Skip to main content

Table 2 Success of triangle network motifs and themes in predicting known MIPS complexes.

From: Triangle network motifs predict complexes by complementing high-error interactomes with structural information

Complementary datatype

Gavin06MATRIX

Gavin06SPOKE

Krogan06

None a

936/166241 = 0.6%

516/10791 = 4.8%

914/33124 = 2.8%

SDDI b

254/2832 = 9.0%

143/521 = 27.4%

254/1182 = 21.5%

Literature co-occurrence c

710/5592 = 12.7%

416/1340 = 31%

502/1876 = 26.8%

Domain co-occurrence d

2004/21876 = 9.2%

892/4268 = 20.9%

1250/4776 = 26.2%

Union of all above

2477/26468 = 9.4%

1446/6129 = 23.6%

1647/6489 = 25.4%

  1. a Second-level indirect relations only
  2. b Structural domain-domain interaction
  3. c PubMed literature co-occurrence of protein mentions
  4. d Pfam domain co-occurrence in IntAct PPIs
  5. Fractions denote True Positive PPIs/All triangle PPIs for triangles or second-level neighbors where all three proteins occur in MIPS complexes. Triangle success in MIPS complex prediction is shown as the triangle edges that overlap with complexes. We consistently notice a lower success rate for Gavin06MATRIX than Gavin06SPOKE, which is explained by the higher number of errors in Gavin06MATRIX.
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us