Table 2.
Comparison of CASCADE to competing clustering methods for 2 biological network data sets (BIOGRID Yeast PPI network, DIP Yeast PPI network).
| Dataset | Method | Cluster | Size | Discard | MIPS (-logp) | GO (-logp) | ||||
| Number | (%) | Function | Location | Complex | mf | cc | bp | |||
| BIOGRID Yeast PPI network | CASCADE | 225 | 19.6 | 18.3 | 3.26 | 2.55 | 5.13 | 4.67 | 4.24 | 3.53 |
| STM | 248 | 18.1 | 16.2 | 2.88 | 2.37 | 4.64 | 4.17 | 3.98 | 3.53 | |
| Maximal clique | 587 | 3.6 | 80.8 | 2.71 | 2.21 | 4.53 | 3.55 | 3.47 | 2.99 | |
| Quasi clique | 431 | 7.4 | 40.9 | 2.97 | 2.03 | 4.89 | 4.16 | 3.88 | 3.02 | |
| Samanta | 289 | 6.7 | 64.8 | 2.63 | 1.61 | 4.59 | 3.48 | 3.29 | 3.01 | |
| MCL | 617 | 6.2 | 29.2 | 2.58 | 1.22 | 3.87 | 4.02 | 3.77 | 2.83 | |
| Chen | 577 | 8.4 | 10.1 | 2.61 | 2.08 | 4.13 | 4.36 | 3.84 | 3.05 | |
| Rives | 217 | 21.5 | 13.5 | 3.04 | 2.34 | 4.22 | 4.14 | 3.97 | 3.03 | |
| SPC | 85 | 54.9 | 13.4 | 1.33 | 0.87 | 2.65 | 2.11 | 2.51 | 2.29 | |
| DIP Yeast PPI network | CASCADE | 50 | 48.1 | 7.3 | 14.1 | 7.84 | 15.8 | 12.1 | 12.8 | 9.09 |
| STM | 60 | 40.1 | 7.8 | 13.0 | 7.23 | 14.2 | 11.8 | 11.9 | 8.04 | |
| Maximal clique | 120 | 5.7 | 98.3 | 10.2 | 7.67 | 10.0 | 8.46 | 10.0 | 6.57 | |
| Quasi clique | 103 | 11.2 | 80.8 | 11.0 | 6.29 | 12.0 | 10.7 | 11.1 | 7.69 | |
| Samanta | 64 | 7.9 | 79.9 | 8.76 | 4.74 | 10.7 | 9.82 | 10.8 | 8.01 | |
| Minimum cut | 114 | 13.5 | 35.0 | 7.97 | 4.58 | 8.56 | 8.19 | 7.87 | 6.21 | |
| Betweenness cut | 180 | 10.3 | 21.0 | 7.89 | 4.06 | 8.59 | 7.02 | 6.98 | 4.88 | |
| MCL | 163 | 9.8 | 36.7 | 8.08 | 3.84 | 9.53 | 7.81 | 8.11 | 6.26 | |
| Chen | 141 | 16.3 | 1.7 | 9.12 | 4.91 | 9.87 | 8.28 | 8.09 | 6.01 | |
| Rives | 42 | 55.3 | 7.8 | 10.1 | 6.88 | 9.52 | 9.61 | 9.59 | 7.42 | |
| SPC | 5 | 47.2 | 6.4 | 5.27 | 2.39 | 5.49 | 6.23 | 5.91 | 5.18 | |
The Number column indicates the number of clusters identified by each method, the Size column indicates the average number of molecular components in each cluster; the Discard(%) indicates the percentage of molecular components not assigned to any cluster. The average -log p values of all detected clusters for MIPS categories (biological function, cellular location, complex) and Gene Ontology (molecular functions (mf), biological process (bp), cellular component (cc)) are shown. Comparisons are performed on the clusters with 5 or more molecular components. The results for Minimum cut and Betweenness cut for the BIOGRID dataset are not shown due to limitation of the available implementation.