brie-factor utility causes error

[linkerst]

I am following the instructions of the brie manual ([(https://brie-rna.sourceforge.io/manual.html)]). At step 4 I run into the following error:

brie-factor -a AS_events/SE.gold.gtf -r GRCm38.p5.genome.fa -c mm10.60way.phastCons.bw -o mouse_features.csv -p 10

loading annotation file... Done.
extracting features for 5227 skipping exon triplets with 10 cores...
Traceback (most recent call last):
File "/myhome /conda-envs/ml_0/bin/brie-factor", line 11, in
sys.exit(main())
File "/myhome /conda-envs/ml_0/lib/python2.7/site-packages/brie/brie_factor.py", line 152, in main
RV = result[g].get()
File "/myhome /conda-envs/ml_0/lib/python2.7/multiprocessing/pool.py", line 567, in get
raise self._value
OSError: [Errno 2] No such file or directory

I receive this error both with my and the example data. I treated the bigWigSummary utility as described in the manual and my multiprocessing packages should be fine.
The first error message looks like the program has trouble finding the relative imports from .utils.gtf_utils import loadgene .

Could you please help me finding the issue?

Thanks
Stephanie

P.S
The code in step 3
brie-event-filter -a AS_events/SE.gff3 -anno_ref gencode.vM12.annotation.gtf -r GRCm38.p5.genome.fa

should be changed to
brie-event-filter -a AS_events/SE.gff3 --anno_ref gencode.vM12.annotation.gtf -r GRCm38.p5.genome.fa

[huangyh09]

Hi Stephanie,

Thanks for using BRIE, and apologize for the delay. I just tested the brie-factor on my machine and it works well, but I can't direct see the problem in your case.

First, I don't think it is the problem with from .utils.gtf_utils import loadgene as the loading annotation file can be finished.

Second, could you try a few things more to diagnose the bug:

• run brie-factor -a AS_events/SE.gold.gtf -r GRCm38.p5.genome.fa -c mm10.60way.phastCons.bw -o mouse_features.csv -p 1 this will use only one CUP so it will avoid the error from multiprocessing.
• run bigWigSummary to see whether you could run it directly.

Could you let me if you try these two things?

Best,
Yuanhua

[linkerst]

Dear Yuanhua,

Thanks for your reply.

• Running BRIE with the –p option leads to the same error I posted above.
• Running bigWigSummary directly in terminal works, e.g running bigWigSummary hg38.phyloP100way.bw chrX 4000 80000 1 has the output 0.0755242.

Best,
Stephanie

[huangyh09]

Hi Stephanie,

Sorry I haven't figured out this error. I suspect this is because of the incompatibility between the python2.7/multiprocessing/pool.py and the system or some other Python packages. I found quite a few similar errors report by searching your error report. You may get some ideas of the bug.

In addition, it is strange to see the same error when you tried -p 1 option, as this mode won't use multiprocessing any more.

Let me know if you have any idea of the bug.

Best,
Yuanhua

[linkerst]

Hi Yuanhua,

I think I managed to fix my problem.

First I created a new conda environment just for BRIE because sometimes I have issues with version conflicts. Afterward I was able to run brie-factor with the –p 1 option without using multiprocessing.

However, I run into a new bug concerning subprocessing. Adding shell=True to the subprocessing command in your get_factor function in the fasta_utils.py file (line 130) solved that problem und brie-factor run successfully.

Best,
Stephanie