Table 4. Comparison of active module detection methods with CTD.
Method | Network Structure | Biological Data | Network Propagation Used? | Optimization/Subgraph Search Performed? | Module Assigned Significance? | Permutation testing? |
---|---|---|---|---|---|---|
CTD | Gaussian graphical models | Functional -omics | YES | YES | YES | NO |
CePa | Pathway Knowledgebases | Gene expression | NO | NO | YES | YES |
DEGraph | Pathway Knowledgebases | Gene expression | NO | YES | YES | NO |
HotNet/HotNet2 | Interaction Knowledgebases | Somatic mutation | YES | YES | YES | YES |
MATISSE | Interaction Knowledgebases | Gene expression | NO | YES | NO | NO |
Multi-Dendrix | Interaction Knowledgebases | Somatic mutation | NO | YES | YES | YES |
NetWalk | Interaction Knowledgebases | Gene expression, RNAi screen | YES | NO | NO* | N/A |
PARADIGM | Pathway Knowledgebases | Copy number, Gene expression, Proteomics | YES | NO | YES | YES |
PRS | Pathway Knowledgebases | Gene expression | NO | NO | YES | YES |
SPIA | Pathway Knowledgebases | Gene expression | NO | NO | YES | YES |
* Edge probabilities are computed as Edge Flux scores. The authors state that statistical procedures on this distribution can be applied, but do not offer module significance inherently.