Proposed by @mkgilson with support from @davidlmobley and @jchodera.
A dataset or set of datasets with drug- or lead-like compounds from drug discovery campaigns is of interest for benchmarking studies.
The source is BindingDB, with a selection 980K compounds.
From @mkgilson:
These are not filtered by number of heavy atoms. If a compound has multiple binding data, then it is repeated across rows. If you sort by BindingDB MonomerID, this will group rows by molecule, which might be convenient. It has a bunch of data and metadata that are not of high interest, but it includes UniProt IDs for the Targets, as requested by @jchodera.
Proposed by @mkgilson with support from @davidlmobley and @jchodera.
A dataset or set of datasets with drug- or lead-like compounds from drug discovery campaigns is of interest for benchmarking studies.
The source is BindingDB, with a selection 980K compounds.
From @mkgilson: