WGCNA Different Module Results

Question

[ninabhatia3]

I'm using WGCNA to find gene modules- I've been using both the code from http://pklab.med.harvard.edu/scw2014/WGCNA.html and from here https://horvath.genetics.ucla.edu/html/CoexpressionNetwork/Rpackages/WGCNA/Tutorials/ and I am getting different results in the number of modules/module sizes. Can anyone explain these discrepancies.

Answer 1

[Sam]

One of the step of WGCNA involves hierachical clustering, which is partially random. Unless you've set the random seed before each run, you'd expect to see slight difference in the number of modules, module size and module membership, especially when your sample size is relatively small

[ninabhatia3]

Thank you so much!

Do you know why the Hovarth code uses this command : dynamicMods = cutreeDynamic(dendro = geneTree, distM = dissTOM, method="hybrid", deepSplit = 2, pamRespectsDendro = FALSE, minClusterSize = minModuleSize); when creating modules stepwise as opposed to this command used in the Harvard version: dynamicMods = cutreeDynamic(dendro = geneTree, method="tree", minClusterSize = minModuleSize);

I ask because using the different commands gives me a vastly different amounts of modules.

[Sam]

I'm not sure. But from the glance of it, those two codes are using two different method (tree vs hybrid). So I am not surprise that they generate a different result