cuffdiff not performing splicing tests and promoter tests

[vaibhav_jain]

Hi all,

I am using cufflinks package to look for differentially spliced exons. The pipeline I followed is tophat -> cufflinks -> cuffmerge/cuffcompare(tried with both) -> cuffdiff.
Result on command line is like this :-

Performed 23758 isoform-level transcription difference tests
Performed 20313 tss-level transcription difference tests
Performed 18306 gene-level transcription difference tests
Performed 17903 CDS-level transcription difference tests
Performed 0 splicing tests
Performed 0 promoter preference tests
Performing 0 relative CDS output tests
Writing isoform-level FPKM tracking
Writing TSS group-level FPKM tracking
Writing gene-level FPKM tracking
Writing CDS-level FPKM tracking
Writing isoform-level count tracking
Writing TSS group-level count tracking
Writing gene-level count tracking
Writing CDS-level count tracking
Writing isoform-level read group tracking
Writing TSS group-level read group tracking
Writing gene-level read group tracking
Writing CDS-level read group tracking
Writing read group info
Writing run info

Command I used to run cuffdiff is :-
./cuffdiff --no-update-check -o day0 -p 2 -N -b /home/Vaibhav/Alt_transcriptome_data/ws230/ws230.fa -u merged_asm/merged.gtf /home/Vaibhav/Alt_transcriptome_data/new/again_X_day0/accepted_hits.bam /home/Vaibhav/Alt_transcriptome_data/new/again_Y_day0/accepted_hits.bam

I have no replicates of the sample. Now I am getting some differential results for gene, cds and isoform but getting nothing for splicing and promoter something I am interested in. I also tried lowering c option till 2, but got the same result.Can someone please suggest me what should I do so that cuffdiff performs splicing tests.
Really in need of the differential spliced exons for my data.

Thanks a lot

[krespim]

hi vaibhav_jain,

did you solve the problem? I having similar issues and can't for the life of me figure out what is going on.

[leifive]

Hi, vaibhav_jain and krespim:
Have you solved the problem? This has puzzled me almost one week. Why cuffdiff performed 0 splicing test?

[krespim]

Hi, vaibhav_jain and krespim:
Have you solved the problem? This has puzzled me almost one week. Why cuffdiff performed 0 splicing test?

Not really a solution, but when I load the results in cuffdiff I do get DE isoforms. Do try that. Maybe it is just a bug in the reporting.

[leifive]

Not really a solution, but when I load the results in cuffdiff I do get DE isoforms. Do try that. Maybe it is just a bug in the reporting.

Thanks for your reply, krespim. But I don't think this is just a bug in the reporting, because there is no "OK" status in my splicing.diff file from Cuffdiff, but NOTEST or LOWDATA instead. And I have tried both combined.gtf and merged.gtf as input of cuffdiff, results are almost the same.

[krespim]

Thanks for your reply, krespim. But I don't think this is just a bug in the reporting, because there is no "OK" status in my splicing.diff file from Cuffdiff, but NOTEST or LOWDATA instead. And I have tried both combined.gtf and merged.gtf as input of cuffdiff, results are almost the same.

Well, I can't help much further but I can tell you what I end up doing. As I am working with species that have relative well annotated genomes/transcriptomes (Homo sapiens and Mus musculus), I forego the initial cufflinks step and go straight to cufdiff after mapping using the gtf provided in iGenomes as a reference. Here are my settings:

# cuffdiff v2.0.2

$cuffdiff -L "GFP,KD" -o $cuff_output -p 8 -u $gtfFile $gfp_a,$gfp_b $srsf1_a,$srsf1_b

And that seems to work.

[leifive]

Well, I can't help much further but I can tell you what I end up doing. As I am working with species that have relative well annotated genomes/transcriptomes (Homo sapiens and Mus musculus), I forego the initial cufflinks step and go straight to cufdiff after mapping using the gtf provided in iGenomes as a reference. Here are my settings:

# cuffdiff v2.0.2

$cuffdiff -L "GFP,KD" -o $cuff_output -p 8 -u $gtfFile $gfp_a,$gfp_b $srsf1_a,$srsf1_b

And that seems to work.

Thanks krespim. I have solved the problem by going back to Cufflinks 1.3.0, and I'm not sure whether there is a bug in the lastest version of Cufflinks or something incompatible in my reference file. However, it's done now.

[krespim]

Update

Thanks krespim. I have solved the problem by going back to Cufflinks 1.3.0, and I'm not sure whether there is a bug in the lastest version of Cufflinks or something incompatible in my reference file. However, it's done now.

I went back to my notes - because I had to run cuffdiff again - and the problem seems to occur when the annotation gft is not a cufflinks output or from iGenomes (not 100% sure about iGenomes though). The reason is that to perform the splicing tests the gtf needs to have some extra annotated features.

Solution: run cuffcompare in the orginal gtf.
cuffcompare -s hg19.fasta -CG -r EnsemblGene-67.gtf EnsemblGene-67.gtf

This will create a cuff* ready annotation file.