Skip to main content

Table 1 Overall performance of MOST in sevenfold cross-validation

From: MOST: most-similar ligand based approach to target prediction

Active data defined by

Performance

pKi ≥ 5

Accuracy

 

Naïve Bayes

Logistic Regression

Random Forests

Morgan

FP2

Morgan

FP2

Morgan

FP2

Explicit Ki

0.927 ± 0.003

0.929 ± 0.003

0.948 ± 0.003

0.949 ± 0.002

0.948 ± 0.001

0.946 ± 0.004

Implicit Ki

0.939 ± 0.002

0.937 ± 0.003

0.948 ± 0.003

0.949 ± 0.003

0.950 ± 0.002

0.949 ± 0.003

MCC

 

Naïve Bayes

Logistic Regression

Random Forests

Morgan

FP2

Morgan

FP2

Morgan

FP2

Explicit Ki

0.352 ± 0.013

0.345 ± 0.017

0.495 ± 0.010

0.484 ± 0.025

0.530 ± 0.017

0.504 ± 0.032

Implicit Ki

0.554 ± 0.008

0.543 ± 0.022

0.558 ± 0.013

0.585 ± 0.023

0.585 ± 0.014

0.560 ± 0.023

pKi ≥ 6

Accuracy

 

Naïve Bayes

Logistic Regression

Random Forests

Morgan

FP2

Morgan

FP2

Morgan

FP2

Explicit Ki

0.848 ± 0.004

0.842 ± 0.002

0.866 ± 0.002

0.862 ± 0.002

0.861 ± 0.004

0.853 ± 0.004

Implicit Ki

0.860 ± 0.004

0.855 ± 0.004

0.867 ± 0.003

0.863 ± 0.003

0.867 ± 0.004

0.862 ± 0.004

MCC

 

Naïve Bayes

Logistic Regression

Random Forests

Morgan

FP2

Morgan

FP2

Morgan

FP2

Explicit Ki

0.561 ± 0.014

0.540 ± 0.009

0.617 ± 0.009

0.602 ± 0.007

0.609 ± 0.013

0.581 ± 0.013

Implicit Ki

0.624 ± 0.011

0.610 ± 0.612

0.632 ± 0.012

0.618 ± 0.011

0.632 ± 0.013

0.618 ± 0.012