[BioC] normalizing incomplete arrays [was: limma error message]

[Gordon Smyth]

As you've deduced, the problem is that your arrays do not a constant block 
size. limma assumes a constant block size for normalizeWithinArrays() and 
imageplot().

Here is a work-around for GenePix data:

for (b in 1:48) {
         i <- RG$genes$Block==b
         MA2 <- normalizeWithinArrays(RG[i,],method="loess")
         if(b==1)
                 MA <- MA2
         else
                 MA <- rbind(MA,MA2)
}

This will perform print-tip-loess normalization even though there are 
irregular numbers of spots in the print-tip groups.

Gordon

>Date: Fri, 29 Jul 2005 16:42:58 -0400
>From: Naomi Altman <naomi at stat.psu.edu>
>Subject: Re: [BioC] limma error message
>To: Jianping Jin <jjin at email.unc.edu>
>Cc: Bioconductor at stat.math.ethz.ch
>Message-ID: <6.2.1.2.2.20050729163707.040c3240 at imap.stat.psu.edu>
>Content-Type: text/plain; charset="us-ascii"; format=flowed
>
>28*28*12*4
>[1] 37632
>
>But you have 34970  genes.
>
>The simplest thing to do is to add the missing lines to your GPR files.
>
>You could probably do this in your favorite text manipulation language but
>with 8 arrays, you might be able to do it manually more quickly.
>
>--Naomi
>
>
>At 12:13 PM 7/29/2005, Jianping Jin wrote:
> >Dear Naomi and BioConductor list,
> >
> >Thanks for your kind reply! But I got the same error message after re-run
> >the program as the following:
> >
> >>targets <- readTargets()
> >>RG <- read.maimages(targets$FileName, source="genepix")
> >Read 0341_3289_A5_052_HO34K_sample-1_HRef_scan2.gpr
> >Read 0341_3290_A5_053_HO34K_sample-2_HRef_scan2.gpr
> >Read 0341_3291_A5_054_HO34K_sample-3_HRef.gpr
> >Read 0341_3292_A5_055_HO34K_sample-4_HRef.gpr
> >Read 0341_3293_A5_056_HO34K_sample-5_HRef.gpr
> >Read 0341_3294_A5_057_HO34K_sample-6_HRef.gpr
> >Read 0341_3295_A5_058_HO34K_sample-7_HRef.gpr
> >Read 0341_3296_A5_059_HO34K_sample-8_HRef.gpr
> >>summary(RG)
> >        Length Class      Mode
> >R       279760 -none-     numeric
> >G       279760 -none-     numeric
> >Rb      279760 -none-     numeric
> >Gb      279760 -none-     numeric
> >targets      1 data.frame list
> >genes        5 data.frame list
> >
> >>dim(RG)
> >[1] 34970     8
> >
> >>getLayout(RG$genes)
> >$ngrid.r
> >[1] 12
> >
> >$ngrid.c
> >[1] 4
> >
> >$nspot.r
> >[1] 28
> >
> >$nspot.c
> >[1] 28
> >
> >attr(,"class")
> >[1] "PrintLayout"
> >
> >>MA <- normalizeWithinArrays(RG, layout=list(ngrid.r=12, ngrid.c=4,
> >nspot.r=28, nspot.c=28))
> >Error: subscript out of bounds
> >
> >I checked the data files and found nsopt.r and nspot.c in the last one
> >third blocks of every files were changed from 28 to 24 and 23
> >respectively. That may be the reason for the error message. Do you have
> >any ways to handle this?
> >
> >Many thanks!
> >
> >JP-
> >
> >xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> >x Jianping Jin Ph.D.            x
> >x Bioinformatics scientist      x
> >x Center for bioinformatics     x
> >x 3133 Bioinformatics Building  x
> >x CB# 7104                      x
> >x University of North Carolina  x
> >x Chapel Hill, NC 27599         x
> >x Tel: (919)843-6105            x
> >x Fax: (919)843-3103            x
> >x E-mail: jjin at email.unc.edu    x
> >xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> >
> >--On Thursday, July 28, 2005 5:17 PM -0400 Naomi Altman
> ><naomi at stat.psu.edu> wrote:
> >
> >>layout=list(ngrid.r= ,ngrid.c= ,nspot.r= , nspot.c= )
> >>
> >>Your array is laid out in blocks according to the print tip configuration
> >>of the printer.
> >>nspot.r is the number of rows in each block
> >>nspot.c is the number of columns in each block
> >>
> >>ngrid.r is the number of rows of blocks
> >>ngrid.c is the number of columns of blocks
> >>
> >>e.g. a typical 8 tip design might have 4 rows and 2 columns of blocks
> >>
> >>If you don't know the right numbers to put in, look at the raw data - the
> >>block number and row and column are output to your expression file.
> >>
> >>--Naomi
> >>
> >>At 04:49 PM 7/28/2005, you wrote:
> >>>Thanks Naomi! Could you explain how to specify parameter "layout"?
> >>>JP-
> >>>
> >>>--On Thursday, July 28, 2005 3:35 PM -0400 Naomi Altman
> >>><naomi at stat.psu.edu> wrote:
> >>>
> >>>>I got this message when I used the wrong layout file.
> >>>>
> >>>>--Naomi
> >>>>
> >>>>At 03:03 PM 7/28/2005, Jianping Jin wrote:
> >>>>>Dear BioConductor list,
> >>>>>
> >>>>>I got an error, Error: subscipt out of bounds, when I started to run a
> >>>>>limma program normalizeWithinArrays(). What I wanted to do was to
> >>>>>normalize eight cDNA slides. >dim(RG): 34970      8.
> >>>>>
> >>>>>I used >debug(normalizeWithinArrays) and then
> >>>>> >normalizeWithinArrays(RG) to debug by either Browse[1]> n or c as
> >>>>>well. But when I retried the program it gave me the same error message
> >>>>>above.
> >>>>>
> >>>>>I will highly appreciate it if any one could give it shot.
> >>>>>
> >>>>>
> >>>>>Best regards!
> >>>>>
> >>>>>JP-
> >>>>>
> >>>>>xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> >>>>>x Jianping Jin Ph.D.            x
> >>>>>x Bioinformatics scientist      x
> >>>>>x Center for bioinformatics     x
> >>>>>x 3133 Bioinformatics Building  x
> >>>>>x CB# 7104                      x
> >>>>>x University of North Carolina  x
> >>>>>x Chapel Hill, NC 27599         x
> >>>>>x Tel: (919)843-6105            x
> >>>>>x Fax: (919)843-3103            x
> >>>>>x E-mail: jjin at email.unc.edu    x
> >>>>>xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> >>>>>
> >>>>
> >>>>Naomi S. Altman                                814-865-3791 (voice)
> >>>>Associate Professor
> >>>>Bioinformatics Consulting Center
> >>>>Dept. of Statistics                              814-863-7114 (fax)
> >>>>Penn State University                         814-865-1348 (Statistics)
> >>>>University Park, PA 16802-2111