--- title: Concept for the data repository subtitle: Revision 10 June 2026 author: Johannes Ranke and Marcel Mathis bibliography: ../inst/REFERENCES.bib format: html: toc: true code-fold: true vignette: > %\VignetteIndexEntry{Concept for the data repository} %\VignetteEngine{quarto::html} %\VignetteEncoding{UTF-8} --- The idea of this data repository is to have a commonly maintained source of consolidated data to be used in the environmental risk assessments of plant protection products, that is - validated by internal (at Agroscope) or external peer review - version controlled, with the possibility to do comparisons to other versions - reproducible, in the sense that each data point can be tracked to a source that is - publicly available and/or - internally available at Agroscope At the time of this writing (March 4 2026), all endpoints contained in the `derappp` data object can be tracked to a publicly available source. Technically, the data are part of the R package `derappp` and are immediately available to the user after installing and loading the package. ## Previous work A number of Agroscope projects and external projects have collected data with similar aims. In the following, some of those projects are briefly sketched and their strengths and weaknesses are discussed. ### PIERIS The Pesticide and Intermediates Ecotoxicological Risk Information System (PIERIS) originally developed at Agroscope and currently used by the Federal Office for the Environment (FOEN) and Agroscope in the context of product registration contains a comprehensive database of information necessary for the environmental risk assessment of products, active substances and their metabolites. It has grown over time and new features and data fields have been added to keep up with regulatory requirements. Therefore, it could be an excellent starting point for the calculation of environmental risk indicators or risk potentials. Strong points: - Available data for all registered active substances should be sufficient to do risk assessments - Data were carefully evaluated, as the data entered were directly relevant to regulatory risk assessments - Unpublished data on products is included, which may not be publicly available elsewhere. The data on solo formulations would be of interest for the evaluation of potentials of the active substances contained. Weak points: - Contains unpublished data, so it needs to be clarified if it can be used and/or published i.e. in supporting information. These unpublished data do not include product compositions, as only the active substance content is given in PIERIS, which is already publicly available e.g. in the Swiss Register of Plant Protection Products. They do include data on products that have potentially not been published elsewhere. - Data on active substances are not up to date with current EFSA conclusions in many cases. They usually reflect the current study situation on the date when these products were evaluated by experts. - For many data types, there is only one endpoint per active substance, e.g. for vapour pressure or water solubility. In other cases, there are two (Koc) or more columns (DT50xxx) but they have not been consistently used, e.g. the Koc column often has the lowest Koc, but it can also be a mean, and the mean Koc column has obviously a mean, but sometimes it has the lowest Koc. - For ecotoxicological endpoints, only the type of source is recorded (GLP study, literature reference, ...). In about 10 000 of the 27 000 cases an additional reference code is given in the table, but this information is really short, and the type of source is not obvious. Likely, in many cases it refers to unpublished regulatory reports. In some cases, references to EFSA conclusions are made in comments. So all in all, the source referencing is incomplete and not consistently formatted. - There is no version control in the sense that changes from one time point to the next are not tracked. ### PPDB/BPDB Agroscope has a subscription to the Pesticide Property Database (PPDB) and Biopesticide Properties Database (BPDB) which includes the delivery of the complete database in MS Access format, containing the data [publicly available](http://sitem.herts.ac.uk/aeru/ppdb/en/index.htm). Strong points: - Data available also on older active ingredients that are not on the market any more. - Environmental fate and ecotoxicology in water and soil are covered. - The data is publicly accessible on the Internet. A licence can be purchased for the use of the data as a database. - Versioning is possible by using the various MS Access versions of the PPDB database. - The subscribed version has detailed data in different soil types and pH ranges for substances for soil absorption and soil degradation parameters Weak points: - Generally, either one value or a conservative value is given for each category. For example, if there are several acute fish toxicity values, only the lowest value is given. - The source is only given in categories (e.g. data used in regulatory assessments). The exact source where a specific endpoint can be found is not given and it is not intended by the maintainers of that database to add that information. - Data can be changed from version to version without notice, e.g. the soil DT50 of certain copper salts was changed from 100 000 days to 0.1 days. - Decisions on which data were included are not documented - Copyright issues may prevent extracts of the PPDB database to be published in the supporting information of scientific publications. ### OpenFoodTox Data source maintained on behalf of EFSA with a web interface and some Excel sheets published on [Zenodo](https://zenodo.org/records/8120114). An [R package](https://github.com/agroscope-ch/openfoodtox) is maintained in the Agroscope github account to facilitate programmatic access to the data. Among other content, there is a large table of so-called reference points, Strong points: - Only data listed in EFSA outputs are included, generally indicating a high degree of reliability - The source (i.e. the corresponding EFSA output) is always given - Reasonable table definitions with clear structure and column labels - Creative Commons Attribution licence allowing to republish the complete dataset if the source is accredited. Weak points: - No environmental fate data included - For each reference point, only one endpoint is used, which is usually the lowest endpoint in its category. Therefore, no geometric mean endpoints or even species sensitivity distributions can be calculated from these data. - Names of biological species are generally in English instead of using the scientific names - The latest version is from 13 September 2023 and it is unclear if and when an update with recent EFSA conclusions will be made ### EcoTox Database maintained by the US EPA. Contains over one million results on more than 12 000 chemicals at the time of this writing (June 2024). Strong points: - Large amount of data, many endpoints for many species for each substance - Publicly available in the form of ASCII text file - Using the data has recently been simplified by the [https://github.com/andschar/standartox] project Weak points: - Only primary sources are included. As far as we have seen so far, this means that data reported in GLP studies as given in EFSA conclusions or EU Renewal Assessment Reports are not included. ## Technical concept The repository structure is that of an R package. Therefore, it is sufficient to install the package to be able to work with the latest version of the data. The central data object for collection of reviewed, version controlled and fully referenced data is the `derappp` object. It contains several tables of data with primary keys and foreign keys, making it possible to ensure referential integrity within the database. ### Peer review The coordination of peer review currently takes place via e-mail exchange among the contributors. In the future, it can make use of issues in a code development system. ### Version control Reviewing differences between versions is made possible by exporting the content of each table in the JSON format which is readable for humans. These JSON files are checked into the git version control system. Using git and the associated tools to show differences between text files, the versions can transparently be compared. ### Source referencing Each endpoint has a reference to a source in the sources table. Chemical identity information was collected using the packages 'webchem' [@szocs2020webchem] and 'chents' [@ranke2026chents] from the Compendium of Pesticide Common Names [@BCPC_Compendium], the PubChem website [@PubChem], the CAS Common Chemistry website [@CAS_Common_Chemistry] and the envipath website [@envipath]. The preferred source of information are EFSA conclusions, e.g. @j.efsa.2006.51r, respectively the Listings of Endpoints (LoEP) that are published as Appendix to such EFSA conclusions, e.g. @j.efsa.2016.4610_LoEP. For the soil toxicity data, the (draft) Renewal Assessment Reports were frequently consulted in addition, as some details that we consider important are often lacking in the LoEP published by EFSA. ## References