Trouble With Rnaseq Analysis

Question: Trouble With Rnaseq Analysis

7.7 years ago by

Hi everybody, I am trying to analyze the differential expression between two RNAseq samples. But I found many troubles aligning my reads. I will describe what I did. First I groomed the FastQ files (2). Then I uploaded the Sorghum genome and aligned the reads to it with Tophat. Aftter that, I tried to use Cufflink with the BAM file of Tophat, using as annotation file an uploaded GTF file and the Sorghum genome, but I received an error message in the three outputs of Cufflink. I tried to align against new brand Maize genome (now at Galaxy), and the same messages. I also converted the BAM file to SAM, but the same. Any advice? What was wrong? Thanks in advance. Cristian

galaxy • 1.3k views

ADD COMMENT • link •

modified 7.7 years ago by Jeremy Goecks • 2.2k • written 7.7 years ago by Cristian Rojas • 70

7.7 years ago by

Jeremy Goecks • 2.2k

Jeremy Goecks • 2.2k wrote:

Cristian, Please share your history with me (History Options --> Share/Publish --> Share with User --> my email) and I'll take a look. Thanks, J.

ADD COMMENT • link written 7.7 years ago by Jeremy Goecks • 2.2k

7.7 years ago by

Jeremy Goecks • 2.2k

Jeremy Goecks • 2.2k wrote:

Cristian, The contig names in your GTF file don't match those in your reference (fasta) file. In order for Cufflinks to use a reference GTF, its contigs names must match those in your reference genome. Best, J.

ADD COMMENT • link written 7.7 years ago by Jeremy Goecks • 2.2k

7.7 years ago by

Jeremy Goecks • 2.2k

Jeremy Goecks • 2.2k wrote:

Cristian, The ">" is a formatting character; what needs to match is the string after the ">" in the genome file and the entries in the contig column of your GTF. Your GTF is quite different that your genome file; your genome file has 10 contigs labeled by number, but your GTF has many, many contig names labelled by numbers and names. For Cufflinks to work, you can either (a) turn off bias correction or (b) restrict entries in your GTF to those that match your reference genome. Finally, please reply all to emails so that all emails remain on list for archival and community purposes. Thanks, J.

ADD COMMENT • link written 7.7 years ago by Jeremy Goecks • 2.2k

Hi Jeremy, I turned off the bias correction and didnt use the genome option, just put the refernde (GTF file), but still the problem. Any help? De: Jeremy Goecks <jeremy.goecks@emory.edu> Para: Cristian Rojas <cristianrojasbr@yahoo.com.ar> CC: galaxy-user@lists.bx.psu.edu Enviado: miércoles, 30 de marzo, 2011 14:25:00 Asunto: Re: [galaxy-user] Trouble with RNAseq analysis Cristian, The ">" is a formatting character; what needs to match is the string after the ">" in the genome file and the entries in the contig column of your GTF. Your GTF is quite different that your genome file; your genome file has 10 contigs labeled by number, but your GTF has many, many contig names labelled by numbers and names. For Cufflinks to work, you can either (a) turn off bias correction or (b) restrict entries in your GTF to those that match your reference genome. Finally, please reply all to emails so that all emails remain on list for archival and community purposes. Thanks, J.

ADD REPLY • link written 7.7 years ago by Cristian Rojas • 70

I was able to run Cufflinks successfully with bias correction turned off; see the history that I've shared with you. Please try one more time and, if it doesn't work, report it using the bug icon next to the broken dataset. Thanks, J.

ADD REPLY • link written 7.7 years ago by Jeremy Goecks • 2.2k

I tried agaian and the same problem. I tuned off the bias correction but mantained the GFT file. May be this is the problem? I didnt find your history. Thanks De: Jeremy Goecks <jeremy.goecks@emory.edu> Para: Cristian Rojas <cristianrojasbr@yahoo.com.ar> CC: galaxy-user@lists.bx.psu.edu Enviado: miércoles, 30 de marzo, 2011 16:16:14 Asunto: Re: [galaxy-user] Trouble with RNAseq analysis I was able to run Cufflinks successfully with bias correction turned off; see the history that I've shared with you. Please try one more time and, if it doesn't work, report it using the bug icon next to the broken dataset. Thanks, J.

ADD REPLY • link written 7.7 years ago by Cristian Rojas • 70

Look for the history I've shared in History Options --> Histories Shared with Me. As requested, if you're still having problems, please report the problematic dataset by clicking on the bug icon. Thanks, J.

ADD REPLY • link written 7.7 years ago by Jeremy Goecks • 2.2k

Please log in to add an answer.

Similar posts • Search »