Table 4.
Common datasets used in benchmarking studies for pocket comparisons.
| Purpose | Name | Content | # Positive (# Negatives) |
|---|---|---|---|
| Pairs of cavities from dissimilar proteins binding identical or similar ligands (positives) and dissimilar ligands (negatives) | APoc set [28] | Diverse | 38,066 (38,066) |
| Barelier et al. [144] | Diverse | 62 | |
| Homogeneous [116] | Diverse | 100 | |
| Kahraman [146]/extended [116] | Cofactor sites | 100/972 | |
| sc-PDB subset [47] | Diverse | 1070 | |
| TOUGH-M1 [145] | Diverse | 505,116 (556,810) |
|
| TOUGH-C1 [128] | Nucleotides, heme, steroid sites | 2218 | |
| Pairs of proteins sharing 3 high affinity ligands (potency < 100 nM) vs. pairs of proteins sharing 3 ligands with divergent affinities | Vertex [133] | Diverse | 6598 (379) |
| Vertex refined [129] | Diverse | 338 (338) |
|
| Pairs of cavities of associated with the same (positives) or different (negatives) functions and fold class | sc-PDB subset [24] | Diverse | 769 (769) |
| sc-PDB subset [121] | Diverse | 766 (766) |
|
| sc-PDB subset [129] | Diverse | 383 (383) |
|
| Intra-family classification | Proteases, kinases, GPCRs, Estrogen receptors [17,20,47,115,148] |
- | |
| Difficult cases | Difficult cases [19,24] | Diverse from experimental validations | 8 |
| Successful applications | ProSPECCTs D7 [38] | Diverse from experimental validations | 115 (56,284) |
| Structures of identical sequences | ProSPECCTs D1 [38] | Diverse | 13,430 (92,846) |
| ProSPECCTs D1.2 [38] | Diverse | 241 (1784) |
|
| NMR structures | ProSPECCTs D2 [38] | Diverse | 7729 (100,512) |
| Artificial sets: random mutations | ProSPECCTs D3 and D4 [38] | Diverse | 13,430 (67,150) |