Table 7.
Algorithm | C~ | C^ | CN | Cr | CF | CR |
---|---|---|---|---|---|---|
ESU (Counts) | 1.0(2,509) | 1.0(5,152) | 1.0(17,457) | 1.0(434) | 1.0(8,095) | 1.0(15,953) |
RAND-ESU | .30 | .32 | .34 | .36 | .34 | .34 |
MFINDER | .78 | .54 | .31 | .38 | .16 | .13 |
EDGEGO-BNM | .97 | .97 | .98 | 1.0 | .99 | .97 |
EDGEBETWEENNESS-BNM | .67 | .64 | .32 | .57 | .22 | .16 |
NMFGO-BNM | .87 | .88 | .78 | .89 | .70 | .73 |
NMF-BNM | .69 | .39 | .23 | .22 | .12 | .90 |
VOLTAGE-BNM | .53 | .38 | .39 | .39 | .32 | .31 |
Except ESU, all algorithms only search 30% of subgraphs in the original network. However, EDGEGO-BNM recovers over 90% of motifs included in functional module. We note that the non-motif types of Cr, CF and CR have a number of instances for this functional match, indicating structural uniqueness is insufficient to discover its biological significance.