Strange maq 7.1 problem

[homeriq5]

Hey all,

I've recently installed maq on my laptop (running ubuntu 9.04 64-bit) and was able to get it up and running with no problem. I ran the following command and it created the appropriate output:

 
maq match test.map refchrom/chr22.bfa FAIRE_data/GM12878/UNCsequencing/raw_data/Rep2/s_2_sequence.bfq

Since maq was a little too much for my laptop to handle, I decided to have my Department's systems administrator install maq on our high performance computing cluster so that I could submit batch jobs to the cluster (using platform LSF). After about 30 seconds a segmentation fault occurs every time using the same command I successfully used on my laptop. It seems to happen right after maq sets the read length and before it loads the reads. Our systems administrator tried to roll back to an older version of maq (6.something) he still got the same problem. This is the email I got from LSF after the cluster tried to run the code and failed.

Job <maq match test.map refchrom/chr22.bfa FAIRE_data/GM12878/UNCsequencing/raw_data/Rep2/s_2_sequence.bfq> was submitted from host <diamond.bios.unc.edu> by user <nrashid>.
Job was executed on host(s) <andesite.bios.unc.edu>, in queue <bat>, as user <nrashid>.
</home/users/nrashid> was used as the home directory.
</home/users/nrashid> was used as the working directory.
Started at Tue Feb 10 17:02:21 2009
Results reported at Tue Feb 10 17:03:33 2009

Your job looked like:

------------------------------------------------------------
# LSBATCH: User input
maq match test.map refchrom/chr22.bfa FAIRE_data/GM12878/UNCsequencing/raw_data/Rep2/s_2_sequence.bfq
------------------------------------------------------------

Exited with exit code 139.

Resource usage summary:

    CPU time   :     20.09 sec.
    Max Memory :      1915 MB
    Max Swap   :      2850 MB

    Max Processes  :         3
    Max Threads    :         3
[B]
The output (if any) follows:

-- maq-0.7.1
[ma_load_reads] loading reads...
[ma_load_reads] set length of the first read as 35.
/home/users/nrashid/.lsbatch/1234303339.152072: line 8: 29828 Segmentation fault      (core dumped) maq match test.map refchrom/chr22.bfa FAIRE_data/GM12878/UNCsequencing/raw_data/Rep2/s_2_sequence.bfq[/B]

Is anyone else having this problem? Any guesses why it would work on my computer but not on LSF? We used the same exact binaries to install the files! Thanks so much!

[rdeborja]

is there a core file from the seg fault?

[homeriq5]

There is a core file, however its 2.8 GB. I'll try to see if I can trim the sizes of my input files so that its a lot smaller. Thanks for the reply!

[bagdevi]

Hi,

I am very new to NGS data assembly and analysis. I am using maq for the first time. When I am running the following command

$ maq match out.map ref.bfa reads.bfq

I encountered Segmentation fault (core dumped) error

The maq.exe.stackdump file contain the followings

Exception: STATUS_ACCESS_VIOLATION at eip=61113783
eax=00000047 ebx=0043DAE7 ecx=00001C4F edx=00000000 esi=00CA8538 edi=00000000
ebp=0028C428 esp=0028C424 program=C:\cygwin\usr\local\bin\maq.exe, pid 5196, thr
ead main
cs=0023 ds=002B es=002B fs=0053 gs=002B ss=002B
Stack trace:
Frame Function Args
0028C428 61113783 (00000000, 00000000, 00000000, 00000001)
0028C458 692C4EB2 (00000000, 0043DAE6, 00000000, 0028C484)
0028CCF8 0040F7E2 (00000004, 612438A8, 49435341, 61100049)
0028CD28 00401CD3 (00000000, 00000000, 0028CD68, 61007038)
0028CD68 61007038 (00000000, 0028CDA4, 61006980, 7EFDE000)
End of stack trace

I am running maq on cygwin environment on a 64 bit windows 7 laptop. both cygwin and maq are the latest versions.

I tried

make clean

but the error came again

Please help.

[CottonJack]

Somebody solved this problem?

[dpryan]

Is there a reason you're still using Maq?

[CottonJack]

Yes,

I`m trying differnet algorithms; I am mapping very short sequences (20-30 bp, mostly) with quite some mismatches.

Any suggestions?

Thanks.

[dpryan]

I would personally switch to bwa, it's the logical successor to Maq, which hasn't been updated in 5 years. Otherwise, you'll need to compile it with debugging and either run it inside gdb or have it dump a core file and load that. I hope you're comfortable with C.