[BioC] Tiling array analysis, ACME problem

[adam_pgsql]

On 1 Jun 2010, at 14:55, Sean Davis wrote:

> 
> 
> On Tue, Jun 1, 2010 at 9:11 AM, adam_pgsql <adam_pgsql at witneyweb.org> wrote:
> 
> Hi,
> 
> I am looking for modules that can analyse overlapping tiling arrays (Agilent) used for expression studies. In testing ACME, the example in the documentation fails with this error:
> 
> [R 2.10.1, OSX 10.6.3]
> 
> > library(ACME)
> > samps <- data.frame(SampleID = 1:3, Sample = paste("Sample", 1:3))
> > dat <- matrix(rnorm(90000), nc = 3)
> > colnames(dat) <- 1:3
> > pos <- rep(seq(1, 10000 * 100, 100), 3)
> > chrom <- rep(paste("chr", 1:3, sep = ""), each = 10000)
> > annot <- data.frame(Chromosome = chrom, Location = pos)
> > a <- new("aGFF", data = dat, annotation = annot, samples = samps)
> > calc <- do.aGFF.calc(a, window = 1000, thresh = 0.95)
> Error in function (classes, fdef, mtable)  :
>  unable to find an inherited method for function "chromosome", for signature "aGFF"
> 
> Am I missing something?
> 
> Yes, but that is my fault.  As of the version of ACME that you are likely using (I can't tell because there is no sessionInfo()), the recommended data structure is the ACMESet rather than an aGFF.  This data structure is what is used in the vignette.  However, the do.aGFF.calc help is INCORRECT and the "x" argument should be an ACMESet object.  So, working through your example (and thanks for the nice, reproducible example):
> 
> samps <- data.frame(SampleID = 1:3, Sample = paste("Sample", 1:3),row.names=1:3)
> dat <- matrix(rnorm(90000), nc = 3)
> colnames(dat) <- 1:3
> starts <- rep(seq(1, 10000 * 100, 100), 3)
> ends <- starts +50 # ACMESet needs ends, also, but could be equal to starts
> chrom <- rep(paste("chr", 1:3, sep = ""), each = 10000)
> a <- new("ACMESet",exprs=dat,
>     phenoData=as(samps,"AnnotatedDataFrame"),
>     featureData=as(data.frame(chromosome=chrom,start=starts,end=ends),"AnnotatedDataFrame"))
> calc = do.aGFF.calc(a,window = 1000, thresh = 0.95)
> 
> I will fix the documentation to make things clearer.  Sorry for the confusion.

ok thanks, that does the trick Sean.

adam