PROJECT DESCRIPTION
BACKGROUND
The main policy landscape of LIFE CONCERT REACH is the EU chemicals regulation, the so called REACH regulation, which in practice compels the industry to assess the safety of the chemicals it produces or imports. The regulation poses several challenges, which the project seeks to address. In particular, the regulation raises the need to use alternative methods to protect environmental and human health, including the application of innovative non-testing methods (NTMs), and to determine the impact of substances with reliable environmental indicators. At the same time, since May 2018 (the deadline for the registration of substances under the REACH regulation), a huge amount of REACH experimental data has become available, which now needs to be better exploited, especially through NTMs.
OBJECTIVES
The main objective of the project was to evaluate the potential impact of chemical substances in Europe, before and after REACH, by integrating experimental data on registered substances into computer data processing (in-silico) tools. In particular, the project established an integrated network of systems offering non-testing methods (NTMs) for REACH.
Quantitative Structure-Activity Relationship (QSAR) models, which the network offers, could be used to support the regulatory assessment of chemicals. The network brought together four tools widely used and supported by authorities and industry: the Danish (Q)SAR database for in-silico models, the VEGA platform, the AMBIT database for the read-across workflow and data from the registered substances, and the OCHEM platform. These were integrated into the largest network of in-silico tools in the world, with the aim of reshaping the strategy of evaluating chemical substances.
RESULTS
The project ended successfully, reaching all its main objectives and results. In particular, it established an integrated network of systems offering non-testing methods (NTM) for REACH purposes. The network includes in-silico models from VEGAHUB, the Danish (Q)SAR database and OCHEM, and read-across workflow and data from the registered substances by AMBIT. It is the largest network of in-silico tools in the world (>450 models plus other tools for read-across), and can support the assessment, by both regulators (assessors) and industries (in the preparation of REACH dossiers), of chemical substances.
More specifically, the following achievements can be mentioned:
- A collection of sets of experimental data obtained from REACH registration (both in AMBIT and in public databases) was extracted and quality checked for structural and biological consistency, resulting in the curation of two big datasets for a series of selected endpoints: melting point, skin sensitisation, skin irritation, steroidogenesis, activated sludge, terrestrial toxicity- earthworm, and eye irritation. A harmonized protocol to relate experimental values of high quality with a structured procedure to document the development of QSARs (Quantitative Structure-Activity Relationship) was created in parallel.
- 30 models were implemented in VEGAHUB, based on already existing tools.
- More than 40 new models (vs. 20 foreseen) were tested and their applicability to the endpoints identified in the data collection exercise examined. More precisely, their predictive power and their suitability for regulatory purposes was assessed and validated on REACH registrations compounds. The various tools were optimised, making them freely available and more user-friendly. A final subset of the 12 best performing models was then selected and implemented in VEGAHUB.
- Rules to be implemented in ToxRead (a read across programme in VEGA) were collected from different sources. Among all the rules collected, the 10 rule sets’ modules that have the best performance and that cover interesting endpoints in terms of regulation and because not yet in the tool were selected and implemented.
- A new programme to use data on registered substances without disclosing proprietary data, bringing together Structural, Toxicological and Grouping Similarities, was developed (VERA) and, with the help of a case study, was fine-tuned ensuring sufficient data to all datasets. Actual implementation within VEGAHUB pursued.
- A weight of evidence (WoE) tool - called SWAN - was completed, integrating results from multiple sources, involving testing on two case studies. This achievement was possible also thanks to the experience gained during the collaboration of partners with EFSA in the assessment of a set of botanicals.
- The web-based “gateway”, integrating VEGAHUB, AMBIT, the Danish (Q)SAR and OCHEM, guides the users through the four in-silico platforms and reports all their predictive software available relative to REACH endpoints. It was developed and made public, also including the predictions of 18 VEGA QSAR models, for more than 600,000 chemical substances.
- The impact of REACH on production and import of chemical substances in 2011-2018, and in particular CMR and PBT substances, has been measured, providing interesting insights to policy makers. The team also analysed the impact of REACH on the Italian market considering the presence of chemicals, their diversity, import and production during the period from 2011 to 2020.
- The tool ClusteringSIM was applied to some endpoints in datasets of registered substances, while the VERA software was applied to carcinogenicity and aquatic toxicity toward fish, allowing the evaluation of families of chemicals with higher and lower concern for those endpoints.
- A series of recommendations for the revision of the description of the model and QSAR results reporting and for the optimisation of the output of the NTM within VEGAHUB was defined, with the help of two case study endpoints, with the aim of improving and facilitating the preparation of REACH dossiers (for industries) and harmonising approaches with authorities.
- Two case studies were assessed for the application of in-silico tools: (i) for authorities - on perfluorinated compounds, conducted in collaboration with various Italian authorities; (ii) for the industry - to demonstrate how to access the models integrated in the gateway, as well as how to analyse the results and report them in IUCLID, for REACH dossier preparation.
Policy makers at national and European levels were constantly engaged and informed on results by different means. These include ECHA, EFSA, JRC, OECD, US EPA, Danish EPA, SCCS (Scientific Committee on Consumer Safety of the EC), EMA (European Medicines Agency), and many others, also out of Europe. As a result of this intensive work, it is quite likely that at least some of the project outputs (i.e., the in-silico tools and datasets) will be integrated within other IT systems managed by regulators: the Danish QSAR Database and VEGA are indeed already available within the OECD Toolbox, VEGA will be linked soon to EFSA for the use within the OpenFoodTox platform of EFSA (and was also offered to the US EPA, though TEST), while SCCS mentioned the project’s Gateway in its official Guidance for cosmetics.
Industries were reached mainly with the support of the beneficiaries KNOELL and SC and, by project end, at least 9 industries (vs. a target of 5) applied the tools developed for the assessment of chemicals in addition to 4 consultant companies (working for chemical companies) which use in silico tools including VEGA for substances assessment. These stakeholders were engaged through workshops and webinars, involving different sectors (e.g., glasses, cosmetics, adhesives, inks and dyes and generally chemical companies, members of Federchimica, the Italian chemicals association).
Finally, the project provided capacity building sessions to various stakeholders, increasing their knowledge on NTMs. In particular, 40 regulators attended a workshop organised by DTU; >100 industrial associations and authorities attended the final workshop’s training, and professionals attended a course about practical applications on cosmetics. In addition, more than 100 students received graduate/postgraduate training.
