Table 1 Statistical parameters of the Py-CoMFA (3D-QSAR) models built using steric field (STE) and 2 Å grid spacing, including Y-randomization (Y-r) test, according to the training and test sets splitting method (random or Kennard-Stone) and the probe atoms and charges (Csp3, + 1; Osp3, − 1; and H, + 1).

From: Py-CoMFA, docking, and molecular dynamics simulations of Leishmania (L.) amazonensis arginase inhibitors

Models

Random

Kennard-Stone

Kennard-Stone

Kennard-Stone

Probe atom

Csp3

Csp3

Osp3

H

Charge

+ 1

+ 1

− 1

+ 1

r2

0.995

0.999

0.992

0.953

SDEC

0.046

0.022

0.061

0.147

q2

0.469

0.573

0.513

0.499

SDEP

0.487

0.444

0.474

0.481

F-test

155.7

197.8

172.6

79.24

r

0.910

0.923

0.918

0.844

r2 (Y-r)

0.992

0.986

0.972

0.890

SDEC (Y-r)

0.056

0.082

0.114

0.225

q2 (Y-r)

0.121

0.103

 − 0.198

 − 0.201

SDEP (Y-r)

0.627

0.643

0.743

0.744

  1. All models presented a P value < 0.0001. r2 = coefficient of determination; SDEC = standard deviation error in calculation; q2 = r2 from leave-one-out (LOO) cross-validation; SDEP = standard deviation error in prediction; r = Pearson correlation coefficient.
Back to article page