Skip to main content

Table 2 Performance of the proposed sequence-based methods on the 9 types of ligands over five-fold cross-validation and comparison with S-SITE

From: Protein ligand-specific binding residue predictions by an ensemble classifier

Ligand w a Method Accuracy (%) Sensitivity (%) Specificity (%) MCC
CU 15 TargetSeq 99.02 51.40 99.69 0.59
S-SITE 97.98 60.37 98.50 0.46
FE 9 TargetSeq 98.83 53.54 99.52 0.57
S-SITE 96.93 59.55 97.49 0.38
FE2 9 TargetSeq 99.20 51.36 99.81 0.63
S-SITE 98.28 42.14 99.00 0.37
ZN 11 TargetSeq 99.01 41.78 99.68 0.50
S-SITE 97.71 56.43 98.20 0.38
SO4 13 TargetSeq 97.79 10.07 99.66 0.19
S-SITE 96.98 14.4 98.73 0.15
PO4 7 TargetSeq 98.09 20.18 99.59 0.31
S-SITE 97.29 27.86 98.63 0.27
ATP 19 TargetSeq 97.14 36.81 99.31 0.48
S-SITE 96.73 48.09 98.48 0.49
FMN 17 TargetSeq 97.23 56.59 99.32 0.66
S-SITE 96.39 66.56 97.92 0.62
HEME 17 TargetSeq 92.62 61.27 95.26 0.53
S-SITE 93.63 58.24 96.61 0.55
  1. aThe optimal window length