A GatingSet object can be exported as a GatingML file or GatingML-based wsp flowJo workspace (version 10) so that they can be loaded into Cytobank or flowJo.

The GatingSet to be exported can be either parsed from Cytobank or flowJo or created by automated gating algorithms from openCtyo. Here we will demontrate the latter.

0.1 Automated gating

0.1.1 Load raw FCS files into GatingSet

library(ncdfFlow)
library(flowWorkspace)
library(CytoML)
dataDir <- system.file("extdata",package="flowWorkspaceData")
#load raw FCS
fs <- load_cytoset_from_fcs(file.path(dataDir,"CytoTrol_CytoTrol_1.fcs"))
gs <- GatingSet(fs)

0.1.2 Compensate and transform

#compensate
comp <- spillover(fs[[1]])[["SPILL"]]
chnls <- colnames(comp)
comp <- compensation(comp)
gs <- compensate(gs, comp)

#transform
trans <- flowjo_biexp_trans()
trans <- transformerList(chnls, trans)
gs <- transform(gs, trans)

Note that the compensation and transformation must be applied directly to GatingSet instead of flowSet/ncdfFlowSet so that these information will be stored in the GatingSet object and exported to gatingML eventually.

0.1.3 Load the gating template and run auto gating

library(openCyto)
#load the original template for tcell panel
tbl <- data.table::fread(system.file("extdata/gating_template/tcell.csv", package = "openCyto"))
#modify some paramters to fit the current data range
tbl[5, gating_args:= "gate_range = c(1e3, 3e3)"]
tbl[c(8,11), gating_args:= "gate_range = c(2e3, 3e3)"]
#write the new template to disc
gtFile <- tempfile()
write.csv(tbl, file = gtFile)
##reload the new template
gt <- gatingTemplate(gtFile, autostart = 1L)
#run the gating
gating(gt, gs)
#hide the gates that are not of interest
toggle.helperGates(gt, gs)
#visualize the gates
library(ggcyto)
autoplot(gs[[1]])