I am in a fix and want people out here to help me.
I have a control and treatment both in replicates so total 12 samples (including replicates). I have run the tophat pipeline (PS: I know i shouldn't have gone for it and should have chosen HISAT or any better pipeline). So after cuffdiff i got a file geneexp.diff that includes log2 fold change values. I set up the threshold cut off of +2 for upregulated and -2 for downregulated and p value less than 0.05
Now, how will i get a matrix to create heatmap ?? i have appx 2000 upregulated genes Do we need to take log2 fold change to create heatmap or fpkm values. my fpkm values range between 0.3 to 5000 is that wrong ?
my excel sheet is as follows test_id gene_id gene locus sample1 sample 2 value1 value 2 log2foldchange p value q value significance
post if you have any further queries.
I have some queries, why run this pipeline while you know there are better ones? Other question, why use p-values and ignore multiple testing? Last question, why not use CummeRbund if you use the (obsolete) tuxedo pipeline?
@b.nota 1) I was told to run this pipeline by my professor while running this one I found that there are better one's available but didnt want to leave this in middle. 2) please elaborate what do you mean by multiple testing or what other parameters to be considered ? 3) I can use CummeRbund but my query is how to get a matrix from my data for heatmap.
Thanks for your answer. Hope to learn more !!
What is your background? If you don't know what multiple testing is, please learn more about omics data analysis first (or ask prof.). There are also many tutorials about how to make heatmaps, which have you tried? Show as what you have tried already.