[BioC] Odd Plotting Behavior with lymphGate flowStats

Nishant Gopalakrishnan ngopalak at fhcrc.org
Tue Aug 3 00:23:56 CEST 2010


Hi Aric,

What you have noticed appears to be a general problem with plotting of
gates with a fill color
in flowViz rather than an error in the calculation of results by the
lymphGate function.

data(GvHD)
dat <- GvHD[1:2]

rgate1 <- rectangleGate("FSC-H"=c(0, 400),"SSC-H"=c(-50, 300))
## works fine
xyplot(`SSC-H` ~ `FSC-H` | Visit, data=dat, filter=rgate1)gate


par.settings=list(gate=list(fill="red", alpha = 0.3))
xyplot(`SSC-H` ~ `FSC-H` | Visit, data=dat, filter=rgate1,
               par.settings = par.settings )
rgate2 <- !rgate1
xyplot(`SSC-H` ~ `FSC-H` | Visit, data=dat, filter= rgate2,
               par.settings = par.settings )

The example above illustrates the same issue where the area outside the
rectangle is not
correctly filled  with color and perhaps needs to be fixed.

Nishant

On 07/30/2010 02:23 PM, Aric Gregson wrote:
> There may be a problem with the lymphGate plot function and potentially
> the entire function. While generating the lymphGate, the plot=TRUE
> option shows the correct population is selected. However, when using
> the non-Data() plotting technique after to show the populations (as
> demonstrated in '...A typical workflow'), the wrong population is shown.
> Here is an example:
>
> lymphfilter <- lymphGate(Data(wf[["biexponential"]]),
>                          channels=c("FSC.A", "SSC.A"),
>                          preselection="APC.H7.A",
>                          filterId="Lymphs",
>                          plot=TRUE,     <- CORRECT POPULATION SHOWN
>                          scale=1.8)  
> add(wf, lymphfilter$n2gate, parent="biexponential")
>
> Then:
> xyplot(`FSC.A` ~ `SSC.A` | Filename, wf[['Lymphs+']]) 
>
> Shows the wrong population. Yet, with the 'usual' Data() plotting
> method the correct population is shown as within the gate:
>
> xyplot(`FSC.A` ~ `SSC.A` | Filename, Data(wf[['biexponential']]),
>        nbin=c(400,400),
>        filter=lymphfilter$n2gate)
>
> Interestingly, changing the above plot to smooth=FALSE results in the
> wrong population being highlighted once again:
>
> xyplot(`FSC.A` ~ `SSC.A` | Filename, Data(wf[['biexponential']]),
>        smooth=FALSE,
>        filter=lymphfilter$n2gate)
>
> However, I have question that the correct population is gated. Based
> upon the medians of each population it seems that the wrong population
> may have been chosen, I'm not sure how to really know for sure.
>
> median(exprs(Data(wf[["Lymphs-"]])[[2,"FSC.A"]]))
> [1] 59620.83
>   
>> median(exprs(Data(wf[["Lymphs+"]])[[2,"FSC.A"]]))
>>     
> [1] 35668.27
>   
>> median(exprs(Data(wf[["Lymphs+"]])[[1,"FSC.A"]]))
>>     
> [1] 34982.41
>   
>> median(exprs(Data(wf[["Lymphs-"]])[[1,"FSC.A"]]))
>>     
> [1] 62609.93
>
> Anyone else seeing this?
>
> Thanks, Aric
>
> sessionInfo()
> R version 2.11.0 (2010-04-22) 
> amd64-portbld-freebsd8.1 
>
> locale:
> [1] C
>
> attached base packages:
>  [1] grid      tools     splines   stats     graphics  grDevices
> utils [8] datasets  methods   base     
>
> other attached packages:
>  [1] Rgraphviz_1.26.0    graph_1.26.0        flowQ_1.9.0        
>  [4] latticeExtra_0.6-11 RColorBrewer_1.0-2  parody_1.6.0       
>  [7] bioDist_1.20.0      KernSmooth_2.23-3   outliers_0.13-2    
> [10] flowStats_1.7.5     cluster_1.12.3      mvoutlier_1.4      
> [13] fda_2.2.2           zoo_1.6-4           flowViz_1.13.1     
> [16] lattice_0.18-5      flowCore_1.15.2     rrcov_1.0-01       
> [19] pcaPP_1.8-2         mvtnorm_0.9-92      robustbase_0.5-0-1 
> [22] Biobase_2.8.0       fortunes_1.3-7     
>
> loaded via a namespace (and not attached):
>  [1] AnnotationDbi_1.10.2 DBI_0.2-5            MASS_7.3-5          
>  [4] RSQLite_0.9-2        annotate_1.26.1      feature_1.2.4       
>  [7] geneplotter_1.26.0   ks_1.6.13            stats4_2.11.0       
> [10] tcltk_2.11.0         xtable_1.5-6
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor
>



More information about the Bioconductor mailing list