ECHA does not intend to produce a list of ‘approved’ models, because the value of a model depends on how it is used. (See FAQ: ‘What makes a good QSAR model?’) Every user of QSARs needs to be aware that QSAR models are only appropriate and reliable for specific sets of chemicals. A highly reliable model will not produce reliable results for chemicals that lie outside the domain of applicability. In addition, models may be suitable for different regulatory functions: risk assessment, classification and labelling or prioritisation, because each makes different demands on the model.
For decades, QSAR models have been used in the USA to evaluate a series of properties of chemicals. Indeed, Section 5 of TSCA (Toxic Substance Control Act) requires a manufacturer and/or importer of a new chemical substance to submit a premanufacture notice (PMN) to the US EPA 90 days before commencing the manufacture or import of a new chemical. Decisions have often been taken without further experimental data. The US EPA instigated and promoted the development and use of a series of QSAR models to predict properties of interest.
In Denmark, the Danish Environmental Protection Agency has developed and used QSARs for regulatory use. (See FAQ: ‘Are QSARs expensive or free to use’.)
QSARs have been used for decades in the development of pharmaceuticals, where a drug is to be developed to achieve a particular biological action. Our interest here is in the reverse use, where the chemical is known, and the biological action is to be predicted. That too has been the focus of research for decades. But here we are concerned specifically with their use for evaluating toxicity within the regulatory framework of REACH, and it is early days.
In practice, the use of in silico methods by European industry within REACH is still limited. There are three key practical issues that can delay their use. They are highly inter-connected.
Unlike industry working within a regulatory context, toxicology and ecotoxicology researchers have the freedom to select and/or develop QSAR models specifically for their scientific use, and do not need to meet the regulatory demands for transparency.
In silico methods, and specifically QSAR models, can be used in a range of ways within research. QSAR models provide statistical evidence of patterns of toxicity across a series of chemicals, and so, for example, can provide an agenda for empirical research into the mechanisms of action by which particular molecular properties generate those observed biological and environmental effects.
The REACH legislation puts the responsibility on industry to provide the necessary toxicity information on each substance which they manufacture, distribute and market, and to assess and manage the risks linked to those substances. This is the principle of ‘no data, no market’. Industry are therefore a key stakeholder in influencing the future use of in silico methods.
For further information, see the page ‘Why use in silico methods?’ on this site. It outlines why industry needs to use in silico methods, or at least know about them.
Many QSAR models have been developed and some are publicly available. Depending on the purpose different models should be used. For instance, for REACH certain requirements have been defined (link to above). For regulatory purposes transparency is highly preferred, in order to provide full information on the model. This also complies with the OECD principles (link). As a consequence, the user should verify if full details on the model are available. For commercial models there may be limitations.
These are some examples of publicly available models developed within the EC projects evaluated within ORCHESTRA.