> On October 15, 2014 at 3:02 PM Bert Brutzel
<bertbrutzel@xxxxxxxxxxxxxx> wrote:
> in the process of splitting a sequencing into two genomes I run into
the
> problem of a fastQ-File with duplicate read names, and duplicated
> sequences with * in them. The Problem occurs when I extract reads from
> the MAF file. This gives me asterix * in the sequence, which in turn
> gives me duplicate entries when combined with unused reads sourced
using
> the info_debrislist.txt. So my question how can I extract reads in
fastQ
> format from a reference mapping without having asterix * in the
sequence.
The gaps (asterisks) can be dealt with by using '-d'.
Now to the general approach:
- to perform a first split of reads into two organisms, I would simply
map the whole data to both organisms at once. That is, have the
reference sequence contain all contigs/genomes from both organisms.
The remaining reads I would assemble de-novo and try to use GC content
or other ways (intron/exon vs. no intron/exon) to try to assign them
to one of the organisms.
- extracting reads: the info directory contains a file named
"*contigreadlist*" which, suprise, contains info regarding which read
was assigned to which contig. Instead of going the miraconvert -M way
one could use that with -n.
- extracting reads: miraconvert -M may still be useful if you want to
extract pre-processed (i.e.: clipped) and partly error corrected
reads. Just remember to use -C.
HTH,
B.
PS: if you need to filter for duplicate sequences (which shouldn't be
in your data anyway): don't do it on sequences, always names.
