Skip to main content


Table 2 Summary of performance of the DFMAC function on the training and test decoy sets.

From: A knowledge-based structure-discriminating function that requires only main-chain atom coordinates

target set Nall Nn Nnn C α RMSD Z-score C.C. F.E.(%) RB1 logPB1 RB10 logPB10 C.C.decoy F.E.decoy(%)
training set 154 115 135 0.764 2.552 0.539 38.9 171.3 -0.78 36.1 -1.38 0.499 27.6
test set 77 59 68 1.174 2.630 0.559 38.7 164.0 -0.75 15.8 -1.41 0.518 25.1
  1. Summarized values are shown by the respective decoy sets. The numbers of total target proteins evaluated (Nall), and correct identification of the native (Nn) or near-native (C α RMSD < 2 Å) (Nnn) structures are shown. The C α RMSD, Z-score, C.C., F.E., RB1, logPB1, RB10, logPB10, C.C.decoy, F.E.decoy are the average of the respective scores of the target proteins evaluated. The definition of each index is described in Methods.