Dear bioc-sig-sequencing,
Following up from my question from last month, I would like to determine an FDR
cutoff for a chipseq experiment. But first, would like to equalize/normalize
the number of reads between treatment & control aligned read input files
(BasicChipSeq.pdf,
http://www.bioconductor.org/workshops/2009/SeattleNov09/ChIP-seq/BasicChipSeq.pdf).
It was previously suggested to use laneSubsample() from chipseq package. This
would require conversion to GenomeData objects.
> load("ctcf.rda")
> load("gfp.rda")
> ctcf
class: AlignedRead
length: 484957 reads; width: 24 cycles
...
> gfp
class: AlignedRead
length: 316176 reads; width: 24 cycles
> ctcf_GD <- as(ctcf, "GenomeData")
> gfp_GD <- as(gfp, "GenomeData")
> laneSubsample(ctcf_GD, gfp_GD, fudge = 0.05)
How to test result of "laneSubsample()" (to check on the new number of reads in
the previously smaller file)?
Finally, it was suggested FDR cutoff can be determined using GenomeData objects
(or GRanges). How to do this?
For example, in "BasicChipSeq.pdf", after import into AlignedRead objects, ctcf
& gfp, then
> library(BSgenome.Mmusculus.UCSC.mm9)
> mouse.chromlens <- seqlengths(Mmusculus)
> > cov.ctcf <- coverage(ctcf, width = mouse.chromlens, extend = 126L)
> ...
How to determine a cutoff starting with GenomeData or GRanges objects rather
than AlignedRead?
Thanks,
P. Terry
[email protected]
[[alternative HTML version deleted]]
_______________________________________________
Bioc-sig-sequencing mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/bioc-sig-sequencing
