search

fogellab / multiWGCNA

an R package for deep mining gene co-expression networks in multi-trait expression data
13 stars 2 forks source link

Permutation #13

Closed GettyScience closed 1 month ago

GettyScience commented 5 months ago

Hi,

Just a quick question since you know your code the best, what do you recommend as the minimum value for the number of permutations to run when using multiWGCNA code and still get reasonably accurate results? I have tended to stick around 1000, but this makes processing much longer. Do you think I could work with less or stick with what I have?

Thank you,

dariotommasini commented 5 months ago

Hi @GettyScience ,

For which analysis, the PreservationComparisons analysis or the PreservationPermutationTest? The PreservationComparisons analysis just runs WGCNA's native module preservation function which is probably fine with 100 permutations, in my hands. Results are largely identical when you run at least 100 permutations. For the PreservationPermutationTest, I would say as many as possible. With that said, it's incredibly slow so I would try 100 first and see how that goes. We did 2000 permutations in the paper but it ran for several days on the cluster.

GettyScience commented 5 months ago

Thank you.

dariotommasini commented 1 month ago

You're welcome!