[BioC] errors in imageplot() in limma

[Gordon Smyth]

Dear Michael,

At 01:47 AM 29/07/2003, michael watson (IAH-C) wrote:
>Hi
>
>I am getting some rather confusing messages from the imageplot() function 
>in limma.  As in my previous mail, I am using a version of Bioconductor 
>1.2 from the website from about 2 weeks ago.
>
>I am reading the data in from a GenePix file but I want to use Medians 
>rather than Means, so I did:
>
> >RG <- 
> read.maimages("F:\\IBDV_Experiment\\same_layout\\lame#9.adjusted.gpr",columns=list(Rf 
> = "F635.Median",Gf = "F532.Median", Rb = "B635.Median", Gb = "B532.Median"))
>
>This seemed to work.  Then:

This is fine.

> >MA <- MA.RG(RG)
>
>Again, this seemed to work fine.  Then I did:
>
> >> imageplot(MA, layout=list(ngrid.r=8,ngrid.c=4,nspot.r=16,nspot.c=16))

This won't work because imageplot doesn't know which array you want to do 
the imageplot for or what statistic you want to plot. (There is no such 
thing as an imageplot of a whole series of arrays, at least no one has 
defined such a thing yet.) As explained in the online help, the first 
argument z to imageplot should be a numeric vector rather than an MAList 
object. Unfortunately there aren't any good examples which make this clear 
in the online documentation yet. For example, if you want to plot the 
log-ratios for the first array, simply use

imageplot(MA$M[,1], layout=list(ngrid.r=8,ngrid.c=4,nspot.r=16,nspot.c=16))

Gordon

>And got the error:
>
>Error in imageplot(MA, layout = list(ngrid.r = 8, ngrid.c = 4, nspot.r = 
>16,  :
>         Number of image spots does not agree with layout dimensions
>
>Now the layout I have given it is 8*4*16*16 = 8192
>
>So I checked:
>
> > length(MA$M)
>[1] 8192
> > length(MA$A)
>[1] 8192
>
>Hmmmmm.... so that is why I am confused, because the number of image spots 
>DOES agree with the layout dimensions....
>
>Any help would be greatly appreciated :-D
>
>Thanks
>Mick