A new dataset called Curate-DTI has been released, focusing on drug-protein bioactivity. This dataset distinguishes itself by including construct-aware protein targets, addressing limitations of using only UniProt accessions which can obscure distinct experimental sequences. Curate-DTI contains over 2.9 million measurements and supports both classification and regression tasks, with associated code and provenance information available. AI
IMPACT Provides a more precise dataset for training and evaluating AI models in drug discovery and protein interaction research.
RANK_REASON The item describes a new dataset and associated processing code published on arXiv, which is a form of research output. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →