Table 2.
Results from a 28 protein benchmark.
| Benchmark results | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| RMSD of first pred. | RMSD of best pred. | 1% RMSD | 1% Energy | ||||||
| Cont | Resamp | Cont | Resamp | Cont | Resamp | Cont | Frag | Resamp | |
| 1di2 | 6.03 | 4.26 | 1.50 | 3.58 | 2.65 | 2.16 | − 138.68 | −138.80 | −132.45 |
| 1dtj | 10.37 | 2.79 | 2.47 | 2.08 | 2.93 | 2.09 | −144.73 | −149.70 | −145.95 |
| 1dcj | 5.22 | 2.50 | 4.99 | 2.50 | 4.13 | 2.44 | −131.90 | −134.96 | −122.43 |
| 1ogw | 4.67 | 3.46 | 3.06 | 3.21 | 3.11 | 3.14 | −152.25 | −152.97 | −146.72 |
| 2reb | 1.33 | 1.27 | 0.74 | 1.07 | 1.23 | 2.04 | −134.20 | −134.89 | −125.89 |
| 2tif | 4.18 | 3.98 | 4.05 | 3.98 | 3.15 | 3.68 | −118.01 | −114.54 | −105.25 |
| 1n0u | 10.56 | 3.73 | 3.14 | 3.73 | 3.76 | 3.11 | −129.61 | −132.85 | −127.63 |
| 1hz6A | 3.15 | 3.50 | 3.12 | 2.53 | 2.41 | 2.08 | −132.58 | −135.33 | −127.00 |
| 1mkyA | 5.08 | 6.21 | 3.64 | 4.88 | 3.75 | 4.04 | −152.32 | −155.47 | −145.47 |
| 1a19A | 3.60 | 11.34 | 3.60 | 11.16 | 3.47 | 5.99 | −176.04 | −179.58 | −168.98 |
| 1a68 | 15.01 | 7.76 | 8.80 | 6.94 | 6.37 | 6.80 | −170.08 | −177.38 | −169.37 |
| 1acf | 11.18 | 4.11 | 11.18 | 2.40 | 6.75 | 4.38 | −233.91 | −241.46 | −222.27 |
| 1aiu | 1.60 | 1.50 | 1.60 | 1.50 | 1.72 | 1.62 | −208.86 | −215.75 | −205.03 |
| 1bm8 | 12.13 | 13.53 | 5.27 | 5.00 | 5.61 | 5.48 | −197.36 | −203.51 | −189.84 |
| 1cc8A | 3.94 | 4.63 | 2.52 | 3.52 | 2.73 | 2.60 | −138.04 | −139.63 | −124.08 |
| 1bq9A | 5.83 | 7.78 | 3.58 | 3.17 | 4.77 | 3.77 | −83.03 | −84.54 | −82.59 |
| 1ctf | 8.97 | 4.13 | 6.08 | 2.67 | 4.20 | 3.03 | −141.52 | −144.24 | −135.21 |
| 1ig5A | 3.73 | 2.82 | 3.73 | 2.71 | 3.01 | 2.32 | −156.67 | −159.15 | −157.16 |
| 1iibA | 10.12 | 4.66 | 3.54 | 4.56 | 3.19 | 3.54 | −203.45 | −204.30 | −190.95 |
| 2ci2I | 9.42 | 6.65 | 6.50 | 6.34 | 4.51 | 5.44 | −123.48 | −127.99 | −111.98 |
| 2chf | 3.96 | 3.06 | 3.08 | 3.06 | 3.59 | 3.00 | −264.57 | −266.76 | −251.24 |
| 1opd | 4.27 | 3.08 | 3.82 | 1.52 | 3.65 | 2.36 | −166.97 | −171.71 | −165.78 |
| 1pgx | 3.10 | 3.66 | 0.867 | 1.71 | 1.61 | 1.34 | −118.16 | −118.70 | −111.27 |
| 1scjB | 2.66 | 6.36 | 2.61 | 6.06 | 2.89 | 3.41 | −132.38 | −136.47 | −124.93 |
| 1tig | 11.66 | 4.17 | 11.14 | 3.06 | 3.91 | 3.04 | −179.02 | −179.10 | −167.46 |
| 1ubi | 9.23 | 3.81 | 3.27 | 3.56 | 3.02 | 2.75 | −141.45 | −143.57 | −138.16 |
| 5croA | 6.36 | 4.22 | 2.92 | 3.37 | 3.26 | 2.96 | −110.43 | −111.41 | −107.56 |
| 4ubpA | 5.20 | 4.10 | 4.92 | 4.10 | 4.26 | 4.41 | −198.71 | −201.91 | −195.62 |
| Mean | 6.52 | 4.75 | 4.13 | 3.71 | 3.56 | 3.32 | −156.37 | −159.17 | −149.94 |
The results in the initial four columns show the RMSD of the first and best-of-five predictions for control (control) and resampled (Presamp) populations. Top five predictions were made by selecting the lowest-energy structures from the five largest clusters. In resampling, fragments were repicked according to the output of the torsion predictor. Beta topology, registers, and contacts were stochastically constrained according to the output of the beta sheet feature predictors. The results in the next two columns show first percentile RMSD for control and resampled populations. The final three columns show first percentile energy for these populations, in addition to models generated using repicked fragments but no beta constraints.