Seqanswers Leaderboard Ad

**mgogol** · 12-30-2009, 07:46 AM

me too

Just wanted to say that the same thing happened to me.

**mhughes** · 01-12-2010, 02:10 PM

ditto

Same issue. Did anyone figure out a fix?

any reason I couldn't just write a script to cull the redundant entries?

Thanks-

**mgogol** · 01-13-2010, 09:33 AM

That would probably work, but which RPKM is right and which would you delete?

**mhughes** · 01-13-2010, 10:05 AM

workable...

I'm working with the Drosophila, and it turns out there's only 2 redundant entires... didn't even need to write a script, I just deleted them by hand.

Not sure which RPKM is right, but I can live with 2 genes being a little wonky, especially because I can always double check by eye if they're something important. Sorry it doesn't help the mammalian folks, but at least CuffCompare can be made to work, if necessary.

**mgogol** · 01-13-2010, 11:40 AM

Oh, in that case, no biggie...

The mouse situation is a bit more complicated. I think I'm just going to not use cuffcompare for now.

**Uwe Appelt** · 07-13-2010, 12:31 AM

Dear all,

i just encountering a very similar problem. I ran cufflinks 0.8.3 on several samples, e.g.:

Code:

cufflinks --num-threads 8 --max-mle-iterations 10000 --GTF ./Homo_sapiens.GRCh37.58.name=id.gtf --output-dir ./2_cufflinks/PE_NN3 ./1_tophat/PE_NN3/accepted_hits.sam

Please note that the GTF-file used originates from

Code:

ftp://ftp.ensembl.org/pub/current_gtf/homo_sapiens/Homo_sapiens.GRCh37.58.gtf.gz

. I, however, changed the original GTF-file scriptually, such that all gene_name's match the according gene_id's (same for transcripts) to make cufflinks/cuffcompare/cuffdiff stick with id's throughout all three analysis stages (i don't like gene_names here). I finally double checked the resulting GTF file for duplicate entries, there are none.

Now, when i run cuffcompare i receive the following error message:

Code:

cuffcompare -T -R -o ./3_cuffcompare/REPL -r ./Homo_sapiens.GRCh37.58.name=id.gtf ./2_cufflinks/PE_NN3/transcripts.gtf $USSC/2_cufflinks/PE_NN5/transcripts.gtf

Code:

Error: duplicate GFF ID 'ENST00000415889' enountered!

(The actual duplicate id reported changes with the sample-composition a try to compare.)

Taking a look into the "transcripts.expr" files produced by cufflinks, it appears that there are indeed duplicate lines (really duplicate!). They match both in transcript_id's as well as in expression values. The according "transcripts.gtf" also produced by cufflinks contains the whole block for duplicate transcripts twice, too. Again, all letters and digits do match, so we're talking about real duplicates reported.

I'm not quite sure, but i feel like, i didn't had this problem in cufflinks 0.8.2.

Any ideas?

Best,
Uwe

**oliviera** · 07-13-2010, 03:22 AM

Dear Winfred,

I think I got a smiliar to yours.
Please a look to this threads and tell me if this is relevant to your problem.

tuning the tophat/cufflink pipeline - SEQanswers

http://seqanswers.com/forums/showthread.php?t=5208&highlight=oliviera

Application of sequencing to RNA analysis (RNA-Seq, whole transcriptome, SAGE, expression analysis, novel organism mining, splice variants)

Olivier

**Uwe Appelt** · 07-13-2010, 06:54 AM

Originally posted by oliviera View Post

Dear Winfred,

I think I got a smiliar to yours.
Please a look to this threads and tell me if this is relevant to your problem.

tuning the tophat/cufflink pipeline - SEQanswers

http://seqanswers.com/forums/showthread.php?t=5208&highlight=oliviera

Application of sequencing to RNA analysis (RNA-Seq, whole transcriptome, SAGE, expression analysis, novel organism mining, splice variants)

Olivier

Thanks for your quick reply, oliviera.

It is, however, the case that the transcript boundaries appear to not differ in their locations. Admittedly the RPKM values do actually differ with regard to the duplicate lines (i claimed them to be equal in my previous post). Please find the two excerpts from "transcripts.expr" and transcripts.gft, respectively. Maybe this tells anyone, what could be wrong with my setup...

transcripts.expr:

Code:

ENST00000513107	408319	5	44388586	44389808	0.02483	1	1	0	0.0744899	0.0442087	1131
ENST00000513107	408321	5	44388586	44389808	0.0192059	1	0.670969	0	0.0592762	0.0341952	1131

transcripts.gtf

Code:

5	Cufflinks	transcript	44388587	44389808	1000	-	.	gene_id "ENSG00000070193"; transcript_id "ENST00000513107"; FPKM "0.0248299624"; frac "1.000000"; conf_lo "0.000000"; conf_hi "0.074490"; cov "0.044209";
5	Cufflinks	exon	44388587	44389158	1000	-	.	gene_id "ENSG00000070193"; transcript_id "ENST00000513107"; exon_number "1"; FPKM "0.0248299624"; frac "1.000000"; conf_lo "0.000000"; conf_hi "0.074490"; cov "0.044209";
5	Cufflinks	exon	44389250	44389808	1000	-	.	gene_id "ENSG00000070193"; transcript_id "ENST00000513107"; exon_number "2"; FPKM "0.0248299624"; frac "1.000000"; conf_lo "0.000000"; conf_hi "0.074490"; cov "0.044209";
5	Cufflinks	transcript	44388587	44389808	1000	-	.	gene_id "ENSG00000070193"; transcript_id "ENST00000513107"; FPKM "0.0192058821"; frac "0.670969"; conf_lo "0.000000"; conf_hi "0.059276"; cov "0.034195";
5	Cufflinks	exon	44388587	44389158	1000	-	.	gene_id "ENSG00000070193"; transcript_id "ENST00000513107"; exon_number "1"; FPKM "0.0192058821"; frac "0.670969"; conf_lo "0.000000"; conf_hi "0.059276"; cov "0.034195";
5	Cufflinks	exon	44389250	44389808	1000	-	.	gene_id "ENSG00000070193"; transcript_id "ENST00000513107"; exon_number "2"; FPKM "0.0192058821"; frac "0.670969"; conf_lo "0.000000"; conf_hi "0.059276"; cov "0.034195";

There is -i triple checked that, after all- definitely only a single entry regarding transcript "ENST00000513107" in the source-GTF file. Below, i attached all GTF-lines that contain the transcript_id "ENST00000513107":

Code:

5	protein_coding	exon	44389250	44389808	.	-	.	 gene_id "ENSG00000070193"; transcript_id "ENST00000513107"; exon_number "1"; gene_name "ENSG00000070193"; transcript_name "ENST00000513107";
5	protein_coding	exon	44388587	44389158	.	-	.	 gene_id "ENSG00000070193"; transcript_id "ENST00000513107"; exon_number "2"; gene_name "ENSG00000070193"; transcript_name "ENST00000513107";
5	protein_coding	CDS	44388587	44388784	.	-	0	 gene_id "ENSG00000070193"; transcript_id "ENST00000513107"; exon_number "2"; gene_name "ENSG00000070193"; transcript_name "ENST00000513107"; protein_id "ENSP00000426406";
5	protein_coding	start_codon	44388782	44388784	.	-	0	 gene_id "ENSG00000070193"; transcript_id "ENST00000513107"; exon_number "2"; gene_name "ENSG00000070193"; transcript_name "ENST00000513107";

So i don't think it to be a transcript boundary problem, even more -if i remember correctly- that this particular problem didn't appear in v0.8.2.

Any other ideas? Thanks in advance!
Uwe

**ChrisL** · 07-15-2010, 02:35 AM

Same problem with "duplicate GFF ID".

For the moment I have switched back to Cufflinks 0.8.2.

**ShaunMahony** · 07-15-2010, 09:25 AM

I also found the same issue with version 0.8.3 but not 0.8.2. It looks like it's introduced by cufflinks when running with the "-G" option.

**Cole Trapnell** · 07-15-2010, 09:30 AM

This looks like a regression in the alignment+GTF record bundling code in Cufflinks. If anybody here can produce a relatively small test set that consistently reproduces this issue, I'll fix it pronto. Sorry for the annoying bug...

**Uwe Appelt** · 07-16-2010, 01:59 AM

Originally posted by Cole Trapnell View Post

This looks like a regression in the alignment+GTF record bundling code in Cufflinks. If anybody here can produce a relatively small test set that consistently reproduces this issue, I'll fix it pronto. Sorry for the annoying bug...

Hi Cole,

first of all, thanks for engagement! I by the way don't think it necessary to apologize for being in BETA stage. IMHO the Trapnell-toolsuite is one of the most complete and consistent ones out there, so thanks for that!

I've prepared a dataset that could hold as a test scenario for the issue discussed above. Hopefully it is somewhat useful, even though its not that "relatively small" ,-) Please find the accepted_hits.sam and the GTF file i used in the following download (1.3GB).

Code:

http://www.gtsg.org/_/COLE.tar.gz

Referring to the cmdline (cufflinks v0.8.3):

Code:

cufflinks --num-threads 8 --max-mle-iterations 10000 --GTF ./Homo_sapiens.GRCh37.58.name=id.gtf --output-dir ./ ./accepted_hits.sam

there should appear two duplicate transcript entries in transcripts.gtf as well as in transcripts.expr:

Code:

ENST00000320936
ENST00000423038

Thanks again and best!
Uwe

**ChrisL** · 07-16-2010, 02:31 AM

I'll second that. A great set of RNA-Seq tools even in their beta form.

Thanks Cole.

Chris

**Cole Trapnell** · 07-16-2010, 02:08 PM

I just posted a pre-release build of Cufflinks 0.8.4 (svn r1370) that addresses this issue. I was able to locally reproduce the bug with Uwe's test set (thanks!), and the build I've posted corrects it (for me, anyways). Please let me know if this crops up again (ideally with another test download to reproduce it).

You should know before trying this build that it is an SVN snapshot and hasn't gone through the normal round of release testing, so you may run into other issues.

This build also includes additional assembler fixes and direct BAM file support. You can now supply BAM files as input to Cufflinks (SAM is still supported).

Please download one of the tarballs with "0.8.4" in the version tag from:

404 Not Found

http://cufflinks.cbcb.umd.edu/downloads/

Topics	Statistics	Last Post
ASHG 2024 Highlights – Part Two by seqadmin Started by seqadmin, Today, 11:09 AM	0 responses 23 views 0 likes	Last Post by seqadmin Today, 11:09 AM
ASHG 2024 Highlights – Part One by seqadmin Started by seqadmin, Today, 06:13 AM	0 responses 20 views 0 likes	Last Post by seqadmin Today, 06:13 AM
Seq-Scope Expands Possibilities for High-Resolution Gene Expression Analysis by seqadmin Started by seqadmin, 11-01-2024, 06:09 AM	0 responses 30 views 0 likes	Last Post by seqadmin 11-01-2024, 06:09 AM
New Model Aims to Explain Polygenic Diseases by Connecting Genomic Mutations and Regulatory Networks by seqadmin Started by seqadmin, 10-30-2024, 05:31 AM	0 responses 21 views 0 likes	Last Post by seqadmin 10-30-2024, 05:31 AM

Seqanswers Leaderboard Ad

Announcement

Cufflinks outputs duplicate GTF entries?

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Latest Articles

ad_right_rmr

News