Skip to main content

Table 3 Prediction results of PLB-SAVE, SiteHound and MPK2 using the HOLO dataset.

From: Protein-ligand binding region prediction (PLB-SAVE) based on geometric features and CUDA acceleration

HOLO Dataset

PLB-SAVE (374 proteins)

SiteHound (374 proteins)

Sensitivity

0.623

0.538

Specificity

0.975

0.975

Accuracy

0.953

0.952

PPV

0.629

0.625

MCC

0.589

0.585

(a)

HOLO Dataset

PLB-SAVE (339 proteins)

MPK2 (339 proteins)

Sensitivity

0.642

0.806

Specificity

0.973

0.875

Accuracy

0.949

0.870

PPV

0.642

0.465

MCC

0.603

0.561

(b)

  1. Calculations were performed only on proteins whose binding regions had been successfully predicted by SiteHound and MPK2. Numbers in bold indicate which software gave the better performance for each parameter. (a) PLB-SAVE compared with SiteHound using 374 proteins. (b) PLB-SAVE compared with MPK2 with using 339 proteins.
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us