Skip to main content

Table 1 Dataset details (a) Golden-standard dataset (b) ChEMBL-based dataset

From: Novel drug-target interactions via link prediction and network embedding

(a) Golden_standarad_dataset

# Drug

# Target

# Negative/unknown DTI

# Positive interaction

Class ratio

Enzyme (E)

445

664

292,554

2926

0.01

Ion channel (IC)

210

204

41,364

1476

0.04

G-protein-coupled receptors (GPCR)

223

95

20,550

635

0.03

Nuclear receptor (NR)

54

26

1314

90

0.07

Total

791

989

777,172

5127

0.01

(b) ChEMBL_dataset

# Drug

# Target

# Real negative DTI†

# Positive interaction*, †

Class ratio

CheMBL

548

556

2057

1721

0.84

 

# Weak DTI**

# Unknown DTI‡

   
 

532

300,378

   
  1. *pChEMBL value ≥ 5.5, **pChEMBL value > 0, †Development-dataset, ‡Experimental-dataset