Alternative sources of data
Good Quality Literature Data
Literature data can include peer reviewed papers and technical documents submitted as part of larger dossiers by industry to provide a weight of evidence under certain circumstances to bodies other than Cefas, for example the US-EPA.
If a supplier wishes to use the information provided in a paper it should be made available to Cefas to be able to check the quality of the publication. A good quality publication should have a detailed protocol of how the tests were carried out and if an internationally accepted test protocol was used (for example if an OECD test protocol was used and whether the test laboratory was GLP accredited).
It is important that literature data is rated Klimisch 2 as determined in Klimisch et al (1997) A Systematic Approach for Evaluating the Quality of Experimental Toxicological and Ecotoxicological Data Regulatory Toxicology and Pharmacology Vol 25 (1):1-5.
Chemical suppliers should provide a Klimisch rating that they have carried out and the reasoning behind this by using the categories in the paper. The Klimisch assessment can be sent in as a word document as an appendix to the HOCNF form. Cefas reserve the right to accept or reject literature data as part of the registration process.
Weight of Evidence
Weight of evidence uses a larger number of literature sources of all types to provide the basis for a reasoned argument. Suppliers should provide the reasoned argument in a single document referring to the literature supporting their argument using multiple literature sources. The document should explain why the literature was chosen, what the literature shows and draw conclusions on how the literature information should be applied in the case of their chemical. The weight of evidence can be from literature sources using the same chemical (as defined by CAS and/or REACH sieve), or similar chemicals via READ-ACROSS
The document itself should be written as simply as possible, should reference the source document and page number of the supporting evidence so that the information is easy to find. If there is contradictory information, this should be highlighted together with a detailed reason why this information should be discounted by the regulatory assessor. Cefas reserve the right to accept or reject weight of evidence as part of the registration process if the information is not clear or not relevant. Chemical suppliers should be prepared to provide more evidence or more detailed explanation if required by the assessor.
‘READ-ACROSS’ argument is stronger if the READ-ACROSS is based on data gap filling whereby the chemical forms part of a series of structurally similar chemicals and the data for the target chemical is missing.
There are a number of computer programs that can be used to make a ‘READ-ACROSS’ argument, these include OECD Toolbox and TOX-READ (available on the OECD Toolbox and VEGA websites), both are free to download. A ‘READ-ACROSS’ argument must follow the ECHA guidelines on submission of non-test data, this will make the assessment of information provided easier to assess by Cefas, any further information or areas where information is lacking can be more quickly identified.
The ECHA guidelines provide a useful check list which will help in deciding whether the information provided is sufficient for registration purposes. Ultimately it is the role of Cefas to determine whether this information is sufficient for OCNS registration purposes. Weight of evidence accepted by other regulatory bodies may not be sufficient for Cefas registration and Cefas reserve the right to reject ‘READ-ACROSS’ and weight of evidence documentation.
Quantitative Structure Activity Relationship (QSAR)
Quantitative structure activity relationships (QSARs) relate chemical structure with a measurable physical or chemical characteristic, such as hydrophobicity, electrophilicity or toxicity, or a combination of these factors. This can be done either by trend analysis or data gap filling depending on the type of data and number of data available.
It is important that any QSAR data has been validated and details are provided on how the QSAR was validated. To see more details related to validations see below.
If submitting QSAR data it is important to fill in the QSAR reporting form. This form is based on the ECHA QSAR reporting form and has been amended to contain important instructions that make filling in the form easier, however full instructions on how to fill in the form are provided in the ECHA guidance document. In some cases such as OECD Toolbox and VEGA the report will be in IUCLID 5 or 6 format which follows the QSAR reporting form layout. Note in many cases these platforms allow free text to be entered into the report and this allows the chemical supplier who has carried out the modelling to add annotation detailing why a particular group of chemicals or species were discarded as part of data trimming. If ‘free text boxes’ do not exist this information should be provided as supplementary information as a word or pdf document.
Data should not be trimmed without explanation. It is important that annotation is made in order that Cefas understands the process that has been used to achieve the outcome.
Appropriate data should be used. If the QSAR outcome is to determine toxicity to fish it is important that only fish data is used, similarly with algae and invertebrate data. It is also important that the appropriate test is used.
It is important that the limitations of QSAR modelling are understood and taken into consideration the chemical type or family.
QSARs are models which answer a specific question based on the physico-chemical properties of a target chemical based on its structure.
Not all algorithms or models are suitable for all chemicals, an example of this would be the use of QSAR data to describe toxicity in UVCB chemicals or surface active chemicals (surfactants).
The data that is used in the QSAR model must be of good quality and be rated at least Klimisch 1 or 2. Ideally the test information should be available as part of the QSAR platform you are using to determine whether the data related to the structural feature is sufficiently strong.