Error in rule read_length_histogram

[ayeletperes]

Hey,

I've been running IgDiscover (version 0.12) with light chain data, however, the run fails each time in this rule read_length_histogram.

I have used different datasets and still get the same error. Also tried to change the number of threads, but that didn't help either.

read_length_histogram: WARNING: QApplication was not created in the main() thread. terminate called without an active exception /bin/sh: line 9: 256081 Aborted (core dumped) /.conda/envs/igdiscover/bin/python -m snakemake stats/readlengths.pdf --snakefile /.conda/envs/igdiscover/lib/python3.6/site-packages/igdiscover/Snakefile --force -j64 --keep-target-files --keep-remote --attempt 1 --force-use-threads --wrapper-prefix None --latency-wait 3 -p --allowed-rules read_length_histogram --notemp --quiet --no-hooks --nolock --mode 1

[marcelm]

This looks very unusual. I’m wondering whether the QApplication warning has something to do with the problem or whether that is inpedendent.

• Are you running this on macOS or Linux?
• Do you have a graphical user interface available?
• Were any other PDF or PNG plots created successfully before the crash occurred?

I’m guessing the problem has something to do with plotting. Plotting is done by matplotlib, and the QApplication warning points to Qt (QApplication is the name of one of the components of Qt). Matplotlib does have a Qt backend, but IgDiscover is supposed to not use it at all, so that is the strange part.

Do you have a matplotlib configuration file named ~/.config/matplotlib/matplotlibrc?

Is your DISPLAY environment variable set? (Find out with echo $DISPLAY. If it prints something, it is set.) If so, try to run IgDiscover with this command:

DISPLAY= igdiscover run ...

(where ... are the options you would normally use).

[ayeletperes]

I'm running on a linux server with a graphical user. It seems that the problem was with the DISPLAY variable. Your solution worked. Added the DISPLAY variable before the run. Thanks!

On Sun, Feb 9, 2020 at 4:22 PM Marcel Martin notifications@github.com wrote:

This looks very unusual. I’m wondering whether the QApplication warning has something to do with the problem or whether that is inpedendent.

• Are you running this on macOS or Linux?
• Do you have a graphical user interface available?
• Were any other PDF or PNG plots created successfully before the crash occurred?

I’m guessing the problem has something to do with plotting. Plotting is done by matplotlib, and the QApplication warning points to Qt (QApplication is the name of one of the components of Qt). Matplotlib does have a Qt backend, but IgDiscover is supposed to not use it at all, so that is the strange part.

Do you have a matplotlib configuration file named ~/.config/matplotlib/matplotlibrc?

Is your DISPLAY environment variable set? (Find out with echo $DISPLAY. If it prints something, it is set.) If so, try to run IgDiscover with this command:

DISPLAY= igdiscover run ...

(where ... are the options you would normally use).

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/NBISweden/IgDiscover/issues/106?email_source=notifications&email_token=AGK4HYJQYS77GIEVFVSXQPLRCAGRVA5CNFSM4KSASAB2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOELGNXCY#issuecomment-583850891, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGK4HYIVOIP4T6U57T5IJHTRCAGRVANCNFSM4KSASABQ .

[marcelm]

Great, glad I could help! I’ll try to find the actual cause of the problem – after all, it should work even when the DISPLAY variable is set.