[BioC] affylmGUI

[James Wettenhall]

Simon,

As we are leading up to a Release version of 1.4, I'll just 
explain what needs to be fixed in affylmGUI and check with Jeff 
that's it OK to fix these things, rather than rushing in and 
changing things...

> We have affylmGUI running on Mac OSX and R2 dev. A couple of things 

Great!

> seem not to work. We can save all the plots except for venn diagrams 
> Object "limmaDataSetNameText" not found

This bug is easy to fix.  Should be able to insert this patch 
for the 1.4 release.  The following line needs to be inserted at 
the beginning of the VennDiagramPlot function in affylmGUI's 
plot.R :

Try(limmaDataSetNameText <- 
  get("limmaDataSetNameText",envir=affylmGUIenvironment))

[Don't get too confused by the Try(...). That's just 
affylmGUI's way of diverting error messages to messageBoxes]

> Error in try(expr, TRUE): No loop to break from, jumping to top level

This appears to be an incorrect use of break() in the 
HeatDiagramPlot function in affylmGUI's plot.R.  The incorrect 
code is :
Try(if (is.na(matchIndex)) break())
[ More code within large if-block ]

It should be:
Try(if (!is.na(matchIndex))
{
  [ More code within large if-block ]
})

> Cannot find package drosgenome1

At the moment, some affylmGUI functions will complain 
if there is no annotation package available for your chip.  
Usually this is just a harmless but annoying message box, on 
which you can click "OK".  You may have to put up with this 
annoying behaviour for the 1.4 Release, but there are some 
related bugs which you encountered around the same time as this
annoying "warning".

The following lines of code in the HeatDiagramPlot 
function in affylmGUI's plot.R will fail if there is
no annotation package, i.e. they are incorrect:
  Try(genelist <- 
    cbind(as.matrix(ls(cdfenv)),as.matrix(geneSymbols)))
...
   Try(heatdiagram(abs(eb$t),fit$coefficients,primary=1,
      critical.primary=primaryCutoff,critical.other=otherCutoff,
      names=substr(geneSymbols,1,20),main=plotTitle))

The first line of code above can be deleted ("genelist" is not 
used in PlotHeatDiagram) and replaced with:
Try(if (length(geneSymbols)>0)
  geneLabels <- geneSymbols
else
  geneLabels <- ls(cdfenv))

and the second can be modified by replacing "geneSymbols" by 
"geneLabels" :
   Try(heatdiagram(abs(eb$t),fit$coefficients,primary=1,
      critical.primary=primaryCutoff,critical.other=otherCutoff,
      names=substr(geneLabels,1,20),main=plotTitle))

> What P value is used to define significant in the venn diagram?

OK, this probably should become part of a dialog box, but I 
don't think it will happen before the 1.4 Release.  The venn 
diagram plotting function calls limma's classifyTestsF which
seems to have a default p-value of 0.01 :
library(limma)
?classifyTestsF

You might get more discussion about that by posting with a 
different subject-line like "limma classifyTestsF p-value"

Regards,
James


--------------------------------------------------------------------------
James Wettenhall                                  Tel: (+61 3) 9345 2629
Division of Genetics and Bioinformatics           Fax: (+61 3) 9347 0852
The Walter & Eliza Hall Institute         E-mail: wettenhall at wehi.edu.au
 of Medical Research,                     Mobile: (+61 / 0 ) 438 527 921    
1G Royal Parade,
Parkville, Vic 3050, Australia
http://www.wehi.edu.au