Filtering multi-mapping reads

[dario]

Hello,

I just started using rseg - Many thanks to the authors to make it
available!

I have a question about the input reads to use for rseg.

As I understand it, rseg takes into account the fact that reads mapped
to
certain regions (deadzones) cannot be reliably considered as correctly
mapped (because of sequence duplication, unassembled genomic regions
etc.).
rseg therefore adjusts its parameters when analyzyng these regions,
right?

Does this imply that it is not necessary (and not recommendable) to
filter
out reads mapping equally well to different genomic positions? (I
usually
exclude reads with mapping quality MAPQ<15 before peak calling.)

Many thanks

Dario

[Song, Qiang]

The input reads should be filtered first, i.e., ambiguously mapped reads 

need to be discarded. As a result, bins with "deadzones"  have fewer reads

than there actually are, which is accounted for adjusting parameters

[dario]

Ok, got it. Many thanks for your reply.
(I might come back with more questions in a separate thread... Bear
with me!)

Dario

On Feb 21, 6:21 pm, "Song, Qiang" <qiang.s...@usc.edu> wrote:
> The input reads should be filtered first, i.e., ambiguously mapped reads
> need to be discarded. As a result, bins with "deadzones"  have fewer reads
> than there actually are, which is accounted for adjusting parameters
>