Hello everyone,

I am trying to remove margin events in my flow data and came across flowDensity. After removing margin events plotting with xyplots fails.

Is this known behaviour? 'boundaryFilter' from flowCore works without problems.

Best,

Ulrik

​library(flowDensity)
library(flowViz)

# Example from 'nmRemove' works
data_dir <- system.file("extdata", package = "flowDensity")
load(list.files(pattern = 'sampleFCS_2', data_dir, full = TRUE))
#Removing margin events of FSC-A and SSC-A channels
no.margin <- nmRemove(f2, c("FSC-A","SSC-A"),talk=TRUE)
plotDens(f2, c("FSC-A","SSC-A"))
# Scatter plot of FSC-A vs. SSC-A after removing margins
plotDens(no.margin, c("FSC-A","SSC-A"))

# Ploting the original flow.frame using xyplots works
xyplot(`FSC-A`~ `SSC-A`, data = f2)
# This fails
xyplot(`FSC-A`~ `SSC-A`, data = no.margin)

sessionInfo()

R version 3.2.3 (2015-12-10)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 15.10

locale:
[1] LC_CTYPE=de_DE.UTF-8       LC_NUMERIC=C               LC_TIME=de_DE.UTF-8        LC_COLLATE=de_DE.UTF-8     LC_MONETARY=de_DE.UTF-8  
[6] LC_MESSAGES=de_DE.UTF-8    LC_PAPER=de_DE.UTF-8       LC_NAME=C                  LC_ADDRESS=C               LC_TELEPHONE=C           
[11] LC_MEASUREMENT=de_DE.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] flowViz_1.34.1    lattice_0.20-33   flowCore_1.36.9   flowDensity_1.4.0

loaded via a namespace (and not attached):
[1] GEOmap_2.3-8        pcaPP_1.9-60        Rcpp_0.12.3         MBA_0.0-8           RColorBrewer_1.1-2  nloptr_1.0.4        DEoptimR_1.0-4    
[8] RSEIS_3.5-2         bitops_1.0-6        tools_3.2.3         lme4_1.1-11         nlme_3.1-125        mgcv_1.8-11         RPMG_2.2-1        
[15] Matrix_1.2-3        graph_1.48.0        parallel_3.2.3      SparseM_1.7         mvtnorm_1.0-5       spam_1.3-0          hexbin_1.27.1     
[22] cluster_2.0.3       IDPmisc_1.1.17      MatrixModels_0.4-1  gtools_3.5.0        caTools_1.17.1      fields_8.3-6        RFOC_3.3-3        
[29] maps_3.1.0          stats4_3.2.3        grid_3.2.3          nnet_7.3-12         robustbase_0.92-5   Biobase_2.30.0      rrcov_1.3-11      
[36] sp_1.2-2            latticeExtra_0.6-28 minqa_1.2.4         gdata_2.17.0        corpcor_1.6.8       car_2.1-1           splines_3.2.3     
[43] gplots_2.17.0       splancs_2.01-38     matrixStats_0.50.1  BiocGenerics_0.16.1 MASS_7.3-45         pbkrtest_0.4-6      Rwave_2.4         
[50] quantreg_5.21       KernSmooth_2.23-15 

It is because "nmRemove" function leaves 'NA' values in the flowFrame 'no.margin', which breaks the 'xyplot' that tries to detect and hide the 'marginal events'. You can disable this behavior by setting 'margin = FLASE',

xyplot(`FSC-A`~ `SSC-A`, data =no.margin, margin = F)

If I were you, I'd contact the author of 'flowDensity' to find out why they didn't remove these 'NA's instead.

I'd also like to mention that 'ggcyto' package is more robust on this. It is currently in Bioc Devel branch (http://bioconductor.org/packages/3.3/bioc/html/ggcyto.html) and will deprecate 'flowViz' eventually. 

If you don't mind to install R.3.3.0, here is how it will plot 'no.margin' without complaining 'NA'.

library(ggcyto)
autoplot(no.margin, x = "FSC-A", y = "SSC-A")