Question: Cuffmerge:Which samples to use for creating a merged gtf file?
0
gravatar for EpiExplorer
3.8 years ago by
EpiExplorer70
New Zealand
EpiExplorer70 wrote:

Hi ,

I have questions regarding merging of files in cuffmerge and usage of groups in cuffdiff. I have 5 samples in condition1 and 4 samples in condition2 and 3 samples are normal samples. I would like carry out differential expression analysis using cuffdiff. So I have following questions:

i) Can cuffdiff carry out differential analysis for three conditions- Condition1, Condition2 and Normal, together? I am asking this to know what is best method to input these conditions in cuffdiff- pairwise or all conditions together?

ii) If I choose to go with pairwise comparisons - Condition1 vs Condtion2 or Normal vs Condition1 or Normal vs Condition2, is it correct to use transcript.gtf files from all 12 samples to create a merged gtf file which can be used further in either of the mentioned comparisons? Thanks for your help.

ADD COMMENTlink modified 3.8 years ago by DG7.1k • written 3.8 years ago by EpiExplorer70

if you have read the manual of cuffdiff, you would have know, that cuffdiff can take as many replicates as needed. from the manual:

cuffdiff [options] <transcripts.gtf> \
<sample1_replicate1.sam[,…,sample1_replicateM.sam]> \
<sample2_replicate1.sam[,…,sample2_replicateM.sam]> … \
[sampleN.sam_replicate1.sam[,…,sample2_replicateM.sam]]

where sampleN in your case =3. cuffdiff will than automatically calculate the DE between all possible constellations.

In this case I would create a gtf file with all samples in one go.

  1. I think it would be best to create a separate gtf file for the pair-wise comparisons. Otherwise you can create false entries in the gtf file, which belongs to different conditions.
ADD REPLYlink modified 3.8 years ago • written 3.8 years ago by Assa Yeroslaviz1.3k
1
gravatar for DG
3.8 years ago by
DG7.1k
DG7.1k wrote:

You should always merge all of the transcript GTF files together with cuffmerge regardless of what comparisons you will do. This essentially creates the master assembled transcriptome that contains all of the possible transcripts. This essentially just acts as a new transcriptome reference for your comparisons.

@Frymor answered your first question, as that information is in the cuffdiff manual

ADD COMMENTlink written 3.8 years ago by DG7.1k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1803 users visited in the last hour