Hi! I am looking for some hints to analyse some data. I have not much experience in this area and it is getting to complicated for me...
I have 6 conditions, each with 3 biological replicas (let's say, 18 samples). (plant samples, 50000 genes) I am using Deseq2 to get differential expressed genes when I compared between 2 samples... But I want to compared the 6 samples (i.e. some heat maps, PCA, clustering, etc)....and this is my problem... - I can get RPKM values for each replica but I can't compare raw RPKM data for each sample because of the deep of sequencing - In some papers, they mention a "standard score (z-score) of Deseq2 normalized read numbers", if I am right, deseq2 count, and then normalized the count according to sequencing of each replica in order to be able to compare between samples. Thus...I guess that to compared all the replicas, this z-score could be a good numerical value for the comparison...but, how could I get it the normalized read numbers or the Z-score?
Thanks
Gonzalo
Thanks! I will take a look...
Hi Michael, I was able to get the values. I loaded the full matrix (6 conditions with 3 replicas of each) and I got the normalizad read count finally only with "counts(dds,normalized=TRUE)".
But then I run DESeq in two ways...the full data set only (6 conditions with 3 replicas of each) and compared two conditions as: res <- results(dds, contrast=c("condition","10dpi2011","4w2011")) but then I extract only the two conditions and run deseq only in these conditions dds2011$condition <- factor(dds2011$condition, levels = c("10dpi2011","4w2011")) ...and I saw that size factors for each replica are different and log2FoldChange are also slightly different (r-square =0,98 for 48.000 genes)...so the question is...how I should proceed for deferentially gene evaluation, with the full table or extracting the conditions for each comparison?
Thanks!
This is one of the FAQ in the vignette.