On Mon, Oct 19, 2009 at 1:01 PM, Martin Morgan <[email protected]> wrote:

> Michael Lawrence wrote:
> > On Mon, Oct 19, 2009 at 10:25 AM, Chris Seidel <[email protected]>
> wrote:
> >
> >> In using gdapply to get coverage over all chromosomes of a
> >> GenomeDataList object:
> >>
> >> cov <- gdapply(extendedReads, coverage)
>
> a different route to this point (<...> representing arguments you'd
> provide) is
>
>  aln <- readAligned(<...>) # or aln <- AlignedRead(<...>)
>  library(BSgenome.Hsapiens.UCSC.hg18)
>  wd <- seqlengths(Hsapiens)
>  ## subset aln and wd so that they have matching names, e.g.,
>  ## it might also be necessary to adjust names, e.g,. dropping '.fa'
>  aln <- aln[chromosome(aln) %in% names(wd)]
>  wd <- wd[names(wd) %in% chromosome(aln)]
>  cvg <- coverage(aln, width=wd, extend=200L)
>
> this also results in a list of Rle's (a SimpleRleList). See
> ?"AlignedRead-class" for details; this is a non-trivial operation, e.g.,
> it maybe useful to separately manage strands.
>
>
Yet another route would be to coerce the 'extendedReads' object to a
RangedData, and then call coverage on that, with the widths.

Something like:
cvg <- coverage(as(extendedReads, "RangedData"), width=wd)

This works even when one does not start with an AlignedRead object (but
usually one will). The reason we do not yet have a coverage method for
'GenomeData' is that we have been treating GenomeData as a generic catch-all
container, for storing per-chromosome objects, that is coerced into data
structures designed for specific use cases.

Michael

Martin
>
>
> >>
> >> how do you supply the "width" argument to coverage? When I try handing
> >> it a vector of chromosome lengths, it complains that it needs a single
> >> value.
> >>
> >>
> > This is the normal behavior of apply functions in R. The additional
> > arguments are passed in whole to the function every time it is invoked.
> We
> > don't have an "mapply" variant of gdapply, so you'll need to do some
> extra
> > work, like keeping track of an index.
> >
> > Perhaps we could make this easier using metadata? We know the genome of
> > 'extendedReads' so 'coverage' could be smart and refer to the appropriate
> > BSgenome package.
> >
> > Michael
> >
> >
> >
> >> -Chris
> >>
> >> _______________________________________________
> >> Bioc-sig-sequencing mailing list
> >> [email protected]
> >> https://stat.ethz.ch/mailman/listinfo/bioc-sig-sequencing
> >>
> >
> >       [[alternative HTML version deleted]]
> >
> > _______________________________________________
> > Bioc-sig-sequencing mailing list
> > [email protected]
> > https://stat.ethz.ch/mailman/listinfo/bioc-sig-sequencing
>
>
> --
> Martin Morgan
> Computational Biology / Fred Hutchinson Cancer Research Center
> 1100 Fairview Ave. N.
> PO Box 19024 Seattle, WA 98109
>
> Location: Arnold Building M1 B861
> Phone: (206) 667-2793
>

        [[alternative HTML version deleted]]

_______________________________________________
Bioc-sig-sequencing mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/bioc-sig-sequencing

Reply via email to