Table 6.
Prediction accuracies for the three data sets shown in Table 1
Data set | Prediction accuracy,*
%†
|
Prediction accuracy,* % assuming all
hydrophobic residues are buried
|
|||||
---|---|---|---|---|---|---|---|
≥95% burial | ≥90% burial | ≥80%‡ burial | ≥80%§ burial | ≥95% burial | ≥90% burial | ≥80% burial | |
Original | 80 | 80 | 84 | 78 | 52 | 62 | 75 |
Test1 | 100 | 60 | 75 | 89 | 46 | 54 | 69 |
Test2 | None found | 100 | 83 | 100 | 44 | 52 | 65 |
Buried residues in each data set are predicted using the criteria in Table 4 and the accuracies are calculated from the percentage of predicted residues that are actually buried in the protein crystal structure.
Percent accuracy = 100% (number of predicted buried residues actually buried)/(total number of predicted buried residues).
Using criteria from Table 4.
Based solely on Hav criteria.
Based solely on Hmom criteria.