Skip to main content

Table 2 Model performance on benchmark datasets

From: Integrating concept of pharmacophore with graph neural networks for chemical property prediction and interpretation

Category

Dataset

# Compoundsa

Task type

# Tasks

Metrics

AttentiveFP

MPNN

RG-MPNN (our model)

Physical chemistry

ESOL

1030

Regression

1

RMSE

0.650 ± 0.123

0.853 ± 0.057

0.605 ± 0.037

FreeSolv

566

Regression

1

RMSE

1.162 ± 0.180

1.255 ± 0.229

0.939 ± 0.067

Lipophilicity

4085

Regression

1

RMSE

0.627 ± 0.055

0.662 ± 0.019

0.579 ± 0.020

Bioactivity

MUV

91,470

Classification

17

ROC-AUC

0.772 ± 0.031

0.740 ± 0.012

0.819 ± 0.011

HIV

38,686

Classification

1

ROC-AUC

0.815 ± 0.022

0.803 ± 0.015

0.824 ± 0.019

BACE

1419

Classification

1

ROC-AUC

0.868 ± 0.024

0.846 ± 0.026

0.889 ± 0.018

Physiology or toxicity

BBBP

1928

Classification

1

ROC-AUC

0.888 ± 0.025

0.824 ± 0.038

0.879 ± 0.035

Tox21

7372

Classification

12

ROC-AUC

0.852 ± 0.025

0.836 ± 0.018

0.873 ± 0.008

ToxCast

8058

Classification

617

ROC-AUC

0.860 ± 0.012

0.848 ± 0.008

0.866 ± 0.009

SIDER

1270

Classification

27

ROC-AUC

0.827 ± 0.008

0.812 ± 0.012

0.825 ± 0.014

ClinTox

1437

Classification

2

ROC-AUC

0.940 ± 0.029

0.941 ± 0.026

0.965 ± 0.011

  1. Note that models with the best performance are in bold
  2. anumber of compounds used in this work