Visualize GatingSet with ggcyto

knitr::opts_chunk$set(message = FALSE, warning = FALSE, error = TRUE)
dataDir <- system.file("extdata",package="flowWorkspaceData")
gs <- load_gs(list.files(dataDir, pattern = "gs_manual",full = TRUE))

By specifying the dimensions through aes and selecting the cell population through subset, ggcyto can easily visualize the gated data stored in GatingSet.

p <- ggcyto(gs, aes(x = CD4, y = CD8), subset = "CD3+") 
# 2d plot 
p <- p + geom_hex(bins = 64)


We can use the instrument range to automatically filter out these outlier cell events

p + ggcyto_par_set(limits = "instrument")

Or by setting limits manually

myPars <- ggcyto_par_set(limits = list(x = c(0,3.5e3), y = c(-10, 4.1e3)))
p <- p + myPars# or xlim(0,3.5e3) + ylim(-10, 4e3) 

To check what kind of visualization parameters can be changed through ggcyto_par_set, simply print the default settings



To plot a gate, simply pass the gate name to the geom_gate layer

p + geom_gate("CD4")

More than one gate can be added as long as they share the same parent and dimensions

p <- p + geom_gate(c("CD4","CD8")) # short for geom_gate("CD8") + geom_gate("CD4")


By default, stats for all gate layers are added through empty geom_stats layer.

p + geom_stats() + labs_cyto("marker")

Note that we choose to only display marker on axis through labs_cyto layer here.

To add stats just for one specific gate, we can pass the gate name to geom_gate

p + geom_stats("CD4")

stats type, background color and position are all adjustable.

p + geom_stats("CD4", type = "count", size = 6,  color = "white", fill = "black", adjust = 0.3)

When 'subset' is not specified, it is at abstract status thus can not be visualized

p <- ggcyto(gs, aes(x = CD4, y = CD8)) + geom_hex() + myPars

unless it is instantiated by the gate layer, i.e. lookup the gating tree for the parent node based on the given gates in geom_gate

p <- p + geom_gate(c("CD4", "CD8"))


With geom_overlay, you can easily overlay the additional cell populations (whose gates are not defined in the current projection) on top of the existing plot.

p + geom_overlay("CD8/CCR7- 45RA+", col = "black", size = 0.1, alpha = 0.4)

geom_overlay automatically determines the overlay type (goem_point or geom_density) based on the number of dimensions specified in ggcyto constructor. Note that we change the default y axis from density to count in order to make the scales comparable for the stacked density layers. They are wrapped with .. because they belong to the computed variables.

p <- ggcyto(gs, aes(x = CD4), subset = "CD3+") + geom_density(aes(y = ..count..))
p + geom_overlay("CD8/CCR7- 45RA+", aes(y = ..count..), fill = "red")


Alternatively, we can choose to plot all children of one specified parent and projections

p <- ggcyto(gs, aes(x = 38, y = DR), subset = "CD4") + geom_hex(bins = 64) + geom_gate() + geom_stats()

Or we can add gate layer to any arbitary node instead of its parent node

ggcyto(gs, subset = "root", aes(x = CD4, y = CD8)) + geom_hex(bins = 64) + geom_gate("CD4") + myPars


Sometime it is helpful to display the axis label in raw scale by inverse transforming the axis without affecting the data

p + axis_x_inverse_trans() + axis_y_inverse_trans()
#add filter (consistent with `margin` behavior in flowViz)
# ggcyto(gs, aes(x = CD4, y = CD8), subset = "CD3+", filter = marginalFilter)  + geom_hex(bins = 32, na.rm = T)

Try the ggcyto package in your browser

Any scripts or data that you put into this service are public.

ggcyto documentation built on Nov. 8, 2020, 5:30 p.m.