The topic of Big Data has strongly entered into predictive and computational toxicology, and is here in order to stay. However, not only the data is big, large are also computational models (the so-called QSAR models) derived from these data. Therefore, the management of computational models in the context of Big Data needs more attention and efforts, so that the models remain transparent and independently verifiable.
In year 2018 the third QsarDB workshop was held. The year culminated with the publication of a scientific article, which documented the "Best Practices for QSAR Model Reporting", which was also a summary of QsarDB team's experience in recent years on how QSAR models are documented in the scientific literature and also made suggestions for improving the current state-of-the-art.
In addition to active development work, digitalization of QSAR models and many presentations, a research article was published that describes the QsarDB data repository and its design principles as a whole.
2014 witnesses many conference presentations, the first workshop (training school), and also the integration of QsarDB repository with the DOI system. In 2014, a scientific article describing the QsarDB data format was published.
This year was special for QsarDB for at least two reasons. The first one is certainly an article in Journal Chemosphere that linked a QSAR model to the repository with a unique persistent digital identifier (HDL). The second was a funding from the EU Regional Development Fund, which significantly boosted activities over the next three years.