Re: Sequence loading ABYSS-P reader.eof() error

[Tony Raymond]

Hi Keith,

Have you been able to reproduce this error message? The error message occurs because ABySS stopped reading the file before it reached the end of the file. This usually occurs due to a malformed file.

What happens if you try running the following:
abyss-tofastq -v filename.fq

Thanks,
Tony

On 2012-08-02, at 5:12 PM, kaboroevich wrote:

> Hello,
>
> Compiling ABySS appears to run smoothy with both google sparce hash and
> MPI support. However, when I run abyss-pe I get the error:
>
> Reading `filename.fq'...
> ABYSS-P: AssemblyAlgorithms.cpp:104: void
> AssemblyAlgorithms::loadSequences(ISequenceCollection*, std::string):
> Assertion `reader.eof()' failed.
> --------------------------------------------------------------------------
> mpirun noticed that process rank 0 with PID 18306 on node nc420 exited
> on signal 11 (Segmentation fault).
> --------------------------------------------------------------------------
> make: *** [genome-1.fa] Error 139
>
> Running only ABYSS-P (based on the "abyss-pe -n" output) on a small 25
> sequence file still returns the error:
>
> $ ABYSS-P -k51 -q3 --coverage-hist=coverage.hist -s foo-bubbles.fa -o
> foo-1.fa foo.fq
> ABySS 1.3.1
> ABYSS-P -k51 -q3 --coverage-hist=coverage.hist -s foo-bubbles.fa -o
> foo-1.fa foo.fq
> Running on 1 processors
> 0: Running on host ngw02
> 0: Reading `foo.fq'...
> ABYSS-P: AssemblyAlgorithms.cpp:104: void
> AssemblyAlgorithms::loadSequences(ISequenceCollection*, std::string):
> Assertion `reader.eof()' failed.
> Segmentation fault (core dumped)
>
> $uname -a
> Linux node02 2.6.32-131.0.15.el6.x86_64 #1 SMP Tue May 10 15:42:40 EDT
> 2011 x86_64 x86_64 x86_64 GNU/Linux
>
> I've tried to find a solution to this problem but cannot find anything
> except for a reference to Mac OSX not reading compressed files. Any
> suggestions would be greatly appreciated.
>
> Regards,
> Keith

[kaboroevich]

Hello Tony,

I ran "abyss-tofastq -v" on a fastq file with a single sequence (4 lines). It printed the file contents then returned the error:

abyss-tofastq: abyss-tofastq.cc:104: void convert(const char*) [with Record = FastqRecord]: Assertion `in.eof()' failed.

I tried again after adding a trailing newline but that reported:

abyss-tofastq: ../Common/StringUtil.h:22: bool chomp(std::string&, char): Assertion `s.length() > 1' failed.

as could be expected. The same eof error also occurs when I try converting fasta files.

Keith

[Tony Raymond]

Hi Keith,

Fastq files can't have empty lines in them. It is good that you can reproduce the error. Please send me the test file you used so I can find out what's going.

Thanks,

Tony

[kaboroevich]

Hi Tony,

I repeated it with many fastq and fasta file variants and it failed each time. Simplest being a single sequence with the sequence and quality being 100 'A's. I've since compiled and ran it on a different machine without issue. 

In this case, is it better I send you one of the compiled executables rather than the arbitrary sequence file? If you would like me to compile it with different settings I could try that too.

Keith

On Wednesday, August 8, 2012 4:12:48 AM UTC+9, tray...@bcgsc.ca wrote:

Hi Keith,

[Tony Raymond]

Hi Keith,

Did you change compilers between machines by any chance? What OS is running on the machines?

I just wanted to see if there was any obvious problem with the sequence file, and if I could do what you did on a different machine. It doesn't sound like it since you ran it fine on a different machine. I likely won't be able to run your executable.

Thanks,

Tony