Skip to main content

Advertisement

Table 3 Evaluation of the Real-SPINE and NetSurfP method on subsets of residues from the CB511 dataset predicted with high reliability.

From: A generic method for assignment of reliability scores applied to solvent accessibility predictions

   Real-SPINE NetSurfP
%Top N RSA ASA P-RSA M-RSA RSA ASA P-RSA M-RSA
10 8372 0.73 0.74 0.16 0.18 0.77 0.79 0.35 0.35
20 16745 0.73 0.74 0.16 0.18 0.79 0.79 0.31 0.31
25 20931 0.73 0.74 0.17 0.19 0.79 0.79 0.30 0.30
50 41863 0.72 0.74 0.18 0.20 0.77 0.77 0.28 0.28
75 62795 0.71 0.73 0.22 0.24 0.74 0.75 0.28 0.28
80 66981 0.71 0.73 0.23 0.25 0.73 0.74 0.28 0.28
90 75354 0.70 0.73 0.25 0.27 0.72 0.73 0.28 0.28
100 83727 0.70 0.73 0.27 0.29 0.70 0.72 0.29 0.29
  1. %Top and N give the percentage and number of residues selected. RSA and ASA give the Pearson's correlation between predicted and target for relative and absolute surface areas, respectively. P-RSA, and M-RSA give the mean predicted and mean measured RSA values, respectively, on the selected subset of residues.