Skip to main content

Table 2 Statistics of predicted complexes by five algorithms with the best parameters optimized on CYC08 and YHTP08 sets and the DIP PPI network as input.

From: Discover Protein Complexes in Protein-Protein Interaction Networks Using Parametric Local Modularity

Algorithm

Gold Standard Sets

(optimized parameters)

CYC08 (236/6.7)

YHTP08 (207/8.2)

COACH

271/113/7.3

271/95/7.3

(affinity threshold)

(0.1)

(0.1)

MCODE

57/25/12.9

57/18/12.9

(VWP)

(0.2)

(0.2)

MCL

830/123/5.9

830/115/5.9

(inflation)

(1.75)

(1.75)

DME

487/44/25.1

503/40/24.7

(density threshold)

(0.97)

(0.96)

miPALM

238/100/7.0

277/88/7.0

(α, δ)

(0.364, 2.40)

(0.374, 2.33)

  1. Total number and average size of gold standard sets are shown in parenthesis. The series of three numbers for each set of predictions are total number of predictions, number of predicted complexes that overlap with known complexes, and average size of predicted complexes. Optimized parameters for each algorithm are shown in parenthesis following prediction numbers.