Skip to main content

Table 2 GraphAlignment and Græmlin performance on empirical bio-molecular networks

From: GraphAlignment: Bayesian pairwise alignment of biological networks

Comparison

Escherichia coli vs. Caulobacter crescentus

Escherichia coli vs. Campylobacter jejuni

Algorithm

GraphAlignment

Græmlin

Blast BBH

GraphAlignment

Græmlin

Blast BBH

NA

445

467

462

354

363

357

NC

319

309 (333)

333

247

241 (253)

253

NO

331

331

331

255

255

255

NC / NA [%]

71.7

66.2 (71.3)

72.1

69.8

66.3 (69.7)

70.9

NC / NO [%]

96.4

93.4 (101)

101

96.9

94.5 (99.2)

99.2

Edge / vertex score

2505 / 2774

-

-

2592 / 2253

-

-

Comparison

Homo sapiens vs. Mus musculus

Homo sapiens vs. Saccharomyces cerevisiae

Algorithm

GraphAlignment

Græmlin

Blast BBH

GraphAlignment

Græmlin

Blast BBH

NA

7919

7907

7862

2369

1213

988

NC

5743

6327

6375

581

869 (882)

808

NO

6402

6402

6402

965

965

965

NC / NA [%]

72.5

80.0 (80.0)

81.1

24.5

71.6 (72.7)

81.8

NC / NO [%]

89.7

98.8 (98.8)

99.6

60.2

90.1 (91.4)

83.7

Edge / vertex score

2034 / 64661

-

-

20025 / 3963

-

-

  1. Protein-protein interaction networks. For Græmlin, the values are calculated using the strict rule. Values obtained following the relaxed rule are given in parentheses. For GraphAlignment, the relative contributions of the edge and node score are also given. Results obtained using BLAST bidirectional best hit are provided for comparison.