[BioC] flowViz Plots Gate Outline and grid.newpage()/viewport Issue

[Florian Hahne]

Aric Gregson wrote:
> On Mon, 13 Apr 2009 17:09:55 -0700
> Florian Hahne <fhahne at fhcrc.org> wrote:
>
>   
>> I don't quite follow why you have to draw the plot twice to get the
>> gate outline. This works on my installation:
>> library(flowViz)
>> data(GvHD)
>> tmp <- GvHD[[1]]
>> rg <- rectangleGate("FSC-H"=c(100, 200), "SSC-H"=c(400, 800))
>> xyplot(`SSC-H`~`FSC-H`, tmp, smooth=FALSE, filter=rg, outline=T)
>>     
>
> Perhaps this has been fixed in 1.7.11? Maybe something else is not
> working properly on my installation (quite possible). In previous
> discussions you had mentioned there were some bugs in the drawing of
> xyplot, so I did not mention this and assumed it was one that you were
> aware of. It has been extremely reproducible on my machine for many
> months.
>   
Hopefully! But if you could provide me with a reproducible example using 
the GvHD data I could make sure. Would be good to fix this before the 
release if it really is a bug.
>   
>> I am also not sure why you want to place individual plots manually. 
>> xyplot knows how to deal with flowSets, so the following works fine:
>> xyplot(`SSC-H`~`FSC-H`, GvHD[1:3], smooth=FALSE, filter=rg, outline=T)
>> and if you want to pass in separate filter objects for each frame you 
>> can do so by means of a list.
>>     
>
> I'm certain there is a better way, but the above is not what I am
> intending to accomplish. I'd like to plot the same flowFrame from a
> flowSet, but with different plots. That is, I'd like to show the gating
> strategy of the flowSet using one flowFrame as an example. I have not
> figured out how lattice can do that without making individual plots and
> then laying them all on a page. 
>   
I see. That is indeed not directly possible with flowViz or lattice 
unless you produce separate lattice objects for each plot and arrange 
them with split, or your self-defined viewports, which boils down to 
pretty much the same thing. I don't see a simple solution to build that 
into flowViz since we are very flowSet centric.
> For example, I have the following plots for flowSet[[6]]:
> 1. FSC.A vs. FSC.H (filterobject1) 
> 2. APC.H7 vs. Violet UV (filterobject2)
> 3. APC.H7 vs. AmCyan (filterobject3)
> 4. SSC.A vs. FSC.A (filterobject4)
> 5. PacBlue vs. Alexa467 (filterobject5)
> 6. densityplot of entire flowSet (filterobject6)
>
> Deepayan Sarkar's book is truly invaluable, but I find it abstract on
> some points. This is purely my fault as I am not intimately familiar
> with lattice, yet. The grid method I used was easiest for me to
> implement and familiar from the ggplot2 methods. I will have another
> look at the split method.
>
> Thanks for any further suggestions.
>
> Aric
>
>   
>> sessionInfo()
>>     
> R version 2.8.0 (2008-10-20) 
> i386-pc-solaris2.11 
>
> locale:
> C
>
> attached base packages:
>  [1] splines   grid      tools     stats     graphics  grDevices
> utils [8] datasets  methods   base     
>
> other attached packages:
>  [1] MASS_7.2-44        RColorBrewer_1.0-2 proto_0.3-8 [5]
> reshape_0.8.2      plyr_0.1.1 xtable_1.5-4       flowViz_1.7.10 [9]
> lattice_0.17-20 flowCore_1.9.18    rrcov_0.4-07       robustbase_0.4-3
> [13] Biobase_2.2.1     
>
> loaded via a namespace (and not attached):
>  [1] AnnotationDbi_1.4.1 DBI_0.2-4           KernSmooth_2.22-22 
>  [4] RSQLite_0.7-1       cluster_1.11.11     feature_1.1-14     
>  [7] geneplotter_1.20.0  graph_1.20.0        latticeExtra_0.5-4 
> [10] stats4_2.8.0       
>   


-- 
Florian Hahne, PhD
Computational Biology Program
Division of Public Health Sciences
Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N, M1-B514
PO Box 19024
Seattle, Washington 98109-1024
206-667-3148
fhahne at fhcrc.org