Identifying disease associated CCANs

[Daniel Gingerich]

In WGCNA, the eigengenes of each module are defined as the first principal component of the module.  Correlation of the module eigengene with disease status can be measured to identify disease associated modules.  

Can the same type of analysis be done with Cicero CCANs?  Specifically, I would run Cicero on my data to obtain the CCANs.  Then, I would do SVD on my data, separately for each CCAN.  Then, measure the correlation of the first principal component projections with disease status using logistic regression and likelihood ratio to control for confounders.