[BioC] simpleaffy question

[James W. MacDonald]

J.delasHeras at ed.ac.uk wrote:
> Quoting "James W. MacDonald" <jmacdon at med.umich.edu>:
> 
>> J.delasHeras at ed.ac.uk wrote:
>>
>>> Quoting "D.Enrique ESCOBAR ESPINOZA" <escobarebio at yahoo.com>:
>>>
>>>
>>>> HI ,
>>>> does someone know if there is a built-inn
>>>> function for adding a legend to the QC grph obtained
>>>> with :
>>>> data(Dilution)
>>>> plot(qc(Dilution))
>>>> or if there is an easy way to add a legend to this graph
>>>> thanks.
>>>
>>>
>>>
>>> In general I find that one or some of these functions help me   
>>> annotate any plot in R:
>>>
>>> ?text
>>> ?mtext
>>> ?title
>>> ?legend
>>>
>>> I presume you probably were looking for 'legend'.
>>
>>
>> Unfortunately legend won't work in this case. The qc() plot is produced
>> after a call to layout(), which cuts the plotting device into four
>> sections. The last section to be plotted is the upper left hand, which
>> contains the names of the QC probesets. If you make a call to legend(),
>> it will put the legend in that very small section, most likely right on
>> top of the names of the QC probesets.
>>
>> To see what I mean, try this code (which comes directly from this
>> plotting function):
>>
>> m <- matrix(c(4, 2, 1, 3), nrow = 2, ncol = 2)
>> layout(m, c(1, 2), c(0.1, 1))
>> layout.show(4)
>>
>> The plots will be added to each box in the order listed.
>>
>> That said, if you want to hack the code it wouldn't be that difficult
>> to  add a call to legend() right after the second call to plot() in the
>> function plot.qc.stats(). This will put a legend in the second layout
>> box, which currently only contains the chip names on the extreme right
>> edge. If you placed the legend "upperleft" and made the font size small
>> enough, it might fit.
>>
>> Luckily, since simpleaffy doesn't use a namespace, you can simply
>> copy-paste the function into an editor, hack away, then source() back
>> in and go.
>>
>> HTH,
>>
>> Jim
> 
> 
> 
> whops!
> 
> I wasn't familiar with the qc() plots, so I assumed they were straight  
> plots. Sorry!
> 
> Your idea to just get into the qc() code and add a call to legend as  
> required seems good and simple enough.
> 
> Is there a function in R that allows you to plot polygons, circles,  
> etc... *anywhere*? Something similar to 'mtext', only not just for text?
> I often multiplot figures, using 'par(mfrow=c(x,y))' among other  
> suitable parameters to fit several plots consecutively into a larger  
> figure. Then I use 'mtext' to place text information anywhere in the  
> final multiplot figure. It works well enough. If there is an  equivalent 
> funtion that you can sue to draw lines, rectangles etc with  in teh same 
> fashion, that would work too.

The only functions I know for plotting shapes require (x,y) coordinates, 
so won't work like mtext(). However, one might be able to come up with a 
reasonable looking legend just using mtext().

Best,

Jim


> 
> Jose


-- 
James W. MacDonald
University of Michigan
Affymetrix and cDNA Microarray Core
1500 E Medical Center Drive
Ann Arbor MI 48109
734-647-5623



**********************************************************
Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues.