Rating | Difficulty | Target | Fraction of Residues1 | % of Residues1 | RMSD1 (Ã…) | Number of Residues2 | RMSD2 (Ã…) | RMSD2 range (Ã…) | Models considered |
|---|
E | 4 | T0110/rbfa | 77/95 | 85 | 3.9 | 80 | 4.0 | 4.0–16.5 | 3 |
G | 5 | T0126/omp | 44/163 | 27 | 4.3 | 60 | 6.9 | 5.5–13.0 | 1,3 |
G | 5 | T0105/sp100 | 39/94 | 41 | 4.3 | 55 | 6.4 | 5.8–12.5 | 4,5 |
G | 5 | T0114/afp1 | 34/87 | 39 | 3.9 | 45 | 6.5 | 5.6–12.1 | 1 |
F | 5 | T0102/as48 | 33/70 | 47 | 4.3 | 70 | 8.9 | 7.3–12.5 | 1–5 |
- | 5 | T0102/as48 | 70/70 | 99 | 5.6 | 70 | 5.6 | 3.7–12.0 | 2 |
F | 5 | T0118/enrn | 28/149 | 22 | 4.5 | 40 | 6.7 | 4.2–10.5 | 2–5 |
E | 6 | T0097/er29 | 58/105 | 55 | 3.7 | 80 | 6.2 | 3.6–13.0 | 4 |
E | 7 | T0091/ybab | 50/109 | 56 | 3.0 |  |  |  | 1–5 |
E | 7 | T0106/sfrp3 | 49/125 | 39 | 4.0 | 70 | 6.2 | 5.2–13.6 | 3,5 |
G | 8 | T0098/sp0a | 47/119 | 40 | 3.5 | 60 | 6.0 | 3.6–11.7 | 2 |
G | 8 | T0086/ubic | 44/164 | 27 | 4.3 | 50 | 6.7 | 5.3–10.8 | 1–5 |
- The targets (column 3) are sorted by their difficulty (column 2) as provided by the CASP4 assessors (determined by the degree of similarity of the target protein to proteins with known structures). Shown also is a subjective evaluation of the quality of the model (column 1; E – excellent, G – good, U – useful, F – failure), the number of residues (over the total) evaluated using the criterion1 provided by the CASP4 assessors which considers non-consecutive Cα atoms in the calculation of the RMSD, the percentage of residues evaluated by criterion1, and the corresponding RMSD1; the number of residues evaluated using our criterion2 which considers only consecutive regions, the corresponding RMSD2, and the RMSD2 ranges for the regions of the same length in all the sampled conformations (some values are missing using this criteria since the experimental result was not provided to the predictors); and the models (out of a total of five) for which the evaluation/result applies. The second row/result given for T102/as48 is actually a "post-diction" made using the NMR secondary structure assignments available to all CASP4 predictors (we used predicted secondary structure with a three-state accuracy of 60% in the original prediction, resulting in a signi cantly worse model). Nine out of eleven predictions are good or excellent.
