Weirdness with spliced reads in samfile.pileup(chr,start,end) given spliced alignments from an RNA-seq bam file.

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Start with an rna-seq bam file
2. Query a region that includes spliced reads overlaping a position using 
pysam.pileup()
3. Try to get the allele frequencies at this position by querying the reads

What is the expected output? What do you see instead?
I see long stretches of base coverage counts from a single basepair comming 
from the entire length of the spliced region of each read that is spliced over 
a given position. There should be a read.is_spliced flag to deal with this in 
the pileup output.

What version of the product are you using? On what operating system?
0.7.4, ubuntu 13.04, python 2.7

Please provide any additional information below.
The pileup lines (when running mpileup) over these regions had characters 
rather than bases that looked something like this ">><>". These somehow get 
interpreted by samfile.pileup wrong as some arbitrary base call, and appear to 
be the same basecall for a given read for the stretch of the splicing event.

Original issue reported on code.google.com by johnthes...@gmail.com on 19 Jul 2013 at 12:01

[GoogleCodeExporter]

Correction to step 2 above, "samfile.pileup()", not pysam.mpileup(), or 
pysam.pileup(). As an intermediate solution I am parsing the output from 
pysam.mpileup() and trashing the odd "><" type characters that pop up in the 
rna-seq files, pretty sure these are from spliced alignments.

Original comment by johnthes...@gmail.com on 19 Jul 2013 at 12:04

[GoogleCodeExporter]

you can go ahead and close this issue. I just realized that if is_indel == 0 
and is_del is true, then the reads are in this class of being refskip. It would 
be great to throw that in the documentation though. Thanks!

Original comment by johnthes...@gmail.com on 19 Jul 2013 at 12:25

[GoogleCodeExporter]

Thanks for reporting this. 

I have added some text to the documentation.

Best wishes,
Andreas

Original comment by andreas....@gmail.com on 17 Sep 2013 at 7:20

• Changed state: WontFix