Entering edit mode
5.5 years ago
afli
▴
190
Hi, in WGCNA manual I learn that the minimum sample size for WGCNA is 15. If I have 10 different treatment, each treatment has three replicates, do I have 30 samples or just 10 samples?
On the other hand, if I want to construct co-expression network of leaf tissue in drought condition, but the available RNA-seq samples of this are fewer than 15, would it be better to add RNA-seq data of leaf tissue in disease condition, salt stress condition and so on? or just keep and analyse the few samples? I wonder the gene regulation mode would maintain in different conditions, if not, add too many different RNA-seq samples may increase noise.
Thank you!
Aifu.
can you show the code?what you are using ,
You have mostly technical replicates ..
you mean three technical replicates is one sample, three biological replicates are three samples?
The value of 15 seems arbitrary to me - I am not sure that it is a hard threshold. If the developer of WGCNA went by the Aztec system of numbers, maybe he would have said that 20 is the minimum (Aztecs had 20-base system of numbers).
I would say that you have 30 samples.
All right, thank you!