Skip to main content

Table 1 Performance of benchmarking detection methods for apo proteins in terms of: (d) distance (FN) false negatives to PDBsum ground-truth cavity centers; (TP) true positives; (FP) false positives; (TN) true negatives; (S v ) sensitivity; (S c ) specificity; (a) accuracy; (r d ) ratio of detected ground-truth cavities; and (C u ) cumulative number of undetected ground-truth cavities

From: GPU-based detection of protein cavities using Gaussian surfaces

  GaussianFinder ConCavity POCASA SURFNET PASS GHECOM Fpocket 3V KVFinder
d[0.0,1.0] 7100 1512 2310 1737 4915 4562 5869 3751 4578
d]1.0,2.0] 188 107 176 1227 884 255 195 417 476
d]2.0,3.0] 227 127 226 520 117 380 207 291 305
d]3.0,4.0] 182 203 297 432 157 413 257 148 193
TP 7697 1949 3009 3014 6073 5610 6528 4607 5552
FP 1033 2103 863 902 325 581 878 1017 979
TN 2045 860 1697 1247 3086 2540 2827 1782 1889
FN 393 207 289 216 44 500 148 428 467
S v 0.951 0.904 0.912 0.933 0.993 0.918 0.978 0.915 0.922
S c 0.664 0.290 0.663 0.580 0.905 0.814 0.763 0.636 0.659
a 0.872 0.549 0.803 0.792 0.961 0.883 0.901 0.815 0.837
r d 0.944 0.239 0.369 0.369 0.745 0.688 0.801 0.565 0.681
C u 453 6201 5141 5136 2077 2540 1622 3543 2598