|
As far as I know, Cuffmerge when we need to perform cuffdiff. In my study I want to classify the samples based on expression level obtained from cufflinks. I am wondering if Cuffmerge step is still needed if we are only interested in expression level ? (no differential expression analysis would be performed). |
|
Hello Setia You might get problems when not running cuffmerge with transcripts that only are found in a couple of the samples. Also with novel transcripts that might not have the start/stop positions in all samples. They might overlap to 99%, but if you are just 1 bp off in either position, a simple script might miss that they are the same transcript. Cuffmerge will fix these things for you. It will write a file containing the union of all transcripts in all samples, and it will do its best to figure out which transcripts are the same or different splice forms. It can also add annotations from a GTF file if you give it one, to annotate all transcripts it can. If you only are interested of previously annotated genes, i guess you could take the RPKM from each sample and skip running cuffmerge. Cheers Martin Hi Martin, Thanks for the answer, so in my case, we have 400 samples, I have to do cuffmerge for all of them together at once by using command: sampe1/transcripts.gtf sampe2/transcripts.gtf ... sampe400/transcripts.gtf as input for cuffmerge? or I have to do it for each sample?
(26 Oct '12, 05:59)
Setia
Yes, you will have to run it once with all the samples in that run. I have never done it with 400 samples, so i have no idea how long time or how much memory it will require :/ I would book a node for 10 days and see how much time and memory it uses.
(26 Oct '12, 09:44)
dahlo ♦♦
Hi Martin, I found out that the isoform.FPKM_tracking for each sample have different number of rows (number of isoforms). Is there any software if i want to combine all samples into a single matrix as we have in gene expression using microarray?
(31 Oct '12, 16:59)
Setia
Hmm, is that not what Cuffmerge is used for? Otherwise i think you will have to write your own script to do it.
(31 Oct '12, 17:01)
dahlo ♦♦
I check the output of cuffmerge, the result is like that.
(01 Nov '12, 16:15)
Setia
|
