Skip to main content

Table 1 Bioactivity datasets assembled from ChEMBL repository and utilized in this study

From: Deep-learning: investigating deep neural networks hyper-parameters and comparison of performance to shallow methods for modeling bioactivity data

Activity class

CHEMBL target id

Number of active inhibitors

Number of decoys

Carbonic anhydrase II,

Class: enzyme, lyase

CHEMBL205

1631

16,310

Cyclin-dependent kinase 2,

Class: protein kinase

CHEMBL301

705

7050

HERG,

Class: Voltage-gated ion channel

CHEMBL240

700

7000

Dopamine D4 receptor,

Class: membrane receptor, GPCR

CHEMBL219

506

5060

Coagulation factor X,

Class: enzyme, serine protease

CHEMBL244

1144

11,440

Cannabinoid CB1 receptor,

Class: membrane receptor, GPCR

CHEMBL218

1911

19,013

Cytochrome P450 19A1,

Class: enzyme, cytochrome P450

CHEMBL1978

621

6210

  1. The ratio of decoys/active per activity class was set to 10:1