SEQanswers (
-   Bioinformatics (
-   -   Problems with cuffmerge (

aniruddha.otago 06-25-2016 04:21 PM

Hi there,

I am running cuffmerge and having this problem:

My command:
cuffmerge -g /home/aniruddha/GTF_files/Human_genome/GRCh37_build_65/Homo_sapiens.GRCh37.65.gtf -p 10 /home/aniruddha/PDL1_RNA-Seq_8_AC/Cuffmerge_8PDcelllines/assembled_tc_list.txt

[Sun Jun 26 12:02:08 2016] Beginning transcriptome assembly merge

[Sun Jun 26 12:02:08 2016] Preparing output location ./merged_asm/
[Sun Jun 26 12:02:23 2016] Converting GTF files to SAM
Error: gtf_to_sam not found on this system. Did you forget to include it in your PATH?

When I see the cufflinks executables I dont see gtf_to_sam .
am I missing something ?

cufflinks-2.2.1.Linux_x86_64]$ ls
AUTHORS README cuffdiff cuffmerge cuffquant gffread isoforms.fpkm_tracking test_data.sam
LICENSE cuffcompare cufflinks cuffnorm genes.fpkm_tracking gtf_to_sam skipped.gtf transcripts.gtf

Thanks for your help.

Brian Bushnell 06-25-2016 04:46 PM

I recommend you avoid the Tuxedo tools. You can get much better results with something else. I won't give a specific alternative because I don't know the goal of your research, but I cannot imagine a question where the Tuxedo suite is the best answer. Or even a good answer.

aniruddha.otago 06-25-2016 09:18 PM

I think I have solved the problem now. For some reason gtf_to_sam tool was missing.
Anyways, for this particular analysis at this stage all I am wanting is to identify differential expression and differentially expressed isoforms. There is hardly any consensus of the best tools to be used. Taxedo suite is off course the kost used one. Anyone have any useful data driven opinion or comments on the use of tools for indetifying DE. ? Also if you have any comments on tools or approaches you think its good for comparing DE of genes with DNA methylation profiles of the same samples. ?
Thanks in advance.

Brian Bushnell 06-25-2016 09:38 PM

DESeq and edgeR are commonly used for differential expression analysis, and both are superior to Tuxedo tools, from what I understand. Cufflinks is incapable of determining that genes specific to the Y chromosome are differentially expressed between human males and females. If that's not sufficient evidence of it being a bad tool, I can't imagine what is. I don't currently have a link to that study, but I can probably get it in a few days when my co-worker who initially found it is available.

All times are GMT -8. The time now is 12:38 PM.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2020, vBulletin Solutions, Inc.