TDC

Why we included this project

Therapeutics Data Commons is about as close as the field gets to a standard starting point for machine learning in drug discovery. It gathers dozens of AI-ready datasets spanning target discovery, activity screening, efficacy, safety, and manufacturing, so a team can pull a benchmark in a few lines of code instead of scraping raw molecular data and reformatting it by hand. The package also includes sensible train/test splits, evaluation metrics, and molecule generation oracles, which makes it much easier to compare methods on equal terms. Anyone working on small molecules, antibodies, vaccines, or drug repurposing will find the curated tasks and leaderboards a solid base for checking whether a new model actually generalizes. It is also an open-science project, and the dataset keeps growing through community contributions, which matters if you plan to build on it long term.