Cannot fetch reads in region

[yang91]

Hi,

I tried brie to my own data, but it reported "Cannot fetch reads in regions: xxx" all the way, and finished with error showed below:

File "/home/yangxx/.local/bin/brie", line 9, in <module>
   load_entry_point('brie==0.1.3', 'console_scripts', 'brie')()
 File "build/bdist.linux-x86_64/egg/brie/brie.py", line 256, in main
 File "/usr/lib/python2.7/multiprocessing/pool.py", line 567, in get
   raise self._value
UnboundLocalError: local variable 'reads' referenced before assignment

My aligner is STAR 2.5.2b, and I already sort bam files during aligner and index with samtools. I also tried to resort output bam files and indexed. These all reported the same error. This is my command line:
brie -a $path/annotation/SE.gold.gtf -f $path/annotation/human_factors.SE.gold.csv -p 10 -s $path/STAR/SRR1033783.Aligned.sortedByCoord.out.bam -o $path/BRIE/SRR1033783. Is there something wrong about my input or parameter?

[huangyh09]

Hi,

That error looks odd, but I can't tell the reason so far. Your command line seems no problem. Could you simply try using one core, i.e., by setting -p 1? It will be slower, but may help me check whether the problem comes from the multiprocessing package.

Thanks and let me know.
Yuanhua

[yang91]

Hi Yuanhua,

I've tried with -p 1 parameter, error seems somewhere different:

[Brie] loading annotation file... Done.
[Brie] loading reads for 12586 genes with 1 cores...
[Brie] [--------------------] 0.1% done in 1.2 sec.Cannot fetch reads in region: 20:38902668-38907054
Traceback (most recent call last):
  File "/usr/local/bin/brie", line 9, in <module>
    load_entry_point('brie==0.1.3', 'console_scripts', 'brie')()
  File "build/bdist.linux-x86_64/egg/brie/brie.py", line 240, in main
  File "build/bdist.linux-x86_64/egg/brie/utils/run_utils.py", line 18, in set_info
  File "build/bdist.linux-x86_64/egg/brie/utils/tran_utils.py", line 270, in set_reads
  File "build/bdist.linux-x86_64/egg/brie/utils/sam_utils.py", line 74, in fetch_reads
UnboundLocalError: local variable 'reads' referenced before assignment

Then I tried with -p 2, the read processing could go to 100% rather than exit with 0.1% when -p 1, but it still reported same error with -p 10 parameter.

While something puzzles me that when during my brie installation, it got stucked by pylab module. You've offer help to forbiddden this function and installation went well ( see in #6 ).But I don't know if it matter. Do you think this error could be lead by this reason?

[huangyh09]

Hi, Thanks for the feedbacks. I think the error is clear now, by looking at the report with -p 1 (this doesn't use the multiple processors, so can directly report the error). The error happens because
No reads can be fetched from region 20:38902668-38907054
20 here means the chr20, while in your reference genome sequence, the chr20 is written as 20.

I would guess there is some incompatibility in your bam file, though I see you already sort and index it. Is it possible to check whether the reads on other genes can be fetched? When you say by using -p 2, all genes can processed, do you mean you can get the output results file? If so, maybe only this gene is incompatible to the reference genome.

Anyway, I have updated the sam_utils.py, to give reads=[] if reads cannot fetched from sam file. You could have a try and shouldn't make things worse. Also, I don't turning pylab off makes any differences here as it is only for plot.

Best
Yuanhua

[yang91]

Oh thank you for your reminding, I've used the human annotaion from brie to run with the mouse aligned bam files, such a fool! Very sorry for bothering you, and thank again for your kindly help!

[huangyh09]

Never mind, enjoy your analysis.