Ontology-driven and artificial intelligence-based repeated dose toxicity testing of chemicals for next generation risk assessment

Projektinformationen

ONTOX

ID Finanzhilfevereinbarung: 963845

Projektwebsite

DOI

10.3030/963845

EK-Unterschriftsdatum 29 Januar 2021

Startdatum 1 Mai 2021

Enddatum 31 Oktober 2026

Finanziert unter

SOCIETAL CHALLENGES - Health, demographic change and well-being

Gesamtkosten

€ 17 447 050,00

EU-Beitrag

€ 17 211 050,00

17 211 050,00

236 000,00

Koordiniert durch

VRIJE UNIVERSITEIT BRUSSEL
Belgium

Periodic Reporting for period 3 - ONTOX (Ontology-driven and artificial intelligence-based repeated dose toxicity testing of chemicals for next generation risk assessment)

Berichtszeitraum: 2024-05-01 bis 2025-10-31

The overall goal of ONTOX is to deliver a generic strategy to create innovative New Approach Methodologies (NAMs) to predict systemic repeated dose toxicity effects of chemicals that, upon combination with tailored exposure assessment, will enable human risk assessment. For proof-of-concept purposes, focus will be put on 3 organs, each with 2 specific types of adversity, namely the liver (steatosis and cholestasis), the kidneys (tubular necrosis and crystallopathy) and the developing brain (neural tube closure and cognitive function defects). This will result in 6 NAMs in total. Chemicals from different application domains will be considered, including from, but not limited to, the pharmaceutical, cosmetics, food and biocide sectors. The 6 NAMs will each consist of a computational system based on advanced Artificial Intelligence (AI), and will be primarily fed by available biological, toxicological, chemical and kinetic data. These data will be consecutively integrated in physiological maps, quantitative Adverse Outcome Pathway (qAOP) networks and ontology frameworks. Data gaps, as identified by the AI system, will be filled by targeted in vitro (cell culture) and in silico (computational) testing. The impact of ONTOX is, however, de facto much broader, as the underlying ontology-driven and AI-based strategy can as such be applied to any type of chemical and systemic repeated dose toxicity effect. ONTOX will thus provide a fully functional and sustainable solution for advancing human risk assessment of chemicals leading to better protection of European citizens without the use of animals.

Highlights of the third reporting period include:
WP1 Published research papers on PM concepts and design, including liver PMs. Adapted a large language model with HU and VHP4Safety to help users navigate complex systems biology diagrams.
WP2 Delivered the first integrated, human-based clinical and mechanistic framework for predicting and characterizing drug-induced hepatic steatosis. Established a systems-toxicology platform for liver and renal injury, enabling mechanistic, predictive, and cross-compound toxicity assessments.
WP3 Implemented computational models in VEGA 1.2.5 (QSAR and fragment-based) and developed an API linking VEGA and ProtoPRED models to ONTOX Hub.
WP4 Explored HT-PBK in contexts including QIVIVE for DNT and integrated kinetic models (in vivo/in vitro) to develop a tiered QIVIVE framework for pesticide DNT assessment.
WP5 Advanced ToxIndex.com into an end-to-end, customizable agentic platform integrated with the ToxTransformer model for risk assessment and data gap filling. Task 5.3 delivered a publicly available, machine-readable nephrotoxicity AOP network. Adapted LleMy, a large language model, to explore complex molecular and systems biology maps, easing access for new users.
WP6 (NIPH-led) initiated a case study on probabilistic risk assessment of PFOA per published protocol. (3RSMC-led) progressed toward full integration of tools and methods in OPRA (ONTOX AI-supported probabilistic risk assessment).
WP7 Fully developed and characterized in vitro test batteries for detecting chemical-induced liver steatosis and cholestasis, including assays targeting molecular initiating and key events.
WP8 Published the tubular necrosis AOP and established in vitro assay batteries for case studies.
WP9 Characterized in vitro assays for DNT assessment and optimized a computational model of mammalian neural tube closure to predict defects using zebrafish embryo transcriptomics.
WP10 Coordinated ONTOX research and non-research activities, chaired ASPIS cluster coordination (ONTOX, PrecisionTox, RISKHUNT3R), and organized the ASPIS Open Symposium 2025 in Athens, Greece.
WP11 Created an ONTOX collection in BioStudies with 58 accessions (physiological maps, AOPs, ToxTemps, transcriptomics), 10 publicly available.
WP12 Delivered audiovisual and written content for dissemination, supported ONTOX meetings and stakeholder engagement. Exceeded dissemination and training targets via high-impact publications, conferences, and trainings, positioning ONTOX as a reference for next-generation risk assessment. Strategic collaborations and upcoming regulator-focused demonstrations ensure sustainability and accelerate uptake of animal-free methodologies.
WP13 Released a new ONTOX Hub and marketplace providing results and services from WP3–6, WP9, and WP11. Collaborated with RISKHUNT3R at ASPIS level to align ASPA and OPRA approaches and participated in workshops to prepare exploitation and sustainability phases.
WP14 Amplified ONTOX advancements via website, social media, ONTOX TV, and newsletters. Added new website sections (“IMPACT” and Working Packages) and launched ONTOX PARCOPEDIA.

Original and innovative aspects of ONTOX presented against the current state-of-the-art include:
1. State-of-the-art: focus on individual critical aspects of systemic repeated dose toxicity and full risk assessment for development of NAMs mainly in separate projects.
Innovation in ONTOX: ontology-driven and AI-based integration of all critical aspects of systemic repeated dose toxicity (biology, toxicology, chemistry and kinetics) and risk assessment (hazard identification, hazard characterisation and exposure assessment) in a single
collaborative project and resulting in 1 NAM per type of adversity.
2. State-of-the-art: frequent insufficient data for the development of AI systems to predict chemical hazard.
Innovation in ONTOX: generation of an advanced AI system for reliable chemical hazard identification fed by data-rich ontologies and read-across structure-activity relationships, which allows inclusion of heterogeneous big data.
3. State-of-the-art: traditional bottom-up approach for hazard identification in animals using high doses of chemicals starting from the apical toxicological outcome at the organism level.
Innovation in ONTOX: top-down approach for hazard identification and next generation risk assessment in human-based NAMs using physiologically relevant doses starting from basic biology at the molecular level.
4. State-of-the-art: 1-to-1 replacement of animal studies for toxicity testing with single non-animal methods.
Innovation in ONTOX: replacement of animal studies for toxicity testing by batteries of in vitro assays and in silico tools encompassing the full biological spectrum and reflecting the in vivo complexity of adverse effects.

The ambition of ONTOX is to strengthen Europe’s world-leading position regarding the development, exploitation, regulation and application of NAMs for human risk assessment of chemicals by empowering:
1. Scientific excellence: to move beyond the state-of-the-art by delivering an unprecedented set of NAMs.
2. Industrial innovation: to generate new value creation opportunities and market potential.
3. Regulatory pioneering: to introduce a paradigm shift in the validation and use of non-animal methods.
4. Societal prioritisation: to increase human chemical safety while fully replacing the use of animals.

logo-regular.jpg

Periodic Reporting for period 3 - ONTOX (Ontology-driven and artificial intelligence-based repeated dose toxicity testing of chemicals for next generation risk assessment)

Herunterladen Den Inhalt der Seite herunterladen