[BioC] bsseq and BSmooth with RRBS

[Kasper Daniel Hansen]

On Thu, Feb 21, 2013 at 11:31 AM, Alex Gutteridge
<alexg at ruggedtextile.com> wrote:
> I'm looking over the documentation for bsseq:
>
> http://www.bioconductor.org/packages/2.11/bioc/html/bsseq.html
>
> And the associated Genome Biology paper:
>
> http://genomebiology.com/content/pdf/gb-2012-13-10-r83.pdf
>
> From what I can see the method should work on reduced representation
> bisulfite sequencing (RRBSeq) datasets as well as WGBSeq, though the
> documentation only mentions WGBSeq. The only issue I can see being the
> smoothing procedure which I thought might struggle in regions with scattered
> CpGs in the RRBSeq.
>
> I will have data in hand shortly to test this myself, but was just curious
> if anyone on the list had run bsseq (successfully or otherwise) on RRBS data
> before?

I don't have any hands on experience with RRBS, but I agree with your
assessment that in principle it should work, depending on how large
regions are actually captured in your experiment.  Mathematically, it
should run out of the box.  Pay close attention to the maxGap
parameter which sets a limit for how far away two neighboring CpGs can
be before not smoothing.

I have had inquires like yours about using it for RRBS data, but I
never heard back about their experience (and as I recall it, it was
much earlier in the development phase, so the software ought to work
much better now).

I would be happy to hear your experience.

Note that currently, (some of) the statistics are really meant for the
case where you have biological replicates, but that question is really
orthogonal to the choice of assay.

Best,
Kasper


> --
> Alex Gutteridge
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives:
> http://news.gmane.org/gmane.science.biology.informatics.conductor