TopHat ERROR: Segment join failed with err = 1

[rebrendi]

Hello all,

I am running TopHat for the first time, but but it gives error. Could you please have a look at the output?

Preparing output location ./tophat_out/
Checking for Bowtie index files
Checking for reference FASTA file
Warning: Could not find FASTA file /bowtie-0.12.5/indexes/xxxxa.fa
Reconstituting reference FASTA file from Bowtie index
Checking for Bowtie
Bowtie version: 0.12.x.0
Checking for Samtools
Samtools Version: 0.1.16
Checking reads
min read length: 36bp, max read length: 36bp
format: fastq
quality scale: phred33 (default)
Mapping reads against xxxx with Bowtie
Joining segment hits
[FAILED]
Error: Segment join failed with err = 1

[rebrendi]

still relevant...

[biznatch]

Look in the log files there may be more detailed errors.

[rebrendi]

Here is what was in the logs:

# reads processed: 29130490
# reads with at least one reported alignment: 23830715 (81.81%)
# reads that failed to align: 4412560 (15.15%)
# reads with alignments suppressed due to -m: 887215 (3.05%)
Reported 89574818 alignments to 1 output stream(s)

long_spanning_reads v1.2.0 (1752)
--------------------------------------------
Opening /dev/null for reading
Opening /dev/null for reading
Opening /dev/null for reading
Opening ./tophat_out/tmp/left_kept_reads.bwtout for reading
Loading reference sequences...
Loading spliced hits...done
Loading junctions...done
Loading deletions...done
Error: could not get read # 34344441 from stream

prep_reads v1.2.0 (1752)
---------------------------
13364 out of 29143854 reads have been filtered out

/usr/local/bin/tophat -r 100 /bowtie-0.12.5/indexes/xxxx data/sequence.txt /usr/local/bin/prep_reads --min-anchor 8 --splice-mismatches 0 --min-report-intron 50 --max-report-intron 500000 --min-isoform-fraction 0.15 --output-dir ./tophat_out/ --max-multihits 40 --segment-length 25 --segment-mismatches 2 --min-closure-exon 100 --min-closure-intron 50 --max-closure-intron 5000 --min-coverage-intron 50 --max-coverage-intron 20000 --min-segment-intron 50 --max-segment-intron 500000 --sam-header ./tophat_out/tmp/stub_header.sam --max-insertion-length 3 --max-deletion-length 3 --inner-dist-mean 100 --inner-dist-std-dev 20 --no-microexon-search --fastq /sequence.txt
bowtie -q --un ./tophat_out/tmp/left_kept_reads_missing.fq --max /dev/null -v 2 -p 1 -k 40 -m 40 bowtie-0.12.5/indexes/xxxx ./tophat_out/tmp/left_kept_reads.fq | /usr/local/bin/fix_map_ordering --fastq ./tophat_out/tmp/left_kept_reads.fq - > .//usr/local/bin/long_spanni/usr/local/bin/long_spanning_reads --min-anchor 8 --splice-mismatches 0 --min-report-intron 50 --max-report-intron 500000 --min-isoform-fraction 0.15 --output-dir ./tophat_out/ --max-multihits 40 --segment-length 25 --segment-mismatches 2 --min-closure-exon 100 --min-closure-intron 50 --max-closure-intron 5000 --min-coverage-intron 50 --max-coverage-intron 20000 --min-segment-intron 50 --max-segment-intron 500000 --sam-header ./tophat_out/tmp/stub_header.sam --max-insertion-length 3 --max-deletion-length 3 --inner-dist-mean 100 --inner-dist-std-dev 20 --no-microexon-search ./tophat_out/tmp/xxxx.fa ./tophat_out/tmp/left_kept_reads.fq /dev/null /dev/null /dev/null ./tophat_out/tmp/left_kept_reads.bwtout > ./tophat_out/tmp/file2SmeWx

[rebrendi]

...and it has also created a directory /tmp, with several large files, some of them seem to look similar to BED format, but none of them has a formal extension, so it seems that the program did not finish working, and these are really temporal files.

Any suggestions?

[biznatch]

Yes those are temp files created by Tophat that are usually automatically deleted after Tophat runs successfully. I'm not sure exactly what's causing your problem, there's just the error "Error: could not get read # 34344441 from stream". Are you using an old version of Tophat? It looks like you're using v1.2.0, maybe try the newest version 1.3.3. I don't know if that will solve the problem but it's the first thing I would try.

[rebrendi]

I am using the latest version of TopHat, 1.3.3.
I just downloaded and installed it yesterday.
I do not know why it lists version 1.2.0.
My bowtie version is very outdated but it seems
that this is not the problem with Bowtie,
because bowtie has mapped the reads.
Can it be because of the memory problem or something like this?

[biznatch]

Ah ok the 1.2.0 must mean Bowtie then. I don't think it's a memory issue, I ran out of memory on a Tophat run and got a -6 error during "Searching for junctions via segment mapping", I think error 1 must be something else, but not sure. How much memory do you have and what species are you aligning to? You should be able to align to the mouse or human genome with under 4GB.

[rebrendi]

yeah, I have more than 4 GB memory.

[rebrendi]

just on case if someone else is getting this error,
I found a similar thread. It seems that this issue
has not been resolved yet, or the issue was different
for different people.

http://seqanswers.com/forums/showthread.php?t=4103

In my case the disc space is not the problem for sure.

[cedance]

rebrendi, About the tophat 1.2.0 and 1.3.3 version difference, are you working on a cluster? if so, the default "tophat" path (installed by your admin) maybe linked to version 1.2.0. This is a very silly thing, but happens more than one might think, as its often overlooked. If so, you'll have to add your tophat path before the default $PATH in your .bash_profile file.