multi mapping parameters and centromere

[ptranvan]

Hi,

I am interested in looking at the mapping rate in the centromeres and was wondering which parameters I can tweak.

For TEs, I read that winAnchorMultimapNmax, outFilterMultimapNmax, and outSAMmultNmax are commonly tweaked.

What is your advice for repetitive regions such as centromeres? I am unsure about the parameters and the values I should set.

Thanks.

[alexdobin]

Hi Patrick,
Yes, increasing these parameters is needed to capture multimappers.

[ptranvan]

Thanks @alexdobin for your quick response.

1. Do you have any recommended values?

2. Also, I am wondering how I can compare coverage between different samples. Do you know any normalization methods?

[alexdobin]

--outFilterMultimapNmax and --outSAMmultNmax should be set to the maximum number of loci you want to map the reads.
--winAnchorMultimapNmax should twice as large.
However, if these values become too large (>~500), the mapping will be very slow.