Table 3.
Quality Comparison of the Decoys Generated by FB5-HMM, the 1st-Order CRF, the 2nd-Order CRF, and Rosetta
|
Test proteins |
FB5-HMM |
1st-order CRF |
2nd-order CRF |
Rosetta |
||||||
|---|---|---|---|---|---|---|---|---|---|---|
| PDB | L | α, β | Good | Best | Good | Best | Good | Best | Good | Best |
| 1FC2 | 43 | 2,0 | 17.1 | 2.6 | 49.1 | 1.94 | 85.4 | 2.28 | 36.5 | 2.72 |
| 1ENH | 54 | 2,0 | 12.2 | 3.8 | 22.4 | 2.32 | 69.3 | 2.21 | 44.8 | 1.23 |
| 2GB1 | 56 | 1,4 | 0.0 | 5.9 | 23.3 | 2.91 | 65.2 | 2.04 | 5.82 | 2.26 |
| 2CRO | 65 | 5,0 | 1.1 | 4.1 | 16.8 | 2.79 | 35.4 | 2.58 | 17.2 | 2.38 |
| 1CTF | 68 | 3,1 | 0.35 | 4.1 | 2.4 | 3.70 | 6.62 | 3.67 | 2.35 | 1.30 |
| 4ICB | 76 | 4,0 | 0.38 | 4.5 | 0.51 | 4.63 | 0.125 | 4.40 | 4.51 | 3.90 |
For each protein, 100,000 decoys are generated by FB5-HMM, while only ∼20,000 decoys by each CRF model and Rosetta. No energy function is used in this comparison. Columns 1–3 list name and PDB code, size and number of α-helices, and β-strands of the test proteins. Columns “Good” and “Best” list the percentage of good decoys (with RMSD ≤ 6 Å) and the RMSD of the best decoys, respectively.