Skip to main content

Table 1 Overview of the PPI datasets used in the experiments

From: Normalized L3-based link prediction in protein–protein interaction networks

Dataset \Number of

Size (MB)

Nodes

PPIs

Cand. PPIs

SIPs

Synthetic PPI dataset

Wang et al.[37]

0.547

8,272

52,922

29,816,060.1

480

Saccharomyces cerevisiae (Yeast)

BioGRID[38]

316

7085

113,116

20,045,849.4

1739

STRING[39]

85.5*

4673

94,529

9,212,026.6

0

MINT[40]

38.3

4049

16,927

5,980,266.7

0

Homo sapiens (Human)

BioGRID[38]

166

24,760

452,684

220,833,040.0

2900

STRING[39]

717*

15,668

308,614

88,982,499.1

12

MINT[40]

55.0

7,534

22,324

15,493,875.9

0

HuRI[41]

161

8109

51,127

21,899,033.2

0

HI-II-14[42]

0.185**

4298

13,868

5,165,263.5

518

Hein et al.[43]

0.368**

5457

28,780

10,939,287.2

1127

Lit-BM-13[42]

0.135**

5545

11,045

8,147,585.2

890

Lit-NB-13[42]

0.064**

3391

4906

2,738,996.0

518

  1. “Cand. PPIs” refers to the mean number of candidate PPIs for its ten sampled datasets
  2. “SIPs” refers to the number of self-interacting proteins
  3. *Denotes the combined file size of multiple essential metadata files
  4. **Indicates that the (pre-processed) dataset was downloaded from the repository of the study by Kovács et al.