Table 1. Molecular Fingerprintsa.
| ID | name | description | features | reference(s) |
|---|---|---|---|---|
| FP1 | AP2D | topological atom pairs | 1211 | (44) |
| FP2 | ASP | all-shortest paths | 26,194 | (45) |
| FP3 | AT2D | topological atom triplets | 56,963 | (44) |
| FP4 | DFS | all-paths (depth-first search) | 48,448 | (46) |
| FP5 | ECFP | extended connectivity fingerprints | 42,672 | (47) |
| FP6 | LSTAR | local path environments | 85,232 | (48) |
| FP7 | MACCS | MDL public keys (166 keys) | 155 | (49) |
| FP8 | PHAP2POINT2D | topological pharmacophore pairs | 17 | (50) |
| FP9 | PHAP3POINT2D | topological pharmacophore triplets | 302 | (50) |
| FP10 | RAD2D | topological molprint-like fingerprints | 92,191 | (48) |
| FP11 | RDKit | topological daylight-like fingerprints | 65,183 | (43,51) |
a
A total of 11 different molecular fingerprints were generated using the jCompoundMapper tool42 or RDKit toolkit (version 2020.09.4)43 for describing the compounds in our datasets. The fourth column, features, represents the total number of features that the jCompoundMapper tool or RDKit toolkit generated for describing our RIKEN high-confidence set (826 compounds). The molecular features that we counted were only those present in the description of at least one compound of this collection, and the molecular fingerprints that required a depth of description were measured at depth 8.