Background The EU REACH legislation introduced stricter requirements for the registration, evaluation, authorisation and restriction of new chemicals, with the aim of improving the protection of human health and the environment. With at least 30 000 new chemicals introduced in subsequent years in Europe, it is questionable whether the number of laboratories is sufficient to cope in a reasonable time with the necessary toxicity testing. Another consequence is that animal testing is likely to increase as companies seek to demonstrate compliance. Some 10 million animals are used every year for laboratory experiments in Europe, mainly by the drug and chemicals industries. Initial estimates are that REACH will increase animal use by 39%; this could be even higher considering the number of pre-registered compounds. To avoid these problems, REACH promotes the use of non-testing methods, but more information is required on how and when to use them. Objectives The ANTARES project aimed to show which non-testing methods (NTMs) can be used to demonstrate compliance with REACH legislation and under what conditions. It sought to bridge the knowledge gap on which methods can be used to avoid animal testing. The project aimed to carry out a preparatory survey of all current methods for assessing compliance with the REACH legislation, and to identify at least 25 NTMs and validate them in terms of toxicity, ecotoxicity and environmental values. The final project objective was to define a unified and organised strategy for the use of multiple NTMs, taking advantage of the strengths of each method and taking into account any conflicts or contradictions emerging from the results. Results The ANTARES project demonstrated the technical viability of non-testing methods (NTMs), with particular reference to Quantitative Structure-Activity Relationships (QSARs), for assessing the toxicity and ecotoxicity of new chemicals submitted under REACH legislation. The innovative character of the project mainly resides in having quantified, for the first time, the efficiency of in-silico NTMs (i.e. those performed on computer or via computer simulation). This provided reliable information on how, and in which cases, the use of software models could be performed, according to the endpoints and to the chemical features of the new chemicals. The project reported good overall performance of the in-silico NTMs tested. The project implemented a survey of current methods used to assess the compliance with REACH legislation. The beneficiaries demonstrated that the private laboratories involved are, in general, unable to perform all the tests required to assess the toxicity and ecotoxicity of new chemicals. The project then identified QSAR models that were available, and defined criteria, methods and datasets to assess their accuracy. The beneficiaries identified 55 QSAR models that were tested for 8 selected endpoints foreseen by the REACH regulation, namely, water solubility, bio-concentration factor (BFC), mutagenicity, ready biodegradability, carcinogenicity, short-term toxicity in fish, Lethal Dose 50 in fish, and short-term toxicity in Daphnia. The accuracy of the models was tested using a dataset of chemical compounds with known toxicity and ecotoxicity values for these 8 endpoints. The tests were carried out both within the applicability domain of the models (using a set of chemical compounds with features similar to the ones used to build the model) and outside it (using a random set of chemical compounds). The project team demonstrated a high level of accuracy in the prediction of toxicity. A complete list of the most accurate models for each endpoint is freely available on the project website (www.antares-life.eu). The project beneficiaries implemented a strategy of integration for the QSAR models, in order to increase both performance and coverage of their applicability. This showed that combining different approaches may produce better results, depending on the endpoints selected. When a certain model prevails over the others, there is no real advantage of integrating multiple models. Conversely, if the results are equally good, and the models are different, it is suitable to apply methods of consensus or hybrid models. In terms of environmental and economic effects, the possibility of substituting conventional test with in-silico NTMs could provide a huge saving in terms of animals and in terms of testing costs for registrants and industries. For that reason, the project results generated a strong interest among the stakeholders both from the private and public sectors. To increase the use of the recommended NTMs, the project conducted targeted communication and awareness-raising activities. The main constraints threating the wide use of NTMs, as outlined by the project findings, are that not all the software models are sufficiently reliable to be immediately used, and that the acceptance of the regulatory bodies towards their use is still very low even when the accuracy is scientifically proven. The work undertaken by the ANTARES project team will be further developed in two ongoing LIFE projects, CALEIDOS (LIFE11 ENV/IT/000295) and PROSIL (LIFE12 ENV/IT/000154), which focus on different aspects of the NTMs in order to further overcome existing barriers to their wide application. The project contributes to the implementation of Directive 86/609/EEC on the protection of animals used for scientific purposes, by promoting alternative approaches to their use. From a general legislative point of view, ANTARES contributes to covering the knowledge gap left by the REACH legislation on the use of QSAR models for the registration of chemicals. Further information on the project can be found in the project's layman report and After-LIFE Communication Plan (see "Read more" section).