Table 6.
Hunter's performance for side chain predictions in homology modelling of structures from CASP7
| Number of targets | RMSD, Å | Contact score | χ1 (± 15°) % | χ1+2 (± 15°) % | |
|---|---|---|---|---|---|
| Hunter | 28 | 4.1 | 17.7 | 63 | 41 |
| TS004 | 4.1 | 17.1 | 60 | 37 | |
| Hunter | 26 | 3.7 | 20.9 | 61 | 40 |
| TS020 | 3.7 | 18.9 | 60 | 38 | |
| Hunter | 27 | 4.0 | 19.2 | 62 | 41 |
| TS186 | 4.0 | 16.0 | 60 | 37 | |
| Hunter | 5 | 3.0 | 27.6 | 69 | 46 |
| TS191 | 3.1 | 20.8 | 64 | 39 | |
| Hunter | 6 | 3.8 | 18.5 | 61 | 37 |
| TS397 | 3.8 | 16.5 | 59 | 38 | |
| Hunter | 6 | 3.9 | 18.0 | 63 | 39 |
| TS556 | 4.0 | 17.8 | 60 | 38 |
The results of the top six groups in CASP7 (according to Read and Chavali [39]) were taken for examining Hunter side chain modelling accuracy in homology modelling. Best models submitted by each group for every target were collected and the average side chain prediction accuracy was determined (side chain RMSD, contact score, chi-angle prediction accuracy). Then, Hunter was used to rebuild side chains on the set same set of submissions, and side chain prediction accuracy was evaluated and presented below. For each set of submissions two rows are given. The first row shows side chain modelling accuracy with Hunter while taking backbone coordinates from the best model submitted by the corresponding group for each target. The second row shows performance in side chain modelling by the group in CASP7. The group name is given here according to the CASP7 experiment.