[BioC] dexseq_count.py SAM "NH" taq

Alejandro Reyes alejandro.reyes at embl.de
Thu Feb 20 13:30:57 CET 2014


Hi Jose,

Both of the solutions that you mention should work.  We need to think of 
a smarter solution for this, but it is difficult when the sam 
specifications can have many interpretations from the aligner authors!

Best regards,
Alejandro


>
>     Hi Alejandro,
>     I have updated my dexseq_count.py and dexseq_prepare_annotation.py
>     script to use it with the -r parameter to ignore the exons with
>     overlap between genes. Now I am using dexseq_count.py like this:
>
>     dexseq_count.py -p yes -s no -a 10 -f sam -r name $SAM $GFF
>
>     *Traceback (most recent call last):*
>
>     *  File "/home/jgarcia/dexseq_count.py", line 225, in <module>*
>
>     *    rs = map_read_pair( af, ar )*
>
>     *  File "/home/jgarcia/dexseq_count.py", line 132, in map_read_pair*
>
>     *    if ar != None and ar.optional_field("NH") > 1:*
>
>     *  File "_HTSeq.pyx", line 1371, in
>     HTSeq._HTSeq.SAM_Alignment.optional_field (src/_HTSeq.c:25899)*
>
>     *KeyError: 'SAM optional field tag NH not found'*
>
>
>     My SAM files do not have indeed the NH tag, but with previous
>     versions of the script it worked. The aligment is made by
>     SOAPSplice that already gives you only single hits in our
>     settings. I know ht-seq_count.py checks as well for the NH tag,
>     but it does not yield an error and simply produces a 0 in
>     multiple-alignments filed at the tail of the count file.
>
>     Do I have to add manually a NH tag =1 or can I use an older
>     version of the dexseq_count.py script to skip this error? I have
>     updated mainly to use the -r option in dexseq_prepare but I do not
>     need using the -r or -f parameters
>
>     My version of DEXSeq is 1.8.0 and I had an older version under
>     1.3.3 that worked
>
>     Thanks
>
>     Jose
>
>
>
>
>
>
> -- 
> Jose M. Garcia Manteiga PhD
> Research Assistant - Data Analysis in Functional Genomics
> Center for Translational Genomics and BioInformatics
> Dibit2-Basilica, 3A3
> San Raffaele Scientific Institute
> Via Olgettina 58, 20132 Milano (MI), Italy
>
> Tel: +39-02-2643-4751
> e-mail: garciamanteiga.josemanuel at hsr.it 
> <mailto:garciamanteiga.josemanuel at hsr.it>



More information about the Bioconductor mailing list