Tophat2 runs ending with truncated files [main_samview] truncated file

[giorgifm]

Dear all,

while running Tophat2 (default parameters, 64 bit centos), we keep encountering the following warning towards the end of the run:

Warning: couldn't remove all temporary files in /path
-----------------------------------------------
[2012-10-17 18:32:54] Run complete: 01:26:19 elapsed
[main_samview] truncated file.
[main_samview] truncated file.
[main_samview] truncated file.

We tried with Tophat v 2.0.4 and Tophat 2.0.5, paired with Bowtie v2beta7 and Samtools 0.1.18

My first guess was "lack of disk space", but we tried with terabytes of space, and the thing is still happening (in a deterministic way, the warnings are identical by number of "[main_samview] truncated file" lines).

The problem is not directly translated into a read loss. The number of read pairs contained in the sum accepted_hits.bam and unmapped.bam is always the same as the input fastq files. Any clue? Should we worry? Has anyone found the same problem?

Federico

[giorgifm]

I may add that the same problem can be found by splitting the input fastq files in smaller pieces. The "[main_samview] truncated file" error keeps appearing multiple times, leaving the tmp folders intact within the Tophat output.

[giorgifm]

It should be added that I reduced this problem by adding unique suffixes to the reads, e.g. /1 to first in pair and /2 to second in pair.

Moreover, the warning ([main_samview] truncated file) is specific to versions of Tophat 2.0.4 and 2.0.5, absent in Tophat 2.0.0

I hope this thread will help future people!

[mckeays]

Hello,

I am having the same problem, I have plenty of disk space but still get two "[main_samview] truncated file." lines at the end of my TopHat2 output. Would love to know exactly what is causing it...

I'd like to try your fix but I'm pretty new to all this so I have a couple of stupid questions: where did you add the \1 and \2 -- in the fastq files? And how did you do it?

Thanks!
Maria

[giorgifm]

Dear Maria,
to reduce the number of "truncated file" lines, I simply added suffixes to read names.

Reads first in pair:

@HWI-ST1210:91152CACXX:8:1101:1247:2126
@HWI-ST1210:91152CACXX:8:1101:1220:2217
@HWI-ST1210:91152CACXX:8:1101:1445:2126

@HWI-ST1210:91152CACXX:8:1101:1247:2126/1
@HWI-ST1210:91152CACXX:8:1101:1220:2217/1
@HWI-ST1210:91152CACXX:8:1101:1445:2126/1

Reads second in pair

@HWI-ST1210:91152CACXX:8:1101:1247:2126
@HWI-ST1210:91152CACXX:8:1101:1220:2217
@HWI-ST1210:91152CACXX:8:1101:1445:2126

@HWI-ST1210:91152CACXX:8:1101:1247:2126/2
@HWI-ST1210:91152CACXX:8:1101:1220:2217/2
@HWI-ST1210:91152CACXX:8:1101:1445:2126/2

However this does NOT seem to be related to the error that generates this warning:

#[2012-11-15 17:47:05] Reporting output tracks
#Warning: couldn't remove all temporary files in /projects/dummy/tophat/tmp/
#-----------------------------------------------
#[2012-11-15 18:10:45] Run complete: 02:08:14 elapsed
#[main_samview] truncated file.
#[main_samview] truncated file.
#[main_samview] truncated file.
#[main_samview] truncated file.
#[main_samview] truncated file.

I checked in the BAM outputs (accepted_hits.bam and unmapped.bam) that reads are NOT omitted, and the BAMs appear to be correct and complete. So this warning shouldn't compromise the work of Tophat, HOWEVER I still don't know what's causing it, except that it appeared between tophat 2.0.0 and tophat 2.0.4, and that it's present for different versions of the samtools as well.

[mckeays]

Thanks very much, I will give it a try. I have noticed that some of my "accepted_hits.bam" files are very small (e.g. 203M) but maybe that is how it should be.

[giorgifm]

Dear Maria, I am assuming you are using Illumina paired reads, right?
In that case, to check if your BAM is complete, you should check if the sum of pair names by merging accepted_hits.bam and unmapped.bam is identical to half the number of reads given as an input.

You can do so playing with samtools view, cut-f1, sort, uniq, wc-l and maybe sed to remove the "/1" and "/2" (the old Casava standard, the new one looks like 1:... and 2:...) from the read names in the unmapped.bam file.

[mckeays]

Aha, yes I have paired reads, I will give this a try also, thank you very much!

[hypno]

I have the same problem but I don't understand how to resolve...The count of the reads gave me a 50% of the sum of line in my input.

I have file _1 and _2 but I not obtain complete .sam file..ANy idea? Thanks for any help!

[ymc]

I have this error also. What are the consequences of this error?