[BioC] Help with the zero.weights option of plotMA3by2 function in Limma package

Gordon Smyth smyth at wehi.EDU.AU
Thu May 4 02:02:59 CEST 2006


Dear Florian,

Thanks for pointing this out. It is a bug and not your mistake. I'll 
fix it for limma 2.7.2.

Best wishes
Gordon

At 08:00 PM 3/05/2006, bioconductor-request at stat.math.ethz.ch wrote:
>Date: Tue, 02 May 2006 13:53:57 -0400
>From: Florian Chain <florian.chain.1 at ulaval.ca>
>Subject: [BioC] Help with the zero.weights option of plotMA3by2
>         function in Limma package
>To: bioconductor at stat.math.ethz.ch
>
>Dear all,
>
>trying to get some insights in a data set resulting from a microarray
>experiment, I used the Limma package and its graphical function plotMA
>and plotMA3by2. I am wondering if there might be a problem with the
>zero.weights option.
>Here are the command lines I have tried, and their results
>  > plotMA(RGflag2[,1], zero.weights=TRUE)
>I got a graphical display, with all data plotted for the first array of
>my data set named RGflag2. RGflag2 includes weights (2 possible values:
>0 and 1).
>
>  > plotMA(RGflag2[,1], zero.weights=FALSE)
>When I use this one, I get a plot but some data are missing ( not
>displayed), which is correct, given the option zero.weights=FALSE
>
>However, when I tried to use plotMA3by2
>  > plotMA3by2(RGflag2, prefix="zero weights included",  zero.weights=TRUE)
>  > plotMA3by2(RGflag2, prefix="zero weights excluded", zero.weights=FALSE)
>
>the images I obtain are identical and correspond to plots without data
>having a 0 as weight value, whatever the setting (TRUE or FALSE).
>In other words, individually the zero.weights option gives the expected
>results, but when dealing with a batch of plots with the plotMA3by2 it
>is like it stays by default on the zero.weights=FALSE.
>Could anyone know if I am doing something wrong or misunderstanding the
>function zero.weights in the plotMA3by2 ?
>
>Thank you,
>Florian
>
>PS: thanks to the creator of R and the packages, by the way.



More information about the Bioconductor mailing list