, governmental agencies, and organisations such as Member State organisations, the OECD, and non-governmental organisations (NGOs).ELIXIR can contribute strongly to the existing infrastructure projects from a cheminformatics and bioinformatics perspective, providing tools and guidelines, linking and harmonizing the ongoing activities, and serving the toxicology users.
For future risk assessment paradigms solely based on human-derived models, and in this way of higher relevance for human adverse effects, 33 various data types will need to be integrated into and cross the conventional boundaries of risk assessment.This involves external exposure assessment (e.g., via workplace or environmental modelling and measurements 34,35 ), internal exposure characterisation (ADME-TK (absorption, distribution, metabolism, and excretion -toxicokinetics) such as via modelling and biomarker-based detection), toxicodynamics on a molecular level, and cell and systems biology.In this way, more data-driven mechanism-based evaluations and supporting data can be integrated into regulatory risk assessment.There will not only be more but also more diverse data such as internal exposure data, that may be inferred from biomonitoring data and/or physiologically-based toxicokinetic modelling to estimate target dose available at the active sites involved in the molecular initiating events of Adverse Outcome Pathways (AOPs).
Another relevant topic is the concept of the exposome, [36][37][38] which aims at characterising lifetime exposure (not only to chemicals in the narrowest sense, but also dietary components, lifestyle factors, environmental exposures, and more) in relation to health outcome.Often, vulnerable periods of life (infancy, childhood, and old age) are investigated, and the evaluation also integrates epidemiology.This is one clear demonstration of the trend that the previously distinct areas of toxicology, drug and product design, and personalized/precision medicine 39 but also environment and health and epidemiology are moving closer together.Data sharing will be increasingly necessary across these disciplines.High throughput data analysis in exposomics, for instance, shares many parallels with metabolomics and other higher-level omics analyses, with an added layer of chemical complexity on top. 38e toxicology community is large and well-established.The current list of proponents of an ELIXIR Toxicology Community only reflects a subset of a much larger community with a lot of commitments to, and activities around open collaboration.It has clear omics, knowledge management and data infrastructure needs to accommodate the increasing wish to predict toxicology without animal testing (e.g., in SEURAT-1, EU-ToxRisk, eTOX, and OBERON 40 ).Foreseeing this need for better infrastructures, the community has previously contacted ELIXIR for collaboration.Various domain-specific projects exist that service the toxicology community with computational and database knowledge (e.g., OpenRiskNet, NanoCommons) that can translate ELIXIR knowledge to the respective communities.These infrastructure projects are the successors of research projects focusing on data management, including diXa, 26 ToxBank 41 and eNanoMapper. 42 benefit the research community, small and medium-sized enterprises (SMEs) and larger industries, and to enable further future support to regulatory applications and upcoming calls (e.g., Green Deal 43,44 ), we need to reach an inclusive ecosystem of data, evaluation, and modelling tools.The current separate consortia from different toxicity-related and neighbouring disciplines already work towards data and knowledge that is FAIR. 6After all, these aspects are essential to efficiently assess the risk of new compounds and materials, as well as combined risks of current stressors (e.g., under the exposome concept).To further accelerate these activities, more toxicology-related data and knowledge need to be linked, such as on physiologically based toxicokinetic (PBTK) modelling, biological pathways describing affected metabolism and cell biology processes, metabolism, metabolic models and metabolism predictions, drug-response, omics (such as biological identity), chemical structures and associated metadata (use, hazard, transformations), QSARs, AOPs, REACH dossiers, and more.Simultaneously, an extension towards the human (preclinical toxicology) discipline should be initiated, in which exposure data are combined with internal exposure and early biomarkers of effect data (e.g., from the European Human Biomonitoring Initiative (HBM4EU) 45 and environmental data from NORMAN 46,47 ) towards pathways of toxicity.Interoperability with the standardization efforts of clinical research by CDISC is also important.However, to reach such interoperable toxicology, resources need to be better integrated.Despite the work of many projects, their FAIR features can still be improved and applying newly developed FAIR metrics will help steer this. 48ven though there is overlap in content with existing ELIXIR Communities (Table 1, key demands specifically fostering the integration for interoperable toxicology and risk assessment include the following roadmap 19,20 ): • chemical structure interoperability challenges (e.g., links to ELIXIR Metabolomics Community 49 ) • metadata, open standards (e.g., links to ELIXIR Interoperability Platform, and TeSS 50 ) • continued ontology development (e.g., links to ELIXIR Interoperability Platform and Ontology Lookup Service) • interoperable computation (e.g., links to the Galaxy 51 and bio.tools 52

communities)
• interactions with other ELIXIR Core Data Resources (e.g., Ensembl and Europe PMC) • interactions with other communities, including nanomedicine and health • deployment of existing tools and modelling approaches on the ELIXIR Compute infrastructure (also allowing future growth towards risk estimations needing Monte Carlo approaches) • integration of ELIXIR AAI • InChI implementation for small molecule data • Spectral database functionality (open implementations) Specifically, we will continue to grow the list of involved toxicology research groups, projects, and ELIXIR Node activities.This Community has already held a joint meeting to select the key priorities and use cases (see Tables 2 and 3), resulting in this positioning paper.The ELIXIR Toxicology Community will continue to expand and search for contributors with relevant expertise as the community and activities mature.By bootstrapping from Open Science approaches developed in aforementioned projects (e.g., Open PHACTS 53 ), the new Community will focus on mutual benefit, an open and inclusive community, solving practical community problems.The goal is not to design domainspecific approaches, but a pragmatic approach that provides FAIR and open tools from the start, allowing all toxicology and neighbouring communities to benefit from these harmonized solutions.The inclusive community will involve the existing sub-communities in pharmaceuticals, e.g., from eTOX, transQST, 54 and eTRANSAFE, 16 cosmetic ingredients (e.g., from SEURAT-1), high and low-volume chemicals (e.g., from HBM4EU and soon from Horizon Europe Partnership for the Assessment of Risk from Chemicals (PARC), 47 or from different Cefic-LRI (Lang-Range Research Initiative) projects), and nanomaterials (via the NanoSafety Cluster), thus building on shared needs and community solutions and strongly aligned to other ELIXIR communities.Open licensing and interfaces (such as ontologies, standards, and formats) will encourage new solutions and collaborations, which will be accessible to any organisation and every project within and outside Europe.This will allow close interoperability with toxicology communities outside Europe that also use open approaches, while at the same time allowing compatibility with closed approaches too.This dual model has been demonstrated successfully in recent projects.A prioritized roadmap is essential; the label "ELIXIR Community" would enable us to set priorities at a level above the individual projects.Existing components that will give this Community an initial boost include: software (e.g., AMBIT with OpenTox API 55 ), databases (e.g., diXa, eNano-Mapper, AMBIT-LRI, NORMAN-SLE, and MassBank), ontologies (e.g., the eNanoMapper ontology, 56 and AOP ontology 57 ), interoperability concepts (e.g., annotation of OpenAPIs, in collaboration with bio.tools, and semantic structural searches with IDSM), teaching/education material (i.e., Bioschemas annotation of outreach activities,  1) and with Core and National Resources.Various existing ELIXIR Communities need similar solutions, e.g., for chemical structure handling, but also toxicology needs proteomics and metabolomics, toxicology involves human data, and ecotoxicology has a significant impact on crops and health.
The following projects have been adopting and integrating FAIR toxicology concepts but need integration with ELIXIR Platforms and Communities: eTOX, NanoCommons (NanoSafety Cluster), EU-ToxRisk, OpenRiskNet, OpenTox Foundation, Open PHACTS Foundation, and the diXa platform.Many other projects have a specific scientific focus but also need integration and some will work on the FAIR concepts.A non-exhaustive list is ACEnano, SmartNanoTox, HeCaToS, NewGeneris, EnviroGenoMarkers, EXPOsOMICS, HELIX, ASAT, 58 PATROLS, and HEALS, as well as new projects, including new Horizon 2020 projects RISK-HUNT3R and HARMLESS and the new Horizon Europe project PARC.Companies and organisations will profit from this Community either as users or as providers of services on top of the infrastructure, including ECHA (FI), Edelweiss Connect (CH), IdeaConsult Ltd. (BG), Misvik Biology (FI), TNO (NL), Seven Past Nine (SI), and the Swedish Academic Consortium for Chemical Safety (SwACCS, SE).Industries showed a strong interest in toxicology, demonstrated by their activities: Cosmetics Europe was participant in the SEURAT-1 cluster; chemical industries (Nanotechnology Industries Association) are a participant of the NSC; chemical branch organisation (Cefic) funds LRI projects around toxicology; and pharmaceutical industries funds toxicology research via IMI projects like eTOX, eTRANSAFE, and Open PHACTS.The Research Data Alliance (RDA) organized a workshop recently about the integration of toxicogenomics resources, 59 and collaboration with international organizations has already been established with, for example, the CompTox Chemicals Dashboard team of the US EPA 60 and PubChem from the US National Institutes of Health. 61Joint activities include an ELIXIR BioHackathon Europe 2022 project around PubChem-compatible data exchange formats 62 and the use of Wikidata and Wikipedia for important chemicals. 63

How to join
The Toxicology Community has a webpage at elixir-europe.org/communities/toxicology where it is possible to join the Community's mailing list and stay up-to-date with community activities such as the annual Face-to-Face or scheduled workshops.These are good opportunities to actively participate in the network.

Disclaimer
• Where authors are identified as personnel of the International Agency for Research on Cancer/World Health Organization, the authors alone are responsible for the views expressed in this article, and they do not necessarily represent the decisions, policy, or views of the International Agency for Research on Cancer/World Health Organization.
• The views expressed in this manuscript are solely those of the authors and do not represent the policies of the U.S. Environmental Protection Agency.Mention of trade names of commercial products should not be interpreted as an endorsement by the U.S. Environmental Protection Agency.
• The views expressed in this manuscript are solely those of the authors and do not represent the policies of and should not be interpreted as an endorsement by the U.S. National Institutes of Health.
models to target-specific predictive models.The article introduces the rapidly developing field of toxicology and the interoperability issues arising in combining chemical and biochemical analysis, biological data, and kinetic modelling.The authors argue for "clear, standardized guidelines for data capture, approaches to integrate and connect across multiple datasets, clear data licensing for these data repositories, and tools that support accessing the data".The ELIXIR Toxicology Community is suggested to 1) connect the different user communities (including industry, governmental agencies, NGOs, and more), 2) strongly contribute to link between, and harmonize the already existing relevant infrastructure projects, resources, and activities, and 3) provide useful tools, guidelines, and training.
Acknowledging the vast number of chemicals that are currently being produced, and the limited knowledge of the consequences of exposure to humans and the environment, this is a timely and important effort.From my impression, an ELIXIR Community is a good tool to drive and coordinate such an international initiative.
I find the article comprehensive and well-written, and their arguments well discussed.However, I have some points that I hope the authors will take into consideration when taking the initiative further: It is early stated that "in this ELIXIR Community, the focus will be primarily on the protection of human health".I would argue that this is an unfavorable restriction for three reasons: It disregards the 'OneHealth approach' recognized by WHO, FAO, OIE, and UNEP, which is defined as "an integrated, unifying approach that aims to sustainably balance and optimize the health of people, animals and ecosystem.It recognizes the health of humans, domestic and wild animals, plants, and the wider environment (including ecosystems) are closely linked and inter-dependent" (One Health High-Level Expert Panel Annual Report 2021).EFSA, ECHA, EMA et al. are also organizing a conference on the topic in June 2022 (https://www.one2022.eu).Seeing that animals and ecosystems are closely linked to human health, for example through foods, their responses to and effects from toxicant exposures are also highly important to study and include in risk assessments.

○
Live animals, as well as animal tissue, cells, and molecules, are often used as models for toxicological studies.As the toxicological defense systems are highly preserved through evolution, they can give a good indication of human outcomes.Thus, it is important to also ensure supporting infrastructure for integrating these as well.

○
Ontologies, standards, and data resources (like omics repositories, data analysis tools, etc.) are often already overlapping.Although I welcome different stakeholders being identified as "user communities", I am concerned by the, from my understanding, lack of industry and governmental/administrative agencies in the author list of this article.I believe that if the aim of improving FAIR data integration for safety to enable ultimate risk assessments is to be met, it is eminent to include non-academic perspectives, experiences, and concerns into the initiative from early on.Experiences from the PARC project could for example be useful in how to approach this.

2.
From my experience, the field of toxicology in general has limited experience with FAIR data 3. management and is not familiar with the resources that are in place.I see that toxicologists' communication, dissemination, and training are mentioned in the proposed roadmap, but I am missing a more targeted approach to engaging with the communities.For example, such activities could include organizing toxicology-specific workshops, presenting at toxicology conferences, and targeted information through different channels.As the proposed ELIXIR Community seems to have a good overview of ongoing initiatives and infrastructures and a good anchoring in the scientific community, a mapping of their target audience and users in the field could be suggested as a starting point.
As mentioned in the paper, systems biology approaches are also getting implemented in toxicology, i.e. systems toxicology (described in Sturla et al., 2014 1 .From my point of view, I am missing mentioning of tools for this, such as the FAIRDOM SEEK platform that integrates ISA-structured metadata and workflows with JWS-based biological systems modelling.However, there might already be other solutions for this that I don't recognize in the paper This wide-ranging review circumscribes an important data integration task with a major catherding dimension.Whilst also appearing herculean, this esteemed author collective fully understands what they are letting themselves in for and I wish them the best for this endeavor.It is, of course, early days, but I will make a few points (whether these might be addressed in a revision and blending in what other reviewers may come up with I will leave to the authors) Harvesting tox data from the extant and future literature seems to neither be specifically addressed nor proposed via direct interactions with the pharmaceutical companies generating most of it.Standardised data from their large historical internal sets only surfaces sparsely and heterogeneously public databases.Companies such as LahsaVtic, Instem, and ToxPlanet claim to have large databases compiled from the literature.Might ChEMBL come into the frame here if they could strategically increase their toxicology data extraction from the literature, both prospectively and retrospectively?(So far with only 24 assays) 1.
My impression is there are simply too many resources mentioned for realistic overarching harmonisation.Could these be cut back to a smaller "active membership" prioritised by the amount of data they have and will continue to generate?2.
Quote "standardized guidelines for data capture, approaches to integrate and connect across multiple databases".As we know, this is going to be the really tough bit.It needs the major sources to share not only their data schema and metadata indexing but push for a level of convergence.It has to be said that ELIXIR does not have a good record on this.Unless I have missed it, none of the core data resources has made any significant changes in their individual (and often decades-old) data models to really enhance interoperability.

3.
PubChem has 113,110 compounds from ~ 30 annotation sources (although most relatively small) indexed under "PubChem Compound TOC: Toxicological Information" and toxicologyrelated submitting sources cover over 0.5 million substances.In addition, the term "toxicology" matches 3,774 BioAssays covering 86,068 compounds.This makes PubChem a de facto toxicology integration hub.The ELIXIR effort should consolidate and expand this, including where an increase in toxicology curation by ChEMBL could feed into BioAssay and thus become interoperative with the 294,000 biological activity data points already in there.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.
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias • You can publish traditional articles, null/negative results, case reports, data notes and more • The peer review process is transparent and collaborative • Your article is indexed in PubMed after passing peer review • Dedicated customer support at every stage • For pre-submission enquiries, contact research@f1000.com

4 . References 1 .
Sturla SJ, Boobis AR, FitzGerald RE, Hoeng J, et al.: Systems toxicology: from basic research to risk assessment.Chem Res Toxicol.2014; 27 (3): 314-29 PubMed Abstract | Publisher Full Text Is the topic of the opinion article discussed accurately in the context of the current literature?Yes Are all factual statements correct and adequately supported by citations?Yes Are arguments sufficiently supported by evidence from the published literature?Yes Are the conclusions drawn balanced and justified on the basis of the presented arguments?Yes Competing Interests: No competing interests were disclosed.Reviewer Expertise: Environmental toxicology, endocrine disruption, bioinformatics, systems toxicology, marine toxicology 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.© 2022 Southan C.This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Christopher Southan 1 Deanery of Biomedical Sciences, University of Edinburgh, Edinburgh, UK 2 Data Sciences,, Medicines Discovery Catapult, Macclesfield, UK

4 .
Is the topic of the opinion article discussed accurately in the context of the current literature?Yes Are all factual statements correct and adequately supported by citations?Yes Are arguments sufficiently supported by evidence from the published literature?Yes Are the conclusions drawn balanced and justified on the basis of the presented arguments?Yes Competing Interests: No competing interests were disclosed.Reviewer Expertise: Bioinformatics and cheminformatics (see LinkedIN)

Table 1 .
Overlap of the proposed ELIXIR Toxicology Community with ELIXIR platforms and communities.The concrete steps forward proposed in this roadmap include: 1. Leverage from open solutions (such as models, ontologies, educational material, and standards) developed by past and ongoing toxicology and ELIXIR projects 2. Connect more closely with the core ELIXIR resources (such as FAIR data and database interoperability), strengthen and connect the inclusive communities that have evolved over the past few years (OpenTox, eNanoMapper, diXa, OpenRiskNet, NORMAN) and older communities like the Federation of European Toxicologists & European Societies of Toxicology (EUROTOX), the Society of Environmental Toxicology and Chemistry (SETAC) and European Environmental Mutagenesis and Genomics Society (EEMGS, formerly known as EEMS) 3. Develop open community standards to support common interest (ontologies, APIs, data formats, deposition databases, and publication recommendations)

Table 2 .
Examples of existing and anticipated collaboration.

Table 3 .
The Toxicology Community roadmap is roughly defined by three themes (steps).Each step comes with a 10-year aim, further detailed with possible work that could be done in ELIXIR activities.

Table 3 .
Continued TeSS), and virtual infrastructures (e.g., OpenRiskNet Virtual Research Environment and the NORMAN Digital Sample Freezing Platform).However, each of these approaches would benefit from integration in the ELIXIR Platforms (see examples in Table