Any heuristic algorithm in tophat?

[zhiqiang]

Hi,

I am playing with the simulated RNA-seq data of the tophat&cufflinks paper published on Nature Protocol. While things go smoothly on my local PC installed with Ubuntu linux, problems occured on big servers installed with SUSE linux:

[Wed Apr 25 14:11:46 2012] Mapping right_kept_reads_seg1 against genome with Bowtie (1/3)
[Wed Apr 25 14:12:18 2012] Mapping right_kept_reads_seg2 against genome with Bowtie (2/3)
[Wed Apr 25 14:12:50 2012] Mapping right_kept_reads_seg3 against genome with Bowtie (3/3)
[Wed Apr 25 14:13:23 2012] Searching for junctions via segment mapping
[Wed Apr 25 14:16:10 2012] Retrieving sequences for splices
[FAILED]
Error: Splice sequence retrieval failed with err =1

After checking the log, I found the problem was located at run of "juncs_db"

juncs_db v1.3.2 (2689)
---------------------------
Error: malformed splice coordinate record

I then checked the input files of juncs_db, and I found that the file "segment.juncs" on the server was different from that generated on my PC. After replacing this file, this step would be OK.

I guess the step prior to "juncs_db" (i.e., segment_juncs) caused this problem, but cannot solve it.

Not sure any heuristic step caused different result files, or hardware or software problems.

Could anyone help?

Thanks.

The OS information: SUSE Linux Enterprise Server 11 SP1 (x86_64), Linux 2.6.32.12-0.7-default #1 SMP 2010-05-20 11:14:20 +0200 x86_64 x86_64 x86_64 GNU/Linux

SAMTOOLS: 0.1.18 built from source
bowtie-0.12.7-linux-x86_64.zip: downloaded from its homepage
tophat-1.3.2.Linux_x86_64.tar.gz: downloaded from its homepage

[zhiqiang]

I seems that the cause is the hardware, i.e., disks, since it runs OK on another disk.