Skip to main content

Table 1 The mean MCC and BDT raw scores and Z-scores obtained by FunFOLD are compared with those obtained by each CASP8 function prediction group.

From: FunFOLD: an improved automated method for the prediction of ligand binding residues using 3D models of proteins

Group ID N Mean score for group Mean score for FunFOLD Increase in mean score Mean Z-score for group Mean Z-score for FunFOLD P-value
(raw score)
P-value
(Z-score)
1 - p-value
(raw score)
1 - p-value
(Z-score)
   MCC BDT MCC BDT MCC BDT MCC BDT MCC BDT MCC BDT MCC BDT MCC BDT MCC BDT
FN407 25 0.768 0.689 0.778 0.702 0.010 0.013 1.105 1.099 1.137 1.134 0.734 0.487 0.699 0.526 0.266 0.513 0.301 0.474
FN293 19 0.742 0.684 0.789 0.748 0.050 0.064 1.016 1.022 1.164 1.254 0.318 0.128 0.285 0.172 0.682 0.872 0.715 0.828
FN202 23 0.670 0.578 0.797 0.754 0.130 0.176 0.845 0.801 1.158 1.259 0.004 0.001 0.006 0.004 0.996 0.999 0.994 0.996
FN417 26 0.485 0.387 0.747 0.702 0.260 0.315 0.335 0.165 1.066 1.134 0.000 0.000 0.000 0.000 1.000 1.000 1.000 1.000
FN034 23 0.488 0.387 0.810 0.762 0.320 0.375 0.249 0.086 1.200 1.283 0.000 0.000 0.000 0.000 1.000 1.000 1.000 1.000
FN209 11 0.465 0.376 0.779 0.694 0.310 0.319 0.230 0.163 1.200 1.207 0.000 0.000 0.000 0.000 1.000 1.000 1.000 1.000
FN163 23 0.438 0.426 0.717 0.668 0.280 0.242 0.209 0.281 1.017 1.066 0.000 0.002 0.000 0.002 1.000 0.998 1.000 0.998
FN057 24 0.405 0.393 0.769 0.718 0.360 0.326 0.084 0.138 1.118 1.184 0.000 0.000 0.000 0.000 1.000 1.000 1.000 1.000
FN450 25 0.358 0.286 0.755 0.715 0.400 0.429 0.073 -0.031 1.072 1.162 0.000 0.000 0.000 0.001 1.000 1.000 1.000 0.999
FN325 25 0.377 0.336 0.778 0.730 0.400 0.394 -0.003 -0.015 1.131 1.203 0.000 0.000 0.000 0.000 1.000 1.000 1.000 1.000
FN437 17 0.209 0.183 0.768 0.716 0.560 0.534 -0.564 -0.603 1.151 1.243 0.000 0.000 0.000 0.000 1.000 1.000 1.000 1.000
FN108 24 0.136 0.133 0.769 0.730 0.630 0.597 -0.682 -0.663 1.121 1.203 0.000 0.000 0.000 0.000 1.000 1.000 1.000 1.000
FN483 21 0.076 0.094 0.723 0.669 0.650 0.575 -0.736 -0.719 1.040 1.063 0.000 0.000 0.000 0.000 1.000 1.000 1.000 1.000
FN242 26 0.126 0.133 0.747 0.702 0.620 0.569 -0.744 -0.704 1.066 1.134 0.000 0.000 0.000 0.000 1.000 1.000 1.000 1.000
FN105 26 0.011 0.044 0.747 0.702 0.740 0.658 -1.069 -0.983 1.066 1.134 0.000 0.000 0.000 0.000 1.000 1.000 1.000 1.000
FN086 24 0.024 0.068 0.781 0.728 0.760 0.660 -1.077 -0.933 1.151 1.222 0.000 0.000 0.000 0.000 1.000 1.000 1.000 1.000
  1. The analysis is based on common subsets of all CASP8 function prediction targets, with a minimum of 10 predictions in common. N, size of the common subset used in the comparison; MCC, Matthews Correlation Coefficient; BDT, the Binding Site Distance Test Score [36].P-value, the p-value for the paired Wilcoxon signed rank sum test using the raw scores; P-value (Z-score), the p-value for the paired Wilcoxon signed rank sum test using the Z-scores; 1 - p-value, 1 minus the p-value for the paired Wilcoxon signed rank sum test using the raw scores; 1-p-value (Z-score), 1 minus the p-value for the paired Wilcoxon signed rank sum test using the Z-scores. The tables are sorted by the Mean MCC Z-score for groups with the best CASP8 groups at the top. The highest mean scores, mean Z-scores, significant p-values and significant 1-p-values are indicated in bold. Server groups are underlined (FN293 and FN105 were extended deadline server groups; no publicly available server could be found for FN293 at the time of writing; group FN202 now have a publicly available server which is comparable to their CASP8 performance [15]).