segmentation fault mapping against Danio rerio Zv9

[alexdobin]

segmentation fault mapping against Danio rerio Zv9

[alexdobin]

Hi Marcel,

please try the 2.4.0c release https://github.com/alexdobin/STAR/releases/tag/STAR_2.4.0c
I believe I fixed this issue (it only affected fasta files, not .fastq), but the git master has not been updated yet.

Cheers
Alex

[marcelm]

Hm, no, that doesn't help, it still segfaults. But as I see it, the tag STAR 2.4.0c is at the commit with id 0f81385, and that is actually the same commit where master is right now.

Cool you've started using Git, is it ok if I suggest something? Perhaps you could add *.o and Depend.list to a .gitignore file. Right now, when I compile STAR and then run git status, I get all these .o files in the output.

[marcelm]

Hi, just some additional information: I've just converted the FASTA file to FASTQ (with all quality values set to 40) and the program still segfaults.

[marcelm]

I've recomplied STAR with -g. Here's the backtrace from within gdb.

#0  stitchWindowAligns (iA=37, nA=37, Score=5369, WAincl=<value optimized out>, tR2=<value optimized out>, tG2=1219749765, trA=..., Lread=5828, WA=0x703a90, R=0x22dadc0 "", 
    Q=0x22ff7e0 '\001' <repeats 200 times>..., G=0x2aaaac04d0d8 "\003\003\001\003\003\001\003\002\002\002\002", sigG=0x2aaaaaefbe78 "C", <incomplete sequence \375>, P=0x6c2430, 
    wTr=0x228a450, nWinTr=0x2276bc0, RA=0x2aae219b3010) at stitchWindowAligns.cpp:87
#1  0x0000000000441215 in stitchWindowAligns (iA=36, nA=37, Score=5367, WAincl=0x6c8170, tR2=5366, tG2=1219749763, trA=..., Lread=5828, WA=0x703a90, R=0x22dadc0 "", 
    Q=0x22ff7e0 '\001' <repeats 200 times>..., G=0x2aaaac04d0d8 "\003\003\001\003\003\001\003\002\002\002\002", sigG=0x2aaaaaefbe78 "C", <incomplete sequence \375>, P=0x6c2430, 
    wTr=0x228a450, nWinTr=0x2276bc0, RA=0x2aae219b3010) at stitchWindowAligns.cpp:337

(33 further recursive calls removed)

#35 0x0000000000441127 in stitchWindowAligns (iA=2, nA=37, Score=1080, WAincl=0x6c8170, tR2=1079, tG2=1219696945, trA=..., Lread=5828, WA=0x703a90, R=0x22dadc0 "", 
    Q=0x22ff7e0 '\001' <repeats 200 times>..., G=0x2aaaac04d0d8 "\003\003\001\003\003\001\003\002\002\002\002", sigG=0x2aaaaaefbe78 "C", <incomplete sequence \375>, P=0x6c2430, 
    wTr=0x228a450, nWinTr=0x2276bc0, RA=0x2aae219b3010) at stitchWindowAligns.cpp:329
#36 0x0000000000441127 in stitchWindowAligns (iA=1, nA=37, Score=977, WAincl=0x6c8170, tR2=976, tG2=1219691293, trA=..., Lread=5828, WA=0x703a90, R=0x22dadc0 "", 
    Q=0x22ff7e0 '\001' <repeats 200 times>..., G=0x2aaaac04d0d8 "\003\003\001\003\003\001\003\002\002\002\002", sigG=0x2aaaaaefbe78 "C", <incomplete sequence \375>, P=0x6c2430, 
    wTr=0x228a450, nWinTr=0x2276bc0, RA=0x2aae219b3010) at stitchWindowAligns.cpp:329
#37 0x0000000000441127 in stitchWindowAligns (iA=0, nA=37, Score=0, WAincl=0x6c8170, tR2=0, tG2=0, trA=..., Lread=5828, WA=0x703a90, R=0x22dadc0 "", 
    Q=0x22ff7e0 '\001' <repeats 200 times>..., G=0x2aaaac04d0d8 "\003\003\001\003\003\001\003\002\002\002\002", sigG=0x2aaaaaefbe78 "C", <incomplete sequence \375>, P=0x6c2430, 
    wTr=0x228a450, nWinTr=0x2276bc0, RA=0x2aae219b3010) at stitchWindowAligns.cpp:329
#38 0x000000000042be9e in ReadAlign::stitchPieces (this=0x2aae219b3010, R=0x6c81b0, Q=0x6c81d0, G=0x2aaaac04d0d8 "\003\003\001\003\003\001\003\002\002\002\002", SA=<value optimized out>, 
    Lread=5828) at ReadAlign_stitchPieces.cpp:405
#39 0x000000000042df0d in ReadAlign::mapOneRead (this=0x2aae219b3010) at ReadAlign_mapOneRead.cpp:88
#40 0x000000000043e057 in ReadAlign::oneRead (this=0x2aae219b3010) at ReadAlign_oneRead.cpp:70
#41 0x0000000000433f37 in ReadAlignChunk::mapChunk (this=0x6dd110) at ReadAlignChunk_mapChunk.cpp:25
#42 0x00000000004327c0 in ReadAlignChunk::processChunks (this=0x6dd110) at ReadAlignChunk_processChunks.cpp:141
#43 0x000000000045760d in mapThreadsSpawn (P=0x6c2430, RAchunk=0x7fffffffaad0) at mapThreadsSpawn.cpp:17
#44 0x000000000040cd30 in main (argInN=<value optimized out>, argIn=<value optimized out>) at STAR.cpp:221

[alexdobin]

Hi Marcel,

sorry I should have looked at your sequence - it's very long - so you need to re-compile STAR with 'make STARlong', which will make proper allocations for long reads. Please let me know how it goes.

Cheers
Alex

[marcelm]

That works, thanks a lot!

Could you perhaps add a check to the standard STAR program such that it prints a message when the read is too long? At least in my case, it would have saved me a few hours of trying to figure out what's going on.

Anyway, cool that it's working now :)

[alexdobin]

Hi Marcel,

thanks for your suggestions, I have set the read length limit at 500b, and added the .gitignore in the the latest patch,

Cheers
Alex

[marcelm]

Great, thanks a lot!