cuffdiff

[4galaxy7]

hi,

so I'm slightly confused about the usage of cuffmerge/cuffdiff. as i see it, cuffmerge merges all of the .bam/.sam output files from tophat to produce a merged .gtf file which you then use as a reference during cuffdiff.

i have like 12 replicates for 3 conditions in my samples (giving 36 in total) and I want to look at differential expression between the 3 conditions. Do i need to somehow merge all 12 samples in each treatment into a file and then compare the 3 merged files, or do I run cuffmerge on all 48 individual files. i don't think the instructions in the readme are very clear!

for the record, I tried doing it like this, using all 48 files like this and it came out with some errors and wouldn't work.

Thanks!

/home/user/Downloads/cufflinks/cuffdiff -o /home/user/Desktop/cuffdiff_out /home/user/Desktop/sam/cuffmerge/beta/merged_beta.gtf 
/home/user/Desktop/sam/other_data/tophat_out/2B84A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/2B84B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/12Y30B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G48A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G48B/accepted_hits.bam,home/user/Desktop/sam/other_data/tophat_out/23Y79A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/23Y79B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W26A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W26B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10B70B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/35Y87A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/35Y87B/accepted_hits.bam
/home/user/Desktop/sam/other_data/tophat_out/2UNA/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/2UNB/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/12Y47A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/12Y47B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G87A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G87B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/23Y70A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/23Y70B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W62A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W62B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/1OR99A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10R99B/accepted_hits.bam
/home/user/Desktop/sam/other_data/tophat_out/2B82A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/2B82B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/12Y51A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/12Y51B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G17A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G17B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/23Y59A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/23Y59B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W35A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W35B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/35Y44A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/35Y44B/accepted_hits.bam

[dpryan]

You use all 36 GTF files with cuffmerge to produce a merged GTF file. That file and the 36 BAM files are then used with cuffdiff.

[GenoMax]

@4galaxy7: You are missing a -L to define groups you want to compare and a -u for the gtf file.

$ /home/user/Downloads/cufflinks/cuffdiff -o /home/user/Desktop/cuffdiff_out -L Cond_1,Cond_2,Cond_3 -u /home/user/Desktop/sam/cuffmerge/beta/merged_beta.gtf  remaining command list of bam files

[4galaxy7]

OK thanks! I'm a bit unclear about separating the different conditions.

So with the L option, do you need to define the different conditions with a name that it then uses to check against the .BAM files. So..

-L cond_1, cond_2, cond_3 and then cond_1_1.bam, cond_1_2.bam, cond_2_1.bam, cond_3_1.bam etc

I.e. how does cuffdiff know which of the different .bam files to allocate to a certain condition?

Thanks!

[GenoMax]

-L just designates labels for conditions you are comparing. Separate Sample replicates by a comma and then files for different conditions by a space. See some additional info here: https://www.biostars.org/p/159868/

Just including a truncated command below. You should provide all options you need.

-L Cond_1,Cond_2,Cond_3 Sample1_rep1.bam,Sample1_rep2.bam,Sample1_rep3.bam Sample2_rep1.bam,Sample2_rep2.bam,Sample2_rep3.bam Sample3_rep1.bam,Sample3_rep2.bam,Sample3_rep3.bam

[4galaxy7]

OK - tried that, but it comes up with an error saying that there is different number of conditions..

user@IT060209:~$ /home/user/Downloads/cufflinks/cuffdiff -o /home/user/Desktop/cuffdiff_out/ -L Cond_1,Cond_2,Cond_3  \
> /home/user/Desktop/sam/other_data/tophat_out/2B84A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/2B84B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/12Y30B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G48A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G48B/accepted_hits.bam,home/user/Desktop/sam/other_data/tophat_out/23Y79A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/23Y79B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W26A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W26B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10B70B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/35Y87A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/35Y87B/accepted_hits.bam \
> /home/user/Desktop/sam/other_data/tophat_out/2UNA/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/2UNB/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/12Y47A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/12Y47B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G87A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G87B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/23Y70A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/23Y70B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W62A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W62B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/1OR99A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10R99B/accepted_hits.bam \
> /home/user/Desktop/sam/other_data/tophat_out/2B82A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/2B82B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/12Y51A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/12Y51B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G17A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/10G17B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/23Y59A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/23Y59B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W35A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/36W35B/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/35Y44A/accepted_hits.bam,/home/user/Desktop/sam/other_data/tophat_out/35Y44B/accepted_hits.bam
You are using Cufflinks v2.2.1, which is the most recent release.
Error: number of labels must match number of conditions

Not sure what I've done wrong because all of the samples are in 3 comma delimited lists like it says they should be..

[dpryan]

"home/user/Desktop/sam/other_data/tophat_out/23Y79A/accepted_hits.bam" should be "/home/user/Desktop/sam/other_data/tophat_out/23Y79A/accepted_hits.bam", though I'm doubt that that's the actual cause of this.

[GenoMax]

@4galaxy7: Have you tried issuing the command on a single line (without the command line breaks \)? It will look long and ugly but so be it. Make the correction Devon caught above.

[blancha]

If this is the exact command that you copy-pasted, there is an extra space after Cond_3.
It can be helpful to use a text editor that displays white spaces, since these errors
are difficult to identify.
There may be other errors, but the command will definitely not work with extra white spaces.

[blancha]

There are other mistakes in your command, beside the extra white space.
I would suggest reading the Nature Protocols article describing the Tuxedo protocol.

Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks

http://www.nature.com/nprot/journal/v7/n3/full/nprot.2012.016.html