[BioC] error in Limma normalization with bc.method="normexp"

[Gordon Smyth]

>Date: Wed, 17 Nov 2004 15:32:39 -0800
>From: "Caimiao Wei" <caimiaow at u.washington.edu>
>Subject: [BioC] error in Limma normalization with bc.method="normexp"
>To: <bioconductor at stat.math.ethz.ch>
>
>  Dear all,
>
>I got this error while I was trying to normalize (first within arrays, then
>between arrays) Agilent arrays  using these code lines (R2.0.0, bioConductor
>1.5, Windows XP):
>
> > #########normalize within arrays
>
> > MA<-normalizeWithinArrays(RG, method="loess", bc.method="normexp",
> > offset=2)
>Corrected array 1
>Corrected array 2
>Corrected array 3
>Corrected array 4
>Corrected array 5
>Corrected array 6
>Corrected array 7
>Corrected array 8
>Corrected array 9
>Corrected array 10
>Corrected array 11
>Corrected array 12
>Corrected array 13
>Corrected array 14
>Corrected array 15
>Corrected array 16
>Corrected array 17
>Corrected array 18
>Corrected array 19
>Corrected array 20
>Corrected array 21
>Corrected array 22
>Corrected array 23
>Corrected array 24
>
> > #########normalize between arrays
>
> > MA.b=normalizeBetweenArrays(MA, method="quantile")
>Error in xy.coords(x, y) : x and y lengths differ
>
>I found that MA$A and MA$M contain a few columns with all "NA" values.

This is almost certainly an indication that the "normexp" algorithm has 
failed for those arrays. The "normexp" algorithm uses a Nelder-Mead 
derivative-free optimisation method. It works most of the time but not on 
all data sets. I do not have a fix at this stage. Sorry. You'll need to 
stick with bg.method="half".

In general you should check you have no missing values before using 
normalizeBetweenArrays() with method="quantile".

Gordon

>Thanks
>Caimiao