Skip to main content

Table 2 The performance of our model on general generative model evaluation metrics (GEM) among five distinct targets: CDK2, EGFR, JAK1, LRRK2, and PIM1

From: ScaffoldGVAE: scaffold generation and hopping of drug molecules via a variational autoencoder based on multi-view graph neural networks

Protein

Validity \(\uparrow\)

Uniqueness1K \(\uparrow\)

Uniqueness5K \(\uparrow\)

Filter \(\uparrow\)

Scaffold uniqueness \(\uparrow\)

Scaffold novelty \(\uparrow\)

Novelty \(\uparrow\)

CDK2

0.9047

0.5989

0.4423

0.8186

0.4880

0.4853

1.0000

EGFR

0.9017

0.6279

0.4762

0.5976

0.5220

0.4989

1.0000

JAK1

0.8967

0.5987

0.4434

0.9059

0.4937

0.4921

1.0000

LRRK2

0.9022

0.6021

0.4457

0.8957

0.4932

0.4748

1.0000

PIM1

0.8949

0.5982

0.454

0.8208

0.5028

0.4898

1.0000

Back to article page

Journal of Cheminformatics

ISSN: 1758-2946

Contact us