Entering edit mode
3.2 years ago
gerastuff
•
0
Hello everyone!
I was just wondering if when doing a WGCNA of many samples with lots of different experimental groups (so for instance this is about 120 different conditions, like serum/serum free, LPS, tx1, tx2 etc.) and having quadruplicates of the data, can you average each gene from its 4 replicates and construct the network from that?
I've been thinking about pro's and cons.. and can't think of a strong reason of why it shouldn't be feasible...
Averaging means reduction of power, I do not see any reason to do that. Do you have concerns towards RAM usage or runtime?