Prodigal Pointer Error, GTDBtk continues

[hjruscheweyh]

Environment

• Installed via pip (include the output of pip list)
• Using a conda environment (include the output of conda list && conda list --revisions)
• Using a Docker container (include the IMAGE ID of the container)

Server information

• CPU: model name : Intel(R) Xeon(R) CPU E7-4850 v4 @ 2.10GHz 128
• RAM: MemTotal: 1056527808 kB
• OS: NAME="Red Hat Enterprise Linux Server" VERSION="7.9 (Maipo)"

Debugging information

• gtdbtk.log has been included (drag and drop the file to upload).
• Genomes have been included (if possible, and there are few).

Additional comments

Hi there

this is more a note on error recovery.

I ran gtdbtk (2.1) on a few thousand MAGs and got this issue:

*** Error in `prodigal': free(): invalid pointer: 0x00007f3992dca010 ***
======= Backtrace: =========
/lib64/libc.so.6(+0x81329)[0x7f3998862329]
prodigal(+0x356f)[0x56438f86256f]
/lib64/libc.so.6(__libc_start_main+0xf5)[0x7f3998803555]
prodigal(+0x4409)[0x56438f863409]
======= Memory map: ========
56438f85f000-56438f861000 r--p 00000000 00:32 358935210122               GTDBTK210/bin/prodigal
56438f861000-56438fa50000 r-xp 00002000 00:32 358935210122              GTDBTK210/bin/prodigal
56438fa50000-5643903f2000 r--p 001f1000 00:32 358935210122               GTDBTK210/bin/prodigal
5643903f3000-5643903f4000 r--p 00b93000 00:32 358935210122               GTDBTK210/bin/prodigal
5643903f4000-5643903f5000 rw-p 00b94000 00:32 358935210122               GTDBTK210/bin/prodigal
564390773000-564390794000 rw-p 00000000 00:00 0                          [heap]
7f398c000000-7f398c021000 rw-p 00000000 00:00 0
7f398c021000-7f3990000000 ---p 00000000 00:00 0
7f3992dca000-7f3992e53000 rw-p 00000000 00:00 0
7f39987e1000-7f39989a5000 r-xp 00000000 fd:00 536872223                  /usr/lib64/libc-2.17.so
7f39989a5000-7f3998ba4000 ---p 001c4000 fd:00 536872223                  /usr/lib64/libc-2.17.so
7f3998ba4000-7f3998ba8000 r--p 001c3000 fd:00 536872223                  /usr/lib64/libc-2.17.so
7f3998ba8000-7f3998baa000 rw-p 001c7000 fd:00 536872223                  /usr/lib64/libc-2.17.so
7f3998baa000-7f3998baf000 rw-p 00000000 00:00 0
7f3998baf000-7f3998cb0000 r-xp 00000000 fd:00 536947083                  /usr/lib64/libm-2.17.so
7f3998cb0000-7f3998eaf000 ---p 00101000 fd:00 536947083                  /usr/lib64/libm-2.17.so
7f3998eaf000-7f3998eb0000 r--p 00100000 fd:00 536947083                  /usr/lib64/libm-2.17.so
7f3998eb0000-7f3998eb1000 rw-p 00101000 fd:00 536947083                  /usr/lib64/libm-2.17.so
7f3998eb1000-7f3998ed3000 r-xp 00000000 fd:00 536872059                  /usr/lib64/ld-2.17.so
7f39990a1000-7f39990a4000 r--p 00000000 00:32 362997430686              GTDBTK210/lib/libgcc_s.so.1
7f39990a4000-7f39990b6000 r-xp 00003000 00:32 362997430686              GTDBTK210/lib/libgcc_s.so.1
7f39990b6000-7f39990b9000 r--p 00015000 00:32 362997430686               GTDBTK210/lib/libgcc_s.so.1
7f39990b9000-7f39990ba000 r--p 00017000 00:32 362997430686               GTDBTK210/lib/libgcc_s.so.1
7f39990ba000-7f39990bb000 rw-p 00018000 00:32 362997430686               GTDBTK210/lib/libgcc_s.so.1
7f39990bb000-7f39990be000 rw-p 00000000 00:00 0
7f39990cc000-7f39990d2000 rw-p 00000000 00:00 0
7f39990d2000-7f39990d3000 r--p 00021000 fd:00 536872059                  /usr/lib64/ld-2.17.so
7f39990d3000-7f39990d4000 rw-p 00022000 fd:00 536872059                  /usr/lib64/ld-2.17.so
7f39990d4000-7f39990d5000 rw-p 00000000 00:00 0
7fff6518c000-7fff6523f000 rw-p 00000000 00:00 0                          [stack]
7fff653a1000-7fff653a3000 r-xp 00000000 00:00 0                          [vdso]
ffffffffff600000-ffffffffff601000 r-xp 00000000 00:00 0                  [vsyscall]

and gtdbtk continued not seeing the error. I assume that checking the exit value of prodigal would solve this issue (see here:

GTDBTk/gtdbtk/biolib_lite/prodigal_biolib.py

Line 144 in 88ab52e

os.system(cmd)

) or run a subprocess with check=True.

best,
hans

[pchaumeil]

Hello,
Thanks for your feedback, we will modify GTDB-Tk so it can detect when Prodigal fails.

[wwood]

This is a very strange bug within prodigal. Get that error when the genome is in /tmp but not when I cp the file to my home directory.

So a workaround to get the genome included is to

mkdir  -p ~/t
TMPDIR=~/t gtdbtk ...

Only happens for some genomes, like 1 in 100 or something. The rest run fine in /tmp.

for context - GTDBtk copies the genomes to TMPDIR before running prodigal