Papyrus - A large scale curated dataset aimed at bioactivity predictions

With the recent rapid growth of publicly available ligand-protein bioactivity data, there is a trove of viable data that can be used to train machine learning algorithms. However, not all data is equal in terms of size and quality, and a significant portion of researcher’s time is needed to adapt the data to their needs. On top of that, finding the right data for a research question can often be a challenge on its own. As an answer to that, we have constructed the Papyrus dataset (DOI: 10.4121/16896406), comprised of around 60 million datapoints. This dataset contains multiple large publicly available datasets such as ChEMBL and ExCAPE-DB combined with several smaller datasets containing high quality data. The aggregated data has been standardised and normalised in a manner that is suitable for machine learning. We show how data can be filtered in a variety of ways, and also perform some baseline quantitative structure-activity relationship analyses and proteochemometrics modeling. Our ambition is this pruned data collection constitutes a benchmark set that can be used for constructing predictive models, while also providing a solid baseline for related research.