Table 6.
Energy rank for heterodimeric structures. Energy rank of the native structure, compared to its decoys, using various energy functions. a23 hetermodimers with their decoys, not used in the parameterization. The 20- and 6-class energy functions used are the ones optimized on OS2. bHeterodimers in the TS1 and TS2 data sets. For the structures in TS1 (bottom 11), the OS1 energy functions are used; for those in TS2 (top 6), the OS2 functions are used (i.e., we show cross-validated results). cFraction of successful series. D and D3 correspond to strong discrimination (native structure ranked first) and weak discrimination (native ranked among top three; see Methods).
| Blind heterodimer seta | TS1, TS2 setsb | ||||||
| PDB ID | 20Cl | Bastolla | 6Cl | PDB ID | 20Cl | Bastolla | 6Cl |
| 1ai7 | 1 | 1 | 1 | 1abr | 4 | 2 | 3 |
| 1as4 | 1 | 1 | 1 | 1aui | 1 | 1 | 1 |
| 1bpl | 1 | 1 | 1 | 1ay7 | 3 | 3 | 2 |
| 1clv | 2 | 1 | 2 | 1blx | 3 | 6 | 3 |
| 1dj7 | 4 | 4 | 4 | 1c1y | 3 | 4 | 3 |
| 1dow | 1 | 1 | 1 | 1dkf | 2 | 47 | 1 |
| 1eud | 8 | 9 | 7 | 1ugh | 1 | 3 | 1 |
| 1f0c | 1 | 1 | 1 | 1efv | 1 | 1 | 1 |
| 1ksg | 5 | 3 | 5 | 1ezq | 1 | 1 | 1 |
| 1ku6 | 3 | 2 | 4 | 1f3v | 67 | 145 | 36 |
| 1lot | 1 | 1 | 1 | 1fle | 19 | 4 | 9 |
| 1lpb | 20 | 8 | 16 | 1hdm | 1 | 1 | 1 |
| 1mhm | 1 | 1 | 1 | 1itb | 1 | 1 | 3 |
| 1mtp | 1 | 1 | 1 | 1phn | 1 | 1 | 1 |
| 1mzw | 3 | 3 | 4 | 1qav | 1 | 2 | 1 |
| 1npe | 4 | 4 | 2 | 1smp | 14 | 69 | 2 |
| 1nw9 | 1 | 10 | 1 | 1stf | 2 | 5 | 1 |
| 1n52 | 10 | 1 | 9 | Dc | 8/17 | 6/17 | 9/17 |
| 1o5m | 3 | 1 | 2 | D3 | 13/17 | 9/17 | 15/17 |
| 1qge | 1 | 1 | 1 | ||||
| 1rke | 1 | 1 | 1 | ||||
| 1rtj | 1 | 2 | 1 | ||||
| 1ubt | 1 | 2 | 1 | ||||
| 4sgb | 5 | 6 | 5 | ||||
| Dc | 13/24 | 13/24 | 13/24 | ||||
| D3 | 17/24 | 18/24 | 16/24 | ||||