Skip to main content
. 2019 Oct 14;14(10):e0223596. doi: 10.1371/journal.pone.0223596

Table 1. Classification (TP, FP, FN) and performance (p, r, F) results with repetitions.

Four cavity detection methods are benchmarked against the ground-truth: Fpocket, GaussianFinder, GHECOM, and KVfinder.

Method APOs HOLOs APOs + HOLOs
TP FP FN p r F TP FP FN p r F p r F
Fpocket 16722 1480 3644 0.92 0.82 0.86 50027 5722 8699 0.90 0.85 0.87 0.90 0.84 0.87
GaussianFinder 16441 2443 1251 0.87 0.93 0.90 45011 7939 2964 0.85 0.94 0.89 0.86 0.94 0.89
GHECOM 11875 5946 5307 0.67 0.69 0.68 33962 19566 13917 0.63 0.71 0.67 0.64 0.70 0.67
KVFinder 8074 1378 7590 0.85 0.52 0.64 23385 3643 19885 0.87 0.54 0.67 0.86 0.53 0.66

Abbreviations:

APOs: apo proteins; HOLOs: holo proteins.

TP: true positives; FP: false positives; FN: false negatives.

p: precision; r: recall; F: F-score.