I have 8 RNA libraries (4 reps per each of two conditions - tissue specific).
I have approx 500 genes upregulated and 100 genes downregulated in the tissue of interest.
I have a matrix of deseq2 normalised counts for all of the genes which I've pulled the differently expressed gene records from.
Rows are gene names, columns represent each library with respective normalised log2 values.
I want to do clustering and to identify which genes are potentially co-expressed in a similar manner. Can anyone guide me through the process. Which software and steps I need to take? I use R and I've figured that with a small sample size wgnca is not valid, however Pearson correlation may be best? Is there a way I can plot a hierarchical clustered tree? I did use hclust but it seems I end up with approx 400 clusters, a lot with very weak evidence. Is there anyway to only create a dendrogram with the highest confidence co-expressed links between subsets of differentially expressed genes?
The plan after is to identify potential upstream promoter regions and motifs.
Many thanks!
