Help please, strange error message

[Bin Wang]

Hi,

I only have paired RNAseq data, when I tried:

run_grit.py -v --rnaseq-reads accepted_hits.bam --reference Ensembl_mm10.gtf --use-reference-tss-exons --use-reference-tes-exons

the error message is as belows:

Traceback (most recent call last):

  File "/soft/python/2.7/2.7.6-vanilla/python/bin/run_grit.py", line 5, in <module>

    pkg_resources.run_script('GRIT==1.1.2', 'run_grit.py')

  File "build/bdist.linux-x86_64/egg/pkg_resources.py", line 528, in run_script

  File "build/bdist.linux-x86_64/egg/pkg_resources.py", line 1401, in run_script

    # we assume here that our metadata may be nested inside a "basket"

  File "/soft/python/2.7/2.7.6-vanilla/python/lib/python2.7/site-packages/GRIT-1.1.2-py2.7-linux-x86_64.egg/EGG-INFO/scripts/run_grit.py", line 547, in <module>

  File "/soft/python/2.7/2.7.6-vanilla/python/lib/python2.7/site-packages/GRIT-1.1.2-py2.7-linux-x86_64.egg/EGG-INFO/scripts/run_grit.py", line 523, in main

  File "/soft/python/2.7/2.7.6-vanilla/python/lib/python2.7/site-packages/GRIT-1.1.2-py2.7-linux-x86_64.egg/EGG-INFO/scripts/run_grit.py", line 482, in iteritems

  File "/soft/python/2.7/2.7.6-vanilla/python/lib/python2.7/site-packages/GRIT-1.1.2-py2.7-linux-x86_64.egg/EGG-INFO/scripts/run_grit.py", line 452, in _build_elements

  File "/soft/python/2.7/2.7.6-vanilla/python/lib/python2.7/site-packages/GRIT-1.1.2-py2.7-linux-x86_64.egg/EGG-INFO/scripts/run_grit.py", line 255, in get_reads

  File "/soft/python/2.7/2.7.6-vanilla/python/lib/python2.7/site-packages/GRIT-1.1.2-py2.7-linux-x86_64.egg/EGG-INFO/scripts/run_grit.py", line 197, in _load_rnaseq_read

  File "build/bdist.linux-x86_64/egg/grit/files/reads.py", line 492, in init

  File "build/bdist.linux-x86_64/egg/grit/files/reads.py", line 124, in read_pairs_are_on_same_strand

ValueError: Reads appear to be a mix of reads on the same and different strands. ({'no_mate': 0, 'same_strand': 16973.0001, 'diff_strand': 83027.0001})

My bam file is the output from tophat.

Any hints are welcome. Thank you.

[Nathan Boley]

Dear Bin,

> File "build/bdist.linux-x86_64/egg/grit/files/reads.py", line 492, in init
> File "build/bdist.linux-x86_64/egg/grit/files/reads.py", line 124, in
> read_pairs_are_on_same_strand
> ValueError: Reads appear to be a mix of reads on the same and different
> strands. ({'no_mate': 0, 'same_strand': 16973.0001, 'diff_strand':
> 83027.0001})

This means that some fraction of your paired end reads (~17%) map to
the same strand, while the other fraction map to opposite strands.
GRIT tries to infer whether they are supposed to map to opposite
strands or they are supposed to map to the same strand, but in some
cases can't determine it.

Are the reads a mixture of different protocols, or are they from the
same protocol? If the former, you should split the different protocols
into separate bam files, and then pass GRIT multiple bam files. If the
latter, you need to determine what the protocol was (e.g. if read
pairs are supposed to map to the same strand or opposite strands) and
then I would filter any read pairs that map incorrectly from the bam
file.

The next release will have an option for specifying the read type
explicitly from the command line.

Best, Nathan

[Nathan Boley]

Dear Bin,

> I only have paired RNAseq data, when I tried:
> run_grit.py -v --rnaseq-reads accepted_hits.bam --reference Ensembl_mm10.gtf
> --use-reference-tss-exons --use-reference-tes-exons

Also, you'll probably have better results with
--use-reference-promoters --use-reference-polyas
--use-reference-junctions instead of --use-reference-tss-exons
--use-reference-tes-exons.

In low coverage experiments the reference junctions provide
connectivity, and I have code which filters promoters and poly(A)s
which aren't expressed, but not filtering code for distal exons.

Best, Nathan