Entering edit mode
8.0 years ago
babak1146
▴
20
Can I make Coexpression Network with 8 sample(4 case and 4 control) expression data? I cannot find more data sample.
Can you? Yes. Will it be reliable/informative? Depends, we'd need more information.
A bit of clarification. While doing WGCNA or any other co-expression network, do we have to use the expression values for both the conditions of only a single condition?
I used Pearson Correlation Coefficients to create an unweighted network. Using 0.9 as threshold, I found that ~90% of interactions are retained. Any possibilities for this? Seemed a bit weird to me.
Use all of the expression values from all of the samples that you have. I'm not entirely sure what you mean by 90% of interactions being retained (compared to what?).
Ok sorry. I missed to mention. 90% of the maximum possible correlation between n genes given by (n*(n-1))/2.