TABLE 1.
Main data sets providing protein-ligand complexes (crystal structures) and corresponding binding affinities. N is the number of protein-ligand complexes (co-crystal structures) with associated binding affinities.
| Data Set | N | Superset | Website |
|---|---|---|---|
| PDBbind 2020 | 19 443 | — | pdbbind.org.cn |
| CASF-2016 | 285 | PDBbind 2016 | pdbbind.org.cn |
| CASF-2013 | 195 | PDBbind 2013 | pdbbind.org.cn |
| CASF-2007 | 195 | PDBbind 2007 | pdbbind.org.cn |
| Binding MOAD 2020 | 15 223 | — | bindingmoad.org |
| CSAR-NCS HiQ | 343 | Binding MOAD + PDBbind | csardock.org |
| CSAR-NCS HiQ Update | 123 | Binding MOAD + PDBbind | csardock.org |
| Astex Diverse Set | 74 | — | doi.org/10.1021/jm061277y |
| BindingDB | 11 442 | — | bindingdb.org |
| D3R GC 4 | 20 | — | drugdesigndata.org |
| D3R GC 3 | 24 | — | drugdesigndata.org |
| D3R GC 2 | 36 | — | drugdesigndata.org |
| D3R GC 2015 | 24 | — | drugdesigndata.org |