Tophat Reporting Output Trakcs Failed!!

Question

[allqwdd]

2012-10-17 01:01:25] Reporting output tracks [FAILED]

Error: bammerge failed to open BAM file tophatSample/tmp/leftkeptreads.mapped.bam

that was the error message Is anyone knows why this error comes up? I think maybe my cufflinks is not right? here is my verison

PATH=$PATH:/home/iscsi/iscsi1/david/tools/boost1510 PATH=$PATH:/home/iscsi/iscsi1/david/tools/bowtie-0.12.8 PATH=$PATH:/home/iscsi/iscsi1/david/tools/tophat-2.0.5.Linuxx8664 PATH=$PATH:/home/iscsi/iscsi1/david/tools/samtools-0.1.7x8664-linux PATH=$PATH:/home/iscsi/iscsi1/david/tools/cufflinks-2.0.2.Linuxx8664 PATH=$PATH:/home/iscsi/iscsi1/david/tools/velvet1.1.05

and here is my command

nohup tophat -o tophatSample -p 50 --fusion-search --keep-fasta-order --bowtie1 --no-coverage-search -r 350 --mate-std-dev 80 -g 1 --fusion-min-dist 100000 --fusion-anchor-length 13 --fusion-ignore-chromosomes chrM /home/iscsi/iscsi1/david/hg19/hg19 /home/iscsi/iscsi1/david/hg19/S30R1100M1.fastq /home/iscsi/iscsi1/david/hg19/S30R1100M_2.fastq &

really appreciate if anyone knows how to solve the problem

[Zev.Kronenberg]

To get an answer you will need to reformat your question. Add context and make it clear what is going on.

Answer 1

[Arun]

First, you are running tophat. I don't understand why you feel cufflinks (note the spelling) is in anyway involved here. Second, You should put in some time and effort in understanding the website and editing a post. Its not difficult at all.

And to answer your question, these are the things I'd try to find out/do
1) Reduce your value for -p parameter (unless you have a cluster with 50 processors and an admin who's nice enough to allow you to make that many temporary files). -p 20 or -p 30 should be good (This creates a lot of temporary files and creating too many isn't good either and it would make things actually slower after the processing part).
2) If 1) isn't the issue, run tophat again with (--keep-tmp) parameter to go and check this file leftkeptreadsmapped.bam and find out if it has any reads in it at all. I tell this because the inner-distance you are using (350) seems quite large than what's obtained from Illumina protocols for me. Probably, many of your reads aren't even mapped.
3) If 2) doesn't work, try removing parameters and running it again (in parallel, probably with fewer threads so as to not kill the cluster), one at a time, to find out which parameter causes the trouble.

[allqwdd]

really appreciated your answer and apologize for my mistake

I'll try your suggestions.

and now I am getting to understand the posting format this website

thanks again