Tutorial: Running the pipeline

Jasme Lee

Introduction

staRgate is an automated gating pipeline to process and analyze flow cytometry data to characterize the lineage, differentiation and functional states of T-cells.

This pipeline is designed to mimic the manual gating of defining flow biomarker positive populations relative to a unimodal background population to include cells with varying intensities of marker expression. This is achieved via estimating the kernel density of the intensity distribution and corresponding derivatives. This pipeline integrates the density gating method in conjuction with some pre-processing steps achieved via the R package openCyto and an optional step with flowAI. The flow data is stored within R as a GatingSet object, which makes it easily transferable to other flow cytometry workflows available on BioConductor.

This vignette will walk through how to run the {staRgate} pipeline starting from importing an flow cytometry standard (FCS) file into R to preprocessing and gating, as well as identifying T-cell subpopulations for downstream analysis. This pipeline returns results at the single-cell level as well as the summarized sample-level data (for the percentages of T-cell cells when identifying the subpopulations).

For illustration purposes, the example FCS file used in this vignette and stored in the package data is a concatenated file limited to the first 30k events/cells acquired to reduce the run time and file size.

After running {staRgate} to gate flow cytometry data, it is recommended to perform some quality checks (QC) on the gate placements to ensure they are reasonable. We suggest to use ridgeplots in addition to the ggcyto::autoplot to visualize the density distributions per marker across samples. When examining a large batch of samples, downsampling, such as to a random sample of 10k CD3+ cells, will make the QC process more manageable. In addition, random spot checks of a few samples would also be helpful QC to detect any edge cases.

Currently in this tutorial, we do not extend to the QC steps and suggest to lean on other examples for how to put together a ridgeplot for example. In the near future, we hope to incorporate some examples for the additional QC steps as well, stay tuned!

Installation

The {staRgate} package relies on a few Biocondcutor R packages. Before installing {staRgate}, first setup Bioconductor and install all dependencies.

Below instructions are instructions for installing Bioconductor

if (!require("BiocManager", quietly = TRUE)) {
  install.packages("BiocManager")
}

The essential dependencies for running the {staRgate} package include: flowCore, and flowWorkspace.

The following packages are needed to fully run the pipeline as shown in this Tutorial, but not required for the {staRgate} code chunks to run: openCyto, flowAI, ggplot2, ggcyto, gt

# Load libraries
library(staRgate)
library(openCyto)
library(flowWorkspace)
#> As part of improvements to flowWorkspace, some behavior of
#> GatingSet objects has changed. For details, please read the section
#> titled "The cytoframe and cytoset classes" in the package vignette:
#> 
#>   vignette("flowWorkspace-Introduction", "flowWorkspace")
library(flowCore)
# Just for plotting in the vignette
library(ggplot2)

# Set up dynamic variables
pt_samp_nm <- "flow_sample_1"

## File path to the FCS file
path_fcs <- system.file("extdata", "example_fcs.fcs", package = "staRgate", mustWork = TRUE)

## File path to the compensation matrix csv file
## Expect format to match flowJo exported version
path_comp_mat <- system.file("extdata", "comp_mat_example_fcs.csv", package = "staRgate", mustWork = TRUE)

## File path for outputs/saving
# Maybe not the best sol, but create a temp dir?
path_out <- tempdir()
# Print the path_out for user to see
path_out
#> [1] "/tmp/RtmpR7t2EM"

## File path Gating template
gtFile <- system.file("extdata", "gating_template_x50_tcell.csv", package = "staRgate", mustWork = TRUE)

## File path to biexp parameters
## Expects 4 columns: full_name, ext_neg_dec, width_basis, positive_dec
## full name should contain the channel/dye name
# 3 remaining cols fill in with desired parameter values
path_biexp_params <- system.file("extdata", "biexp_transf_parameters_x50.csv", package = "staRgate", mustWork = TRUE)

## File path to positive peak thresholds
path_pos_peak_thresholds <- system.file("extdata", "pos_peak_thresholds.csv", package = "staRgate", mustWork = TRUE)

Input files and function parameters

In order to run the pipeline, the user must have the data in flow cytometry standard (FCS) format. This is usually the output from flowJo. All of these input files are expected to be comma-separated values (csv) files. Please see the inst folder of the package for examples of the formats.

• Compensation matrix from manual gating-
  • Matrix where the column and row names correspond to the channel names, cell values correspond to the spillover correction to be applied.
  • This can be exported as a csv in flowJo’s options
• Biexponential transformation parameters-
  • This is a table specifying the parameters (negative decades, width basis and positive decades) to be applied to the listed channels
• Gating template
  • A gating template is required to run the pre-gating via openCyto
  • The package includes a gating template tailored for gating this panel of T-cell markers.
  • For examples of how to modify the gating template, please refer to the openCyto documentation.
• Bin size
  • From a systematic grid search, we have found bin sizes of 40 to 50 works well for the density gating

A few things to keep in mind when debugging/iterating through gating:

• If saving at the same path with same name (i.e., rerunning the same code), the GatingSet folder from the flowWorkspace::save_gs() command needs to be deleted for openCyto to save again, otherwise, will encounter an error related to an invalid path from the flowWorkspace::save_gs() function

Example

Below is an example of gating 1 FCS sample.

Optional: include a samp_metadata for saving out some properties of the FCS file and from the processing steps (e.g., the compensation matrix, the number of cells in the root population)

# Optional to save the samp_metadata
samp_metadata <- list()

Import FCS

# Read in gating template
dtTemplate <- data.table::fread(gtFile)

# Load the FCS
gt_tcell <- openCyto::gatingTemplate(gtFile)
#> expanding pop: -/++/-
#> Adding population:fsc_ssc_qc
#> Adding population:nonDebris
#> Adding population:singlets
#> Adding population:cd14-cd19-
#> Adding population:live
#> Adding population:cd3
#> Adding population:cd4+
#> Adding population:cd8+
#> Adding population:cd4+cd8+
#> Adding population:cd4-cd8+
#> Adding population:cd4+cd8-
#> Adding population:cd4-cd8-

cs <- flowWorkspace::load_cytoset_from_fcs(path_fcs)

# **Optional** -
# For consistency among different naming conventions for the live/dead
# change "Viability" or "L/D" "LD" or "live" to = "L_D" for consistency on gtemplates etc
{
  if (any(c("Viability", "L/D", "L_D", "live", "Live") %in% flowWorkspace::markernames(cs))) {
    aux_log <- flowWorkspace::markernames(cs) == "Viability" | flowWorkspace::markernames(cs) == "L/D" | flowWorkspace::markernames(cs) == "L_D"

    flowWorkspace::markernames(cs) <- replace(markernames(cs), aux_log, "LD")
  }
}

# Create a GatingSet of 1 sample
gs <- flowWorkspace::GatingSet(cs)

# Check- how many cells is in the FCS file?
n_root <- flowWorkspace::gh_pop_get_count(gs, "root")

n_root
#> [1] 30000

# Add n_root to samp_metadata
samp_metadata$n_root <- n_root

Compensation

# Check no comp applied
chk_cm <- flowWorkspace::gh_get_compensations(gs)

chk_cm
#> NULL

# Apply comp
gs <- getCompGS(gs, path_comp_mat = path_comp_mat)

# Can check that the comp was applied
chk_cm <- flowWorkspace::gh_get_compensations(gs)

# Not aware of an accessor that we can use for this
head(methods::slot(chk_cm, "spillover"))
#>                 AF700-A       APC-A   APC-f750-A      BB515-A    BB660-A
#> AF700-A     1.00000e+00 2.48301e-02  2.49284e-01  0.009679220 0.00268438
#> APC-A       1.32334e-01 1.00000e+00  3.37338e-02 -0.000160633 0.01598550
#> APC-f750-A  8.68539e-02 1.38732e-01  1.00000e+00 -0.001767070 0.00222729
#> BB515-A    -1.15670e-06 8.78380e-06 -3.46330e-06  1.000000000 0.00737752
#> BB660-A     8.87227e-02 5.93124e-01  2.00381e-02  0.014618200 1.00000000
#> BB700-A     1.56721e-01 1.03493e-01  5.16101e-02  0.021794700 0.23758800
#>               BB700-A     BB790-A    BUV395-A     BUV496-A     BUV563-A
#> AF700-A    0.04732480 0.017307700 0.001813870  1.81722e-03  1.45041e-03
#> APC-A      0.00443383 0.001225970 0.000471851 -5.18352e-05 -3.25396e-05
#> APC-f750-A 0.00135711 0.052727800 0.007016350 -3.55365e-04 -1.00153e-03
#> BB515-A    0.00160540 0.000259797 0.000488654  9.92731e-03  2.45334e-03
#> BB660-A    0.29925900 0.072568700 0.000750601  4.69872e-04  5.91621e-03
#> BB700-A    1.00000000 0.305046000 0.001371840  5.33863e-04  7.68645e-03
#>               BUV615-A    BUV661-A     BUV737-A     BUV805-A      BV421-A
#> AF700-A    0.002494550 0.002203370  1.38716e-01  0.033104200  5.11994e-05
#> APC-A      0.001271670 0.114363000  2.21613e-02  0.005274360 -4.09868e-04
#> APC-f750-A 0.000055023 0.013948700  3.17988e-02  0.249248000 -7.88098e-03
#> BB515-A    0.001525590 0.000265054 -1.52650e-06 -0.000011894 -3.13395e-04
#> BB660-A    0.013022400 0.199251000  3.28067e-02  0.008413150 -8.18947e-04
#> BB700-A    0.012374800 0.042845800  1.02680e-01  0.034781800 -1.40587e-03
#>                BV480-A     BV510-A      BV570-A     BV605-A     BV650-A
#> AF700-A    0.003390200 2.60901e-03  2.45899e-03 0.003585150 0.004457680
#> APC-A      0.000113668 4.37466e-05  5.85607e-05 0.000850790 0.138441000
#> APC-f750-A 0.003118890 1.52763e-03 -9.31873e-04 0.000173277 0.016350600
#> BB515-A    0.000190708 3.34812e-03  8.84301e-04 0.000579157 0.000126033
#> BB660-A    0.000388549 2.12014e-03  2.71519e-02 0.049095400 0.692174000
#> BB700-A    0.000586408 3.01256e-03  3.65599e-02 0.059112300 0.173298000
#>                BV711-A      BV750-A      BV786-A         PE-A   PE-CF594-A
#> AF700-A    2.12730e-01  8.48399e-02  4.39476e-02  0.006840740  2.92636e-02
#> APC-A      2.85134e-02  9.73321e-03  4.17643e-03  0.000280757  5.41964e-03
#> APC-f750-A 5.86523e-03  6.38759e-02  2.26547e-01 -0.002460560  7.43340e-03
#> BB515-A    8.56130e-06 -1.02317e-05 -1.94972e-05 -0.000126331 -7.83382e-05
#> BB660-A    1.61692e-01  4.23041e-02  2.02054e-02  0.000124760  1.78335e-03
#> BB700-A    5.56872e-01  1.49641e-01  8.27345e-02  0.000280215  7.49651e-04
#>                PE-Cy5-A   PE-Cy5.5-A     PE-Cy7-A
#> AF700-A     1.64829e-02  3.26051e-01  1.03233e-01
#> APC-A       2.67747e-01  8.10869e-02  2.15633e-02
#> APC-f750-A  4.93700e-02  2.25955e-02  6.09145e-01
#> BB515-A    -6.68526e-05 -3.01727e-05 -1.21999e-05
#> BB660-A     7.25687e-02  2.16664e-02  5.19117e-03
#> BB700-A     2.13325e-02  6.47322e-02  1.92574e-02

# **Optional**- save the comp mat matrix
# Can grab with the @spillover from the compensation object
samp_metadata$comp_mat <-
  methods::slot(chk_cm, "spillover")

Transformation

The transformation applied to all channels is the same: biexponential with extra negative decades = 0.5, positive decades = 4.5 and width basis = -30

The structure of the table of parameters (as .csv format) should be first column for the flurochrome names corresponding to the panel, followed by the parameters.

tbl_biexp_params <-
  utils::read.csv(path_biexp_params) |>
  janitor::clean_names(case = "all_caps")

# **Optional**-- saving out the table of parameters as format that's easier to read
# Below there is an option to save it out as the transformerList which is what
# getBiexpTransformGS() relies on to supply to flowWorkspace::transform
samp_metadata$biexp_params <- tbl_biexp_params

# The biexp table
gt::gt(tbl_biexp_params)

FULL_NAME	EXT_NEG_DEC	WIDTH_BASIS	POSITIVE_DEC
BUV395-A	0.5	-30	4.5
BUV496-A	0.5	-30	4.5
BUV563-A	0.5	-30	4.5
BUV615-A	0.5	-30	4.5
BUV661-A	0.5	-30	4.5
BUV737-A	0.5	-30	4.5
BUV805-A	0.5	-30	4.5
BV421-A	0.5	-30	4.5
BV480-A	0.5	-30	4.5
BV510-A	0.5	-30	4.5
BV570-A	0.5	-30	4.5
BV605-A	0.5	-30	4.5
BV650-A	0.5	-30	4.5
BV711-A	0.5	-30	4.5
BV750-A	0.5	-30	4.5
BV786-A	0.5	-30	4.5
BB515-A	0.5	-30	4.5
BB660-A	0.5	-30	4.5
BB700-A	0.5	-30	4.5
BB790-A	0.5	-30	4.5
PE-A	0.5	-30	4.5
PE-CF594-A	0.5	-30	4.5
PE-Cy5-A	0.5	-30	4.5
PE-Cy5.5-A	0.5	-30	4.5
PE-Cy7-A	0.5	-30	4.5
APC-A	0.5	-30	4.5
AF700-A	0.5	-30	4.5
APC-f750-A	0.5	-30	4.5

Currently, the package only supports biexponetial transformation for all channels with the getBiexpTransformGS() function. However, the user may choose to create a transformation list explicitly if other transformations (e.g., archsin) are desired.

Note that the flowWorkspace package also allows for an automated transformation calculation “guessing” appropriate parameters. We chose to explicitly specify the biexponential transformation with fixed parameters for all channels to match the manual gating strategy used in flowJo for a more direct comparison of {staRgate} to the manual gating results.

## **Optional**-- to check what pre-transformed data against post
# Check no transformation before
chk_tf <- flowWorkspace::gh_get_transformations(gs)

chk_tf
#> list()

dat_pre_transform <-
  flowWorkspace::gh_pop_get_data(gs) |>
  flowCore::exprs()


# Apply biexp trans
gs <- getBiexpTransformGS(gs, path_biexp_params = path_biexp_params)


## **Optional**-- to check what pre-transformed data against post
# Check no transformation before
chk_tf <- flowWorkspace::gh_get_transformations(gs)

# This is how 1 transformation specification looks like
chk_tf[[1]]
#> function (x, deriv = 0) 
#> {
#>     deriv <- as.integer(deriv)
#>     if (deriv < 0 || deriv > 3) 
#>         stop("'deriv' must be between 0 and 3")
#>     if (deriv > 0) {
#>         z0 <- double(z$n)
#>         z[c("y", "b", "c")] <- switch(deriv, list(y = z$b, b = 2 * 
#>             z$c, c = 3 * z$d), list(y = 2 * z$c, b = 6 * z$d, 
#>             c = z0), list(y = 6 * z$d, b = z0, c = z0))
#>         z[["d"]] <- z0
#>     }
#>     res <- stats:::.splinefun(x, z)
#>     if (deriv > 0 && z$method == 2 && any(ind <- x <= z$x[1L])) 
#>         res[ind] <- ifelse(deriv == 1, z$y[1L], 0)
#>     res
#> }
#> <bytecode: 0x55e1ade4bf90>
#> <environment: 0x55e1ad014b08>
#> attr(,"type")
#> [1] "biexp"
#> attr(,"parameters")
#> attr(,"parameters")$channelRange
#> [1] 4096
#> 
#> attr(,"parameters")$maxValue
#> [1] 262144
#> 
#> attr(,"parameters")$neg
#> [1] 0.5
#> 
#> attr(,"parameters")$pos
#> [1] 4.5
#> 
#> attr(,"parameters")$widthBasis
#> [1] -30

dat_post_transform <-
  flowWorkspace::gh_pop_get_data(gs) |>
  flowCore::exprs()

## **Optional**-- to check that the transformation worked on all provided channels!
summary(dat_pre_transform)
#>      FSC-A            FSC-H            FSC-W            SSC-A       
#>  Min.   :  1492   Min.   :  9988   Min.   : 54676   Min.   :  1033  
#>  1st Qu.: 21138   1st Qu.: 18451   1st Qu.: 99971   1st Qu.: 11850  
#>  Median : 97082   Median : 63860   Median :153544   Median : 26879  
#>  Mean   : 95408   Mean   : 62836   Mean   :152152   Mean   : 31716  
#>  3rd Qu.:137767   3rd Qu.: 90685   3rd Qu.:186615   3rd Qu.: 40599  
#>  Max.   :262144   Max.   :262144   Max.   :262144   Max.   :262144  
#>      SSC-H            SSC-W           AF700-A              APC-A          
#>  Min.   :  1231   Min.   : 21321   Min.   :-40597.32   Min.   :-83668.63  
#>  1st Qu.: 11045   1st Qu.: 70248   1st Qu.:   -89.72   1st Qu.:   -27.74  
#>  Median : 24490   Median : 86074   Median :   -50.31   Median :    80.28  
#>  Mean   : 26669   Mean   : 91965   Mean   :    51.63   Mean   :   893.85  
#>  3rd Qu.: 35312   3rd Qu.:100669   3rd Qu.:    -2.40   3rd Qu.:   442.46  
#>  Max.   :262144   Max.   :262144   Max.   :117999.96   Max.   :258429.72  
#>    APC-f750-A            BB515-A             BB660-A         
#>  Min.   :-15427.155   Min.   : -8649.57   Min.   :-81303.97  
#>  1st Qu.:   -51.782   1st Qu.:   -11.09   1st Qu.:     8.37  
#>  Median :    -0.517   Median :   105.76   Median :   100.04  
#>  Mean   :    86.131   Mean   :   234.41   Mean   :   154.55  
#>  3rd Qu.:    89.942   3rd Qu.:   244.52   3rd Qu.:   235.17  
#>  Max.   : 10246.643   Max.   :133731.45   Max.   :123396.73  
#>     BB700-A             BB790-A              BUV395-A        
#>  Min.   :-77947.38   Min.   :-171716.41   Min.   :-34770.30  
#>  1st Qu.:   -52.99   1st Qu.:    -18.98   1st Qu.:   -32.52  
#>  Median :    49.92   Median :    138.56   Median :   960.00  
#>  Mean   :   209.92   Mean   :    901.38   Mean   :  6796.03  
#>  3rd Qu.:   373.89   3rd Qu.:   1124.65   3rd Qu.: 11336.71  
#>  Max.   : 38183.27   Max.   : 196894.52   Max.   :101130.94  
#>     BUV496-A            BUV563-A            BUV615-A        
#>  Min.   :-21624.89   Min.   : -4940.99   Min.   :  -819.99  
#>  1st Qu.:   -59.50   1st Qu.:   -58.47   1st Qu.:   -25.87  
#>  Median :    72.92   Median :    29.19   Median :    94.06  
#>  Mean   :   508.88   Mean   :   131.69   Mean   :   305.05  
#>  3rd Qu.:  1018.48   3rd Qu.:   163.32   3rd Qu.:   299.30  
#>  Max.   : 12533.95   Max.   :193698.25   Max.   :163601.52  
#>     BUV661-A            BUV737-A            BUV805-A        
#>  Min.   :-19976.12   Min.   :-34474.75   Min.   : -2277.25  
#>  1st Qu.:  -117.70   1st Qu.:   -20.66   1st Qu.:   -67.60  
#>  Median :   -41.57   Median :    81.00   Median :    -7.75  
#>  Mean   :    98.23   Mean   :   860.45   Mean   :   615.80  
#>  3rd Qu.:    61.91   3rd Qu.:  1764.78   3rd Qu.:   136.72  
#>  Max.   :142146.95   Max.   : 98435.47   Max.   :111528.16  
#>     BV421-A             BV480-A             BV510-A         
#>  Min.   :-34983.16   Min.   : -1604.91   Min.   :-62086.07  
#>  1st Qu.:   -10.76   1st Qu.:   -49.19   1st Qu.:   -10.44  
#>  Median :   113.66   Median :    67.76   Median :    83.43  
#>  Mean   :  1321.77   Mean   :   664.87   Mean   :   193.69  
#>  3rd Qu.:   461.01   3rd Qu.:   227.83   3rd Qu.:   197.83  
#>  Max.   :233484.27   Max.   :228909.97   Max.   :205399.70  
#>     BV570-A             BV605-A             BV650-A         
#>  Min.   : -2557.12   Min.   :-15426.25   Min.   : -3029.89  
#>  1st Qu.:   -37.91   1st Qu.:   -21.88   1st Qu.:   -13.74  
#>  Median :    81.17   Median :   172.21   Median :   161.84  
#>  Mean   :   656.24   Mean   :   485.53   Mean   :  1616.83  
#>  3rd Qu.:   250.13   3rd Qu.:   424.23   3rd Qu.:  2921.42  
#>  Max.   :280203.25   Max.   :221790.52   Max.   :196991.55  
#>     BV711-A             BV750-A             BV786-A         
#>  Min.   : -3053.64   Min.   :-16241.38   Min.   :-13683.76  
#>  1st Qu.:   -48.11   1st Qu.:   -86.54   1st Qu.:   -59.14  
#>  Median :    33.28   Median :    15.93   Median :    21.86  
#>  Mean   :  2247.52   Mean   :  1152.19   Mean   :   393.86  
#>  3rd Qu.:   236.58   3rd Qu.:  2227.86   3rd Qu.:   750.71  
#>  Max.   :228820.50   Max.   :141663.83   Max.   :181618.50  
#>       PE-A              PE-CF594-A            PE-Cy5-A        
#>  Min.   :-158344.33   Min.   :-156548.30   Min.   :-14296.08  
#>  1st Qu.:    -74.52   1st Qu.:    -62.45   1st Qu.:   -55.40  
#>  Median :     82.19   Median :     67.09   Median :    14.74  
#>  Mean   :    367.16   Mean   :    679.63   Mean   :    97.37  
#>  3rd Qu.:    302.49   3rd Qu.:    342.16   3rd Qu.:   100.11  
#>  Max.   :  25212.26   Max.   :  95713.01   Max.   :207472.58  
#>    PE-Cy5.5-A           PE-Cy7-A              Time   
#>  Min.   :-24382.69   Min.   :  -489.16   Min.   :48  
#>  1st Qu.:     6.11   1st Qu.:   -37.00   1st Qu.:48  
#>  Median :   120.63   Median :    -2.74   Median :48  
#>  Mean   :   229.33   Mean   :   100.07   Mean   :48  
#>  3rd Qu.:   287.22   3rd Qu.:    34.86   3rd Qu.:48  
#>  Max.   :170867.83   Max.   :196990.38   Max.   :48
summary(dat_post_transform)
#>      FSC-A            FSC-H            FSC-W            SSC-A       
#>  Min.   :  1492   Min.   :  9988   Min.   : 54676   Min.   :  1033  
#>  1st Qu.: 21138   1st Qu.: 18451   1st Qu.: 99971   1st Qu.: 11850  
#>  Median : 97082   Median : 63860   Median :153544   Median : 26879  
#>  Mean   : 95408   Mean   : 62836   Mean   :152152   Mean   : 31716  
#>  3rd Qu.:137767   3rd Qu.: 90685   3rd Qu.:186615   3rd Qu.: 40599  
#>  Max.   :262144   Max.   :262144   Max.   :262144   Max.   :262144  
#>      SSC-H            SSC-W           AF700-A             APC-A         
#>  Min.   :  1231   Min.   : 21321   Min.   :-16892.6   Min.   :-35262.5  
#>  1st Qu.: 11045   1st Qu.: 70248   1st Qu.:   840.9   1st Qu.:   959.7  
#>  Median : 24490   Median : 86074   Median :   915.9   Median :  1169.3  
#>  Mean   : 26669   Mean   : 91965   Mean   :   990.0   Mean   :  1415.6  
#>  3rd Qu.: 35312   3rd Qu.:100669   3rd Qu.:  1009.3   3rd Qu.:  1682.2  
#>  Max.   :262144   Max.   :262144   Max.   :  3812.1   Max.   :  4091.4  
#>    APC-f750-A         BB515-A           BB660-A          BB700-A        
#>  Min.   :-6157.6   Min.   :-3266.9   Min.   :-34254   Min.   :-32822.4  
#>  1st Qu.:  913.1   1st Qu.:  992.3   1st Qu.:  1030   1st Qu.:   910.8  
#>  Median : 1013.0   Median : 1216.9   Median :  1206   Median :  1111.3  
#>  Mean   : 1120.5   Mean   : 1262.0   Mean   :  1245   Mean   :  1252.8  
#>  3rd Qu.: 1187.5   3rd Qu.: 1446.7   3rd Qu.:  1433   3rd Qu.:  1611.5  
#>  Max.   : 2933.3   Max.   : 3856.7   Max.   :  3828   Max.   :  3408.8  
#>     BB790-A            BUV395-A           BUV496-A          BUV563-A      
#>  Min.   :-72814.7   Min.   :-14407.4   Min.   :-8800.9   Min.   :-1685.2  
#>  1st Qu.:   976.8   1st Qu.:   950.4   1st Qu.:  898.2   1st Qu.:  900.2  
#>  Median :  1276.0   Median :  2015.1   Median : 1155.4   Median : 1071.1  
#>  Mean   :  1518.0   Mean   :  1975.3   Mean   : 1407.2   Mean   : 1139.9  
#>  3rd Qu.:  2081.6   3rd Qu.:  2970.2   3rd Qu.: 2040.1   3rd Qu.: 1318.8  
#>  Max.   :  3994.6   Max.   :  3757.0   Max.   : 3006.8   Max.   : 3988.7  
#>     BUV615-A          BUV661-A          BUV737-A           BUV805-A     
#>  Min.   :  79.94   Min.   :-8097.7   Min.   :-14281.3   Min.   :-549.1  
#>  1st Qu.: 963.39   1st Qu.:  789.3   1st Qu.:   973.6   1st Qu.: 882.7  
#>  Median :1195.18   Median :  932.8   Median :  1170.7   Median : 998.8  
#>  Mean   :1278.48   Mean   : 1032.8   Mean   :  1526.5   Mean   :1204.5  
#>  3rd Qu.:1522.04   3rd Qu.: 1134.4   3rd Qu.:  2265.8   3rd Qu.:1272.7  
#>  Max.   :3928.57   Max.   : 3878.5   Max.   :  3747.4   Max.   :3792.0  
#>     BV421-A            BV480-A          BV510-A            BV570-A      
#>  Min.   :-14498.2   Min.   :-262.4   Min.   :-26057.5   Min.   :-668.5  
#>  1st Qu.:   992.9   1st Qu.: 918.1   1st Qu.:   993.5   1st Qu.: 939.9  
#>  Median :  1231.3   Median :1145.6   Median :  1175.3   Median :1171.0  
#>  Mean   :  1445.3   Mean   :1265.6   Mean   :  1209.3   Mean   :1305.1  
#>  3rd Qu.:  1699.8   3rd Qu.:1422.0   3rd Qu.:  1375.6   3rd Qu.:1454.8  
#>  Max.   :  4055.3   Max.   :4048.2   Max.   :  4009.6   Max.   :4121.1  
#>     BV605-A           BV650-A          BV711-A          BV750-A       
#>  Min.   :-6157.2   Min.   :-870.1   Min.   :-880.3   Min.   :-6504.8  
#>  1st Qu.:  971.2   1st Qu.: 987.1   1st Qu.: 920.2   1st Qu.:  846.9  
#>  Median : 1333.7   Median :1316.3   Median :1079.1   Median : 1045.2  
#>  Mean   : 1347.4   Mean   :1626.2   Mean   :1371.0   Mean   : 1523.6  
#>  3rd Qu.: 1664.4   3rd Qu.:2463.9   3rd Qu.:1435.1   3rd Qu.: 2358.3  
#>  Max.   : 4037.0   Max.   :3994.7   Max.   :4048.1   Max.   : 3877.2  
#>     BV786-A             PE-A            PE-CF594-A          PE-Cy5-A      
#>  Min.   :-5414.0   Min.   :-67111.6   Min.   :-66345.6   Min.   :-5675.2  
#>  1st Qu.:  898.9   1st Qu.:   869.6   1st Qu.:   892.6   1st Qu.:  906.1  
#>  Median : 1056.8   Median :  1172.9   Median :  1144.3   Median : 1042.9  
#>  Mean   : 1338.9   Mean   :  1248.4   Mean   :  1320.0   Mean   : 1091.7  
#>  3rd Qu.: 1910.2   3rd Qu.:  1526.2   3rd Qu.:  1575.1   3rd Qu.: 1206.4  
#>  Max.   : 3965.8   Max.   :  3259.6   Max.   :  3737.4   Max.   : 4013.2  
#>    PE-Cy5.5-A       PE-Cy7-A           Time   
#>  Min.   :-9977   Min.   : 302.8   Min.   :48  
#>  1st Qu.: 1026   1st Qu.: 941.7   1st Qu.:48  
#>  Median : 1244   Median :1008.6   Median :48  
#>  Mean   : 1289   Mean   :1053.8   Mean   :48  
#>  3rd Qu.: 1506   3rd Qu.:1082.1   3rd Qu.:48  
#>  Max.   : 3944   Max.   :3994.7   Max.   :48

## **Optional**-- save out the transformerList
samp_metadata$transformerList <- chk_tf

Pre-gating

In this context, pre-gating is defined as gating from the root population up to CD3+, or CD4+/CD8+ subsets

Then we will pass the data to the density gating step.

The flowAI step serves as a quality control (QC) to match the first Time gate step that is typically done in manual gating. It is possible, however, that the user may choose to skip this step if flowAI excludes too many cells.

The first step of the gating template is a QC step that is especially important to include if the user chooses to exclude the flowAI step.

## Optional - flowAI step 
# set to TRUE if want to include
flowAI_yn <- FALSE

# Run flowAI if TRUE, run it
if (flowAI_yn) {
  # Check if the dir exists
  if (!dir.exists(glue::glue("{path_out}/flowAI_results"))) {
    dir.create(glue::glue("{path_out}/flowAI_results"))
  }

  # For sample
  qc_flowai <- flowAI::flow_auto_qc(flowWorkspace::gh_pop_get_data(gs),
    folder_results = here::here(glue::glue("{path_out}/flowAI_results"))
  )

  # Convert to a flowSet in order to convert to GatingSet
  qc_flowai <- flowCore::flowSet(qc_flowai)

  # Rnemae the sample
  sampleNames(qc_flowai) <- pt_samp_nm

  # convert the flowFrame obj returned in test_qc to a GatingSet to pass to openCyto
  gs_qc <- flowWorkspace::GatingSet(qc_flowai)

  # Remove to save memory
  rm(qc_flowai)

  # Save data
  save_gs(gs_qc,
    path = paste0(glue::glue("{path_out}/data/GatingSet/{pt_samp_nm}_flowAI_qc_{Sys.Date()}"))
  )
} else {
  # if no need to run flowAI, just set gs_qc as gs from previous transformation step
  gs_qc <- gs
  rm(gs)
}

# Pre-gating up to CD4/8+ with `r BiocStyle::Biocpkg("openCyto")`
## Set seed using today's date
set.seed(glue::glue({
  format(Sys.Date(), format = "%Y%m%d")
}))

openCyto::gt_gating(gt_tcell, gs_qc)
#> Gating for 'fsc_ssc_qc'
#> done!
#> done.
#> Gating for 'nonDebris'
#> done!
#> done.
#> Gating for 'singlets'
#> done!
#> done.
#> Gating for 'cd14-cd19-'
#> done!
#> done.
#> Gating for 'live'
#> done!
#> done.
#> Gating for 'cd3'
#> done!
#> done.
#> Gating for 'cd8+'
#> done!
#> done.
#> Gating for 'cd4+'
#> done!
#> done.
#> Population 'cd4-cd8-'
#> done.
#> Population 'cd4+cd8-'
#> done.
#> Population 'cd4-cd8+'
#> done.
#> Population 'cd4+cd8+'
#> done.
#> finished.

## Check autoplot
ggcyto::autoplot(gs_qc[[1]])

Extract intensity matrix

# Extract intensity matrix from GatingSet object
## Grab marker names from GatingSet for labeling col names in intensity matrix
## Can skip this step if you know the exact namings of the marker names in your FCS files
# In that case, supply a string is fine:
# marker_chnl_names = c("CD45RA", "CCR7", "LAG3", ...)
marker_chnl_names <-
  flowWorkspace::gh_pop_get_data(gs_qc) |>
  flowWorkspace::markernames() |>
  data.frame("marker_full" = _) |>
  tibble::rownames_to_column(var = "chnl") |>
  # clean up the names
  dplyr::mutate(
    marker_full = janitor::make_clean_names(marker_full, replace = c("-" = "", "_" = "", " " = "")) |> toupper()
  ) |>
  # Reorder the marker channel names to start with CD3, CD4, CD8 then the rest
  dplyr::arrange(match(marker_full, c("CD3", "CD4", "CD8"))) |> 
  # Clean up
  dplyr::mutate(marker_full = 
                  dplyr::case_when(marker_full == "FOX_P3" ~ "FOXP3",
                                   .default = marker_full))

# For our Tcell panel, we only want to apply the density gating on
# the `markers_to_gate` markers
# Again, this can be made specified explicitly as string
markers_to_gate <-
  marker_chnl_names$marker_full |> (\(x){
    x[!(x %in% c("CD3", "CD4", "CD8", "LD", "CD1419"))]  
  })()

## Grab the intensity matrix from GatingSet
## the gh_pop_get_indices grabs the 0/1 for whether gated as CD3
intensity_dat <-
  # grab all indices from opencyto gating to be complete
  flowWorkspace::gh_pop_get_indices(gs_qc, y = "fsc_ssc_qc") |>
  cbind(
    flowWorkspace::gh_pop_get_data(gs_qc) |>
      flowCore::exprs(),
    "fsc_ssc_qc" = _
  ) |>
  # Add nondebris
  cbind(
    "nonDebris" = flowWorkspace::gh_pop_get_indices(gs_qc, y = "nonDebris")
  ) |>
  # singlets
  cbind(
    "singlets" = flowWorkspace::gh_pop_get_indices(gs_qc, y = "singlets")
  ) |>
  # cd14-19-
  cbind(
    "cd14_neg_19_neg" = flowWorkspace::gh_pop_get_indices(gs_qc, y = "cd14-cd19-")
  ) |>
  # add live
  cbind(
    "live" = flowWorkspace::gh_pop_get_indices(gs_qc, y = "live")
  ) |>
  # Add cd3+
  cbind(
    "cd3_pos" = flowWorkspace::gh_pop_get_indices(gs_qc, y = "cd3")
  ) |>
  # add on the cd4 and cd8 0/1s
  cbind(
    "cd4_pos" = flowWorkspace::gh_pop_get_indices(gs_qc, y = "cd4+")
  ) |>
  cbind(
    "cd8_pos" = flowWorkspace::gh_pop_get_indices(gs_qc, y = "cd8+")
  ) |>
  tibble::as_tibble() |>
  # Rename with marker names
  dplyr::rename(stats::setNames(marker_chnl_names$chnl, as.character(marker_chnl_names$marker_full))) |>
  dplyr::mutate(
    cd4_pos_cd8_pos = dplyr::case_when(
      cd3_pos == 1 & cd4_pos == 1 & cd8_pos == 1 ~ "cd4_pos_cd8_pos",
      cd3_pos == 1 & cd4_pos == 1 & cd8_pos == 0 ~ "cd4_pos_cd8_neg",
      cd3_pos == 1 & cd4_pos == 0 & cd8_pos == 1 ~ "cd4_neg_cd8_pos",
      cd3_pos == 1 & cd4_pos == 0 & cd8_pos == 0 ~ "cd4_neg_cd8_neg"
    )
  )

## Preview of intensity matrix
head(intensity_dat)
#> # A tibble: 6 × 44
#>   `FSC-A` `FSC-H` `FSC-W` `SSC-A` `SSC-H` `SSC-W`  KI67  TBET GZM_B   PD1  LAG3
#>     <dbl>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1  14579.  12925.  91156.   6783.   6899.  59440.  898.  995.  949.  922. 1184.
#> 2  47535.  36802. 124831.  21010.  20507.  79210.  732. 1178.  775.  914. 1169.
#> 3 117789.  81878. 183615.  41468.  38820.  93061.  814. 2613. 1960. 1372. 1330.
#> 4 230540.  99239. 192036.  48270.  23680. 129099.  821. 1006. 1156. 1675. 1581.
#> 5  65184.  43167. 138985.  25413.  21602.  88339.  654. 1252. 1004. 1069. 1257.
#> 6 138285.  96765. 182364.  37571.  34888.  93352.  705. 2678.  984. 1034. 1011.
#> # ℹ 33 more variables: CD127 <dbl>, CD38 <dbl>, CD45RA <dbl>, CD4 <dbl>,
#> #   ICOS <dbl>, CD25 <dbl>, TIM3 <dbl>, CD27 <dbl>, CD8 <dbl>, CD57 <dbl>,
#> #   CXCR5 <dbl>, LD <dbl>, CD1419 <dbl>, CCR4 <dbl>, CCR7 <dbl>, HLADR <dbl>,
#> #   CD3 <dbl>, CD28 <dbl>, TIGIT <dbl>, EOMES <dbl>, CTLA4 <dbl>, FOXP3 <dbl>,
#> #   GITR <dbl>, Time <dbl>, fsc_ssc_qc <dbl>, nonDebris <dbl>, singlets <dbl>,
#> #   cd14_neg_19_neg <dbl>, live <dbl>, cd3_pos <dbl>, cd4_pos <dbl>,
#> #   cd8_pos <dbl>, cd4_pos_cd8_pos <chr>

Apply density gating

The suggested strategy is based on all CD3+, but this can be customized based on the string corresponding to the column name supplied to subset_col in the get_density_gates function

The suggested number of bins for density estimation is 40.

For illustration purposes, we will only apply density gating on a few markers.

# Density gating parameters
peak_r <- 10
bin_i <- 40

neg_intensity_thres <- -1000

# select a few markers to gate
example_markers <- c("LAG3", "CCR7", "CD45RA", "EOMES")

# Read in positive peak thresholds
pos_thres <- utils::read.csv(path_pos_peak_thresholds)

# Show the exact output from the get_density_gates() function
dens_gates_pre <-
  dplyr::filter(intensity_dat, cd3_pos %in% c(0, 1)) |>
  getDensityGates(
    intens_dat = _,
    marker = example_markers,
    subset_col = "cd3_pos",
    bin_n = bin_i,
    peak_detect_ratio = peak_r,
    pos_peak_threshold = pos_thres,
    neg_intensity_threshold = neg_intensity_thres
  )

dens_gates_pre
#> # A tibble: 2 × 5
#>   cd3_pos  LAG3  CCR7 CD45RA EOMES
#>     <dbl> <dbl> <dbl>  <dbl> <dbl>
#> 1       0 1328. 1452.  1375. 1533.
#> 2       1 1634. 1768.  1486. 1741.

# Since we apply density gating on CD3+ cells but
# Would like to calculate subpopulations with CD4+ and CD8+ as
# the starting parent population, we need to add corresponding rows
dens_gates <-
  dplyr::filter(dens_gates_pre, cd3_pos == 1) |>
  tibble::add_row() |>
  tibble::add_row() |>
  tibble::add_row() |>
  dplyr::mutate(cd4_pos_cd8_pos = c("cd4_neg_cd8_neg", "cd4_pos_cd8_neg", "cd4_neg_cd8_pos", "cd4_pos_cd8_pos")) |>
  tidyr::fill(cd3_pos, dplyr::all_of(example_markers), .direction = "down")


# View updated gates with the col for CD4/CD8
dens_gates
#> # A tibble: 4 × 6
#>   cd3_pos  LAG3  CCR7 CD45RA EOMES cd4_pos_cd8_pos
#>     <dbl> <dbl> <dbl>  <dbl> <dbl> <chr>          
#> 1       1 1634. 1768.  1486. 1741. cd4_neg_cd8_neg
#> 2       1 1634. 1768.  1486. 1741. cd4_pos_cd8_neg
#> 3       1 1634. 1768.  1486. 1741. cd4_neg_cd8_pos
#> 4       1 1634. 1768.  1486. 1741. cd4_pos_cd8_pos

# try get indicator col
example_intensity_gated <-
  getGatedDat(intensity_dat,
    subset_col = "cd4_pos_cd8_pos",
    cutoffs = dens_gates
  )


# Plot the gate for visual
intensity_dat |>
  dplyr::filter(cd3_pos == 1) |>
  # additional step to remove large intensity values only when density gating.
  # Still kept in the data
  dplyr::filter(!(dplyr::if_any(dplyr::all_of(markers_to_gate), ~ .x < neg_intensity_thres))) |>
  ggplot() +
  geom_density(aes(LAG3)) +
  geom_vline(
    data = dens_gates,
    aes(xintercept = LAG3),
    color = "blue",
    linetype = "dashed"
  ) +
  labs(subtitle = "Distribution of LAG3 intensity on all CD3+. Gate identifed by {staRgate} in blue.")

# If by CD4/CD8,
intensity_dat |>
  dplyr::filter(cd3_pos == 1) |>
  # additional step to remove large intensity values only when density gating.
  # Still kept in the data
  dplyr::filter(!(dplyr::if_any(dplyr::all_of(markers_to_gate), ~ .x < neg_intensity_thres))) |>
  ggplot() +
  geom_density(aes(LAG3)) +
  geom_vline(
    data = dens_gates,
    aes(xintercept = LAG3),
    color = "blue",
    linetype = "dashed"
  ) +
  labs(subtitle = "Distribution of LAG3 intensity on all CD3+. Gate identifed by {staRgate} in blue.") +
  facet_wrap(~cd4_pos_cd8_pos)

# For CCR7
intensity_dat |>
  dplyr::filter(cd3_pos == 1) |>
  # additional step to remove large intensity values only when density gating.
  # Still kept in the data
  dplyr::filter(!(dplyr::if_any(dplyr::all_of(markers_to_gate), ~ .x < neg_intensity_thres))) |>
  ggplot() +
  geom_density(aes(CCR7)) +
  geom_vline(
    data = dens_gates,
    aes(xintercept = CCR7),
    color = "blue",
    linetype = "dashed"
  ) +
  labs(subtitle = "Distribution of CCR7 intensity on all CD3+. Gate identifed by {staRgate} in blue.")

# If by CD4/CD8,
intensity_dat |>
  dplyr::filter(cd3_pos == 1) |>
  # additional step to remove large intensity values only when density gating.
  # Still kept in the data
  dplyr::filter(!(dplyr::if_any(dplyr::all_of(markers_to_gate), ~ .x < neg_intensity_thres))) |>
  ggplot() +
  geom_density(aes(CCR7)) +
  geom_vline(
    data = dens_gates,
    aes(xintercept = CCR7),
    color = "blue",
    linetype = "dashed"
  ) +
  labs(subtitle = "Distribution of CCR7 intensity on all CD3+. Gate identifed by {staRgate} in blue.") +
  facet_wrap(~cd4_pos_cd8_pos)

# For CD45RA
intensity_dat |>
  dplyr::filter(cd3_pos == 1) |>
  # additional step to remove large intensity values only when density gating.
  # Still kept in the data
  dplyr::filter(!(dplyr::if_any(dplyr::all_of(markers_to_gate), ~ .x < neg_intensity_thres))) |>
  ggplot() +
  geom_density(aes(CD45RA)) +
  geom_vline(
    data = dens_gates,
    aes(xintercept = CD45RA),
    color = "blue",
    linetype = "dashed"
  ) +
  labs(subtitle = "Distribution of CD45RA intensity on all CD3+. Gate identifed by {staRgate} in blue.")

# If by CD4/CD8,
intensity_dat |>
  dplyr::filter(cd3_pos == 1) |>
  # additional step to remove large intensity values only when density gating.
  # Still kept in the data
  dplyr::filter(!(dplyr::if_any(dplyr::all_of(markers_to_gate), ~ .x < neg_intensity_thres))) |>
  ggplot() +
  geom_density(aes(CD45RA)) +
  geom_vline(
    data = dens_gates,
    aes(xintercept = CD45RA),
    color = "blue",
    linetype = "dashed"
  ) +
  labs(subtitle = "Distribution of CD45RA intensity on all CD3+. Gate identifed by {staRgate} in blue.") +
  facet_wrap(~cd4_pos_cd8_pos)

Getting percentage data

We can then summarize the single-cell level data to counts and percentages of cells for all combinations of markers.

For the subpopulations, the denominator is defined as the parent population and numerator is the population of interest out of the parent population. The $n_d$ refers to the number of markers considered for the denominator and $n$ for the number of markers considered for the numerator.

For the 29-marker panel, if the denominator is specified as the CD4 and CD8 subsets, then $n_d = 2$ and $n = 23$ for the markers of interest.

The getPerc function allows user to list the markers of interest for the numerator and denominator

In the example below, we will consider CD4 and CD8 subsets as the key parent populations of interest (denominator) and the three markers we gated on using c("LAG3", "CCR7", "CD45RA") (numerator markers).

The additional arguments expand_num and expand_denom generates different lists of subpopulations to calculate counts/percentages for:

• expand_num: should calculations consider up to pairs of numerator markers included?,
• expand_denom: should the calculations consider combinations of each numerator marker and parent populations specified in the denominator?

Currently, we support the four scenarios listed below:

expand_num	expand_denom	The subpopulations to generate	Examples	Expected number of subpopulations1	Expected number of subpopulations for the 29-marker T-cell panel
FALSE	FALSE	Numerator: Positive and negative for each marker specified Denominator: Combinations of positive and negative for markers specified	Numerator specified to (“LAG3”) and denominator specified as (“CD4”, “CD8”): LAG3+/- of CD4- & CD8- LAG3+/- of CD4+ & CD8- LAG3+/- of CD4+ & CD8+ LAG3+/- of CD4- & CD8+	2n*(2nd)	184
TRUE	FALSE	Numerator: Positive and negative for each marker specified, and combinations of positive/negative for pairs of markers At least 2 markers must be included in the numerator Denominator: Combinations of positive and negative for marker(s)	Numerator specified as (“LAG3”, “KI67”) and denominator as (“CD4”): LAG3+/- of CD4+/- KI67+/- of CD4+/- LAG3- KI67- of CD4+/- LAG3+ KI67- of CD4+/- LAG3- KI67+ of CD4+/- LAG3+ KI67+ of CD4+/-	2n*(2nd) + 2nd*(n choose 2)*(22)	4232
FALSE	TRUE	Numerator: Positive and negative for each marker specified At least 2 markers must be included in the numerator Denominator: Combinations of positive and negative for marker(s), and combinations of positive and negative for marker(s) in denominator with one marker from the numerator	Numerator specified as (“LAG3”, “KI67”) and denominator as (“CD4”) LAG3+/- of CD4+/- KI67+/- of CD4+/- LAG3+/- of CD4-KI67- LAG3+/- of CD4-KI67+ LAG3+/- of CD4+KI67- LAG3+/- of CD4+KI67+ KI67+/- of CD4-LAG3- KI67+/- of CD4-LAG3+ KI67+/- of CD4+LAG3- KI67+/- of CD4+LAG3+	2n*(2nd) + 4*2nd*n*(n-1)	8280
TRUE	TRUE	Numerator: Positive and negative for each marker specified, and combinations of positive/negative for pairs of markers At least 3 markers must be included in the numerator Denominator: Combinations of positive and negative for marker(s), and combinations of positive and negative for marker(s) in denominator with one marker from the numerator	Numerator specified as (“LAG3”, “KI67”, “CTLA4”) and denominator as (“CD4”) LAG3+/- of CD4+/- KI67+/- of CD4+/- CTLA4+/- of CD4+/- LAG3+/- of CD4-KI67- LAG3+/- of CD4-KI67+ LAG3+/- of CD4+KI67- LAG3+/- of CD4+KI67+ KI67+/- of CD4-LAG3- KI67+/- of CD4-LAG3+ KI67+/- of CD4+LAG3- KI67+/- of CD4+LAG3+ …. LAG3- KI67- of CD4+/- LAG3+ KI67- of CD4+/- LAG3- KI67+ of CD4+/- LAG3+ KI67+ of CD4+/- … LAG3+ KI67+ of CD4+/- & CTLA4+/- …	2n*(2nd) + 2nd*(n choose 2)*(22) + 4*2nd*n*(n-1) + 22*n*((n-1) choose 2)*(2nd+1)	182344
1 n, number of markers specified in the numerator; nd, number of markers specified in the denominator

The keep_indicators argument provides the 0/1 for which marker is considered in the numerator and denominator for each subpopulation. This is especially useful when merging onto other data that does not have the same format.

For example, when matching strings: “CD4+ & CD8- of CD3+” is different from “CD8- & CD4+ of CD3+” and “CD4+ and CD8- of CD3+”

Below is we show examples of each of the four combinations for the expand_num and expand_denom arguments.

For the example of when expand_num = FALSE and expand_denom = FALSE, keep_indicators = TRUE to illustrate the columns we get for the _POS and _POS_D. Other examples use keep_indicators = FALSE.

Expand example for expand_num = FALSE and expand_denom = FALSE, and keep_indicators = TRUE

example_perc1 <-
  # Should only count the CD3+ cells
  dplyr::filter(example_intensity_gated, cd3_pos == 1) |> 
  getPerc(
    intens_dat = _,
    num_marker = example_markers,
    denom_marker = c("CD4", "CD8"),
    expand_num = FALSE,
    expand_denom = FALSE,
    keep_indicators = TRUE
  )

# For display only, group based on the denominators and
# simplify the names to be numerators
example_perc1 |> 
  tidyr::separate_wider_delim(subpopulation,
    delim = "_OF_",
    names = c("num", "denom"),
    cols_remove = FALSE
  ) |>
  dplyr::mutate(denom = paste("Denom = ", denom)) |>
  dplyr::group_by(denom) |>
  dplyr::select(-subpopulation) |>
  gt::gt() |>
  gt::fmt_number(
    columns = "perc",
    decimals = 1
  )

num	n_num	n_denom	perc	LAG3_POS	CCR7_POS	CD45RA_POS	EOMES_POS	CD4_POS_D	CD8_POS_D
Denom = CD4_NEG_CD8_NEG
LAG3_NEG	196	271	72.3	0	NA	NA	NA	0	0
CCR7_NEG	220	271	81.2	NA	0	NA	NA	0	0
CD45RA_NEG	21	271	7.7	NA	NA	0	NA	0	0
EOMES_NEG	59	271	21.8	NA	NA	NA	0	0	0
LAG3_POS	75	271	27.7	1	NA	NA	NA	0	0
CCR7_POS	51	271	18.8	NA	1	NA	NA	0	0
CD45RA_POS	250	271	92.3	NA	NA	1	NA	0	0
EOMES_POS	212	271	78.2	NA	NA	NA	1	0	0
Denom = CD4_NEG_CD8_POS
LAG3_NEG	1711	2302	74.3	0	NA	NA	NA	0	1
CCR7_NEG	1699	2302	73.8	NA	0	NA	NA	0	1
CD45RA_NEG	413	2302	17.9	NA	NA	0	NA	0	1
EOMES_NEG	648	2302	28.1	NA	NA	NA	0	0	1
LAG3_POS	591	2302	25.7	1	NA	NA	NA	0	1
CCR7_POS	603	2302	26.2	NA	1	NA	NA	0	1
CD45RA_POS	1889	2302	82.1	NA	NA	1	NA	0	1
EOMES_POS	1654	2302	71.9	NA	NA	NA	1	0	1
Denom = CD4_POS_CD8_NEG
LAG3_NEG	8289	8615	96.2	0	NA	NA	NA	1	0
CCR7_NEG	1395	8615	16.2	NA	0	NA	NA	1	0
CD45RA_NEG	1830	8615	21.2	NA	NA	0	NA	1	0
EOMES_NEG	7800	8615	90.5	NA	NA	NA	0	1	0
LAG3_POS	326	8615	3.8	1	NA	NA	NA	1	0
CCR7_POS	7220	8615	83.8	NA	1	NA	NA	1	0
CD45RA_POS	6785	8615	78.8	NA	NA	1	NA	1	0
EOMES_POS	815	8615	9.5	NA	NA	NA	1	1	0
Denom = CD4_POS_CD8_POS
LAG3_NEG	139	276	50.4	0	NA	NA	NA	1	1
CCR7_NEG	152	276	55.1	NA	0	NA	NA	1	1
CD45RA_NEG	35	276	12.7	NA	NA	0	NA	1	1
EOMES_NEG	71	276	25.7	NA	NA	NA	0	1	1
LAG3_POS	137	276	49.6	1	NA	NA	NA	1	1
CCR7_POS	124	276	44.9	NA	1	NA	NA	1	1
CD45RA_POS	241	276	87.3	NA	NA	1	NA	1	1
EOMES_POS	205	276	74.3	NA	NA	NA	1	1	1

Expand example for expand_num = TRUE and expand_denom = FALSE, and keep_indicators = FALSE

example_perc2 <-
  # Should only count the CD3+ cells
  dplyr::filter(example_intensity_gated, cd3_pos == 1) |>
  getPerc(
    intens_dat = _,
    num_marker = example_markers,
    denom_marker = c("CD4", "CD8"),
    expand_num = TRUE,
    expand_denom = FALSE,
    keep_indicators = FALSE
  )

# For display only, group based on the denominators and
# simplify the names to be numerators
example_perc2 |>
  tidyr::separate_wider_delim(subpopulation,
    delim = "_OF_",
    names = c("num", "denom"),
    cols_remove = FALSE
  ) |>
  dplyr::mutate(denom = paste("Denom = ", denom)) |>
  dplyr::group_by(denom) |>
  dplyr::select(-subpopulation) |>
  gt::gt() |>
  gt::fmt_number(
    columns = "perc",
    decimals = 1
  )

num	n_num	n_denom	perc
Denom = CD4_NEG_CD8_NEG
LAG3_NEG	196	271	72.3
CCR7_NEG	220	271	81.2
CD45RA_NEG	21	271	7.7
EOMES_NEG	59	271	21.8
LAG3_POS	75	271	27.7
CCR7_POS	51	271	18.8
CD45RA_POS	250	271	92.3
EOMES_POS	212	271	78.2
LAG3_NEG_CCR7_NEG	147	271	54.2
LAG3_NEG_CD45RA_NEG	17	271	6.3
LAG3_NEG_EOMES_NEG	58	271	21.4
LAG3_NEG_CCR7_POS	49	271	18.1
LAG3_NEG_CD45RA_POS	179	271	66.1
LAG3_NEG_EOMES_POS	138	271	50.9
CCR7_NEG_CD45RA_NEG	17	271	6.3
CCR7_NEG_EOMES_NEG	20	271	7.4
CCR7_NEG_LAG3_POS	73	271	26.9
CCR7_NEG_CD45RA_POS	203	271	74.9
CCR7_NEG_EOMES_POS	200	271	73.8
CD45RA_NEG_EOMES_NEG	7	271	2.6
CD45RA_NEG_LAG3_POS	4	271	1.5
CD45RA_NEG_CCR7_POS	4	271	1.5
CD45RA_NEG_EOMES_POS	14	271	5.2
EOMES_NEG_LAG3_POS	1	271	0.4
EOMES_NEG_CCR7_POS	39	271	14.4
EOMES_NEG_CD45RA_POS	52	271	19.2
LAG3_POS_CCR7_POS	2	271	0.7
LAG3_POS_CD45RA_POS	71	271	26.2
LAG3_POS_EOMES_POS	74	271	27.3
CCR7_POS_CD45RA_POS	47	271	17.3
CCR7_POS_EOMES_POS	12	271	4.4
CD45RA_POS_EOMES_POS	198	271	73.1
Denom = CD4_NEG_CD8_POS
LAG3_NEG	1711	2302	74.3
CCR7_NEG	1699	2302	73.8
CD45RA_NEG	413	2302	17.9
EOMES_NEG	648	2302	28.1
LAG3_POS	591	2302	25.7
CCR7_POS	603	2302	26.2
CD45RA_POS	1889	2302	82.1
EOMES_POS	1654	2302	71.9
LAG3_NEG_CCR7_NEG	1165	2302	50.6
LAG3_NEG_CD45RA_NEG	344	2302	14.9
LAG3_NEG_EOMES_NEG	587	2302	25.5
LAG3_NEG_CCR7_POS	546	2302	23.7
LAG3_NEG_CD45RA_POS	1367	2302	59.4
LAG3_NEG_EOMES_POS	1124	2302	48.8
CCR7_NEG_CD45RA_NEG	329	2302	14.3
CCR7_NEG_EOMES_NEG	258	2302	11.2
CCR7_NEG_LAG3_POS	534	2302	23.2
CCR7_NEG_CD45RA_POS	1370	2302	59.5
CCR7_NEG_EOMES_POS	1441	2302	62.6
CD45RA_NEG_EOMES_NEG	125	2302	5.4
CD45RA_NEG_LAG3_POS	69	2302	3.0
CD45RA_NEG_CCR7_POS	84	2302	3.6
CD45RA_NEG_EOMES_POS	288	2302	12.5
EOMES_NEG_LAG3_POS	61	2302	2.6
EOMES_NEG_CCR7_POS	390	2302	16.9
EOMES_NEG_CD45RA_POS	523	2302	22.7
LAG3_POS_CCR7_POS	57	2302	2.5
LAG3_POS_CD45RA_POS	522	2302	22.7
LAG3_POS_EOMES_POS	530	2302	23.0
CCR7_POS_CD45RA_POS	519	2302	22.5
CCR7_POS_EOMES_POS	213	2302	9.3
CD45RA_POS_EOMES_POS	1366	2302	59.3
Denom = CD4_POS_CD8_NEG
LAG3_NEG	8289	8615	96.2
CCR7_NEG	1395	8615	16.2
CD45RA_NEG	1830	8615	21.2
EOMES_NEG	7800	8615	90.5
LAG3_POS	326	8615	3.8
CCR7_POS	7220	8615	83.8
CD45RA_POS	6785	8615	78.8
EOMES_POS	815	8615	9.5
LAG3_NEG_CCR7_NEG	1221	8615	14.2
LAG3_NEG_CD45RA_NEG	1700	8615	19.7
LAG3_NEG_EOMES_NEG	7627	8615	88.5
LAG3_NEG_CCR7_POS	7068	8615	82.0
LAG3_NEG_CD45RA_POS	6589	8615	76.5
LAG3_NEG_EOMES_POS	662	8615	7.7
CCR7_NEG_CD45RA_NEG	927	8615	10.8
CCR7_NEG_EOMES_NEG	811	8615	9.4
CCR7_NEG_LAG3_POS	174	8615	2.0
CCR7_NEG_CD45RA_POS	468	8615	5.4
CCR7_NEG_EOMES_POS	584	8615	6.8
CD45RA_NEG_EOMES_NEG	1461	8615	17.0
CD45RA_NEG_LAG3_POS	130	8615	1.5
CD45RA_NEG_CCR7_POS	903	8615	10.5
CD45RA_NEG_EOMES_POS	369	8615	4.3
EOMES_NEG_LAG3_POS	173	8615	2.0
EOMES_NEG_CCR7_POS	6989	8615	81.1
EOMES_NEG_CD45RA_POS	6339	8615	73.6
LAG3_POS_CCR7_POS	152	8615	1.8
LAG3_POS_CD45RA_POS	196	8615	2.3
LAG3_POS_EOMES_POS	153	8615	1.8
CCR7_POS_CD45RA_POS	6317	8615	73.3
CCR7_POS_EOMES_POS	231	8615	2.7
CD45RA_POS_EOMES_POS	446	8615	5.2
Denom = CD4_POS_CD8_POS
LAG3_NEG	139	276	50.4
CCR7_NEG	152	276	55.1
CD45RA_NEG	35	276	12.7
EOMES_NEG	71	276	25.7
LAG3_POS	137	276	49.6
CCR7_POS	124	276	44.9
CD45RA_POS	241	276	87.3
EOMES_POS	205	276	74.3
LAG3_NEG_CCR7_NEG	79	276	28.6
LAG3_NEG_CD45RA_NEG	26	276	9.4
LAG3_NEG_EOMES_NEG	50	276	18.1
LAG3_NEG_CCR7_POS	60	276	21.7
LAG3_NEG_CD45RA_POS	113	276	40.9
LAG3_NEG_EOMES_POS	89	276	32.2
CCR7_NEG_CD45RA_NEG	19	276	6.9
CCR7_NEG_EOMES_NEG	33	276	12.0
CCR7_NEG_LAG3_POS	73	276	26.4
CCR7_NEG_CD45RA_POS	133	276	48.2
CCR7_NEG_EOMES_POS	119	276	43.1
CD45RA_NEG_EOMES_NEG	21	276	7.6
CD45RA_NEG_LAG3_POS	9	276	3.3
CD45RA_NEG_CCR7_POS	16	276	5.8
CD45RA_NEG_EOMES_POS	14	276	5.1
EOMES_NEG_LAG3_POS	21	276	7.6
EOMES_NEG_CCR7_POS	38	276	13.8
EOMES_NEG_CD45RA_POS	50	276	18.1
LAG3_POS_CCR7_POS	64	276	23.2
LAG3_POS_CD45RA_POS	128	276	46.4
LAG3_POS_EOMES_POS	116	276	42.0
CCR7_POS_CD45RA_POS	108	276	39.1
CCR7_POS_EOMES_POS	86	276	31.2
CD45RA_POS_EOMES_POS	191	276	69.2

Expand example for expand_num = FALSE and expand_denom = TRUE, and keep_indicators = FALSE

example_perc3 <-
  # Should only count the CD3+ cells
  dplyr::filter(example_intensity_gated, cd3_pos == 1) |>
  getPerc(
    intens_dat = _,
    num_marker = example_markers,
    denom_marker = c("CD4", "CD8"),
    expand_num = FALSE,
    expand_denom = TRUE,
    keep_indicators = FALSE
  )

# For display only, group based on the denominators and
# simplify the names to be numerators
example_perc3 |> 
  tidyr::separate_wider_delim(subpopulation,
    delim = "_OF_",
    names = c("num", "denom"),
    cols_remove = FALSE
  ) |>
  dplyr::mutate(denom = paste("Denom = ", denom)) |>
  dplyr::group_by(denom) |>
  dplyr::select(-subpopulation) |>
  gt::gt() |>
  gt::fmt_number(
    columns = "perc",
    decimals = 1
  )

num	n_num	n_denom	perc
Denom = CD4_NEG_CD8_NEG
LAG3_NEG	196	271	72.3
CCR7_NEG	220	271	81.2
CD45RA_NEG	21	271	7.7
EOMES_NEG	59	271	21.8
LAG3_POS	75	271	27.7
CCR7_POS	51	271	18.8
CD45RA_POS	250	271	92.3
EOMES_POS	212	271	78.2
Denom = CD4_NEG_CD8_POS
LAG3_NEG	1711	2302	74.3
CCR7_NEG	1699	2302	73.8
CD45RA_NEG	413	2302	17.9
EOMES_NEG	648	2302	28.1
LAG3_POS	591	2302	25.7
CCR7_POS	603	2302	26.2
CD45RA_POS	1889	2302	82.1
EOMES_POS	1654	2302	71.9
Denom = CD4_POS_CD8_NEG
LAG3_NEG	8289	8615	96.2
CCR7_NEG	1395	8615	16.2
CD45RA_NEG	1830	8615	21.2
EOMES_NEG	7800	8615	90.5
LAG3_POS	326	8615	3.8
CCR7_POS	7220	8615	83.8
CD45RA_POS	6785	8615	78.8
EOMES_POS	815	8615	9.5
Denom = CD4_POS_CD8_POS
LAG3_NEG	139	276	50.4
CCR7_NEG	152	276	55.1
CD45RA_NEG	35	276	12.7
EOMES_NEG	71	276	25.7
LAG3_POS	137	276	49.6
CCR7_POS	124	276	44.9
CD45RA_POS	241	276	87.3
EOMES_POS	205	276	74.3
Denom = CD4_NEG_CD8_NEG_LAG3_NEG
CCR7_NEG	147	196	75.0
CD45RA_NEG	17	196	8.7
EOMES_NEG	58	196	29.6
CCR7_POS	49	196	25.0
CD45RA_POS	179	196	91.3
EOMES_POS	138	196	70.4
Denom = CD4_NEG_CD8_POS_LAG3_NEG
CCR7_NEG	1165	1711	68.1
CD45RA_NEG	344	1711	20.1
EOMES_NEG	587	1711	34.3
CCR7_POS	546	1711	31.9
CD45RA_POS	1367	1711	79.9
EOMES_POS	1124	1711	65.7
Denom = CD4_POS_CD8_NEG_LAG3_NEG
CCR7_NEG	1221	8289	14.7
CD45RA_NEG	1700	8289	20.5
EOMES_NEG	7627	8289	92.0
CCR7_POS	7068	8289	85.3
CD45RA_POS	6589	8289	79.5
EOMES_POS	662	8289	8.0
Denom = CD4_POS_CD8_POS_LAG3_NEG
CCR7_NEG	79	139	56.8
CD45RA_NEG	26	139	18.7
EOMES_NEG	50	139	36.0
CCR7_POS	60	139	43.2
CD45RA_POS	113	139	81.3
EOMES_POS	89	139	64.0
Denom = CD4_NEG_CD8_NEG_CCR7_NEG
LAG3_NEG	147	220	66.8
CD45RA_NEG	17	220	7.7
EOMES_NEG	20	220	9.1
LAG3_POS	73	220	33.2
CD45RA_POS	203	220	92.3
EOMES_POS	200	220	90.9
Denom = CD4_NEG_CD8_POS_CCR7_NEG
LAG3_NEG	1165	1699	68.6
CD45RA_NEG	329	1699	19.4
EOMES_NEG	258	1699	15.2
LAG3_POS	534	1699	31.4
CD45RA_POS	1370	1699	80.6
EOMES_POS	1441	1699	84.8
Denom = CD4_POS_CD8_NEG_CCR7_NEG
LAG3_NEG	1221	1395	87.5
CD45RA_NEG	927	1395	66.5
EOMES_NEG	811	1395	58.1
LAG3_POS	174	1395	12.5
CD45RA_POS	468	1395	33.5
EOMES_POS	584	1395	41.9
Denom = CD4_POS_CD8_POS_CCR7_NEG
LAG3_NEG	79	152	52.0
CD45RA_NEG	19	152	12.5
EOMES_NEG	33	152	21.7
LAG3_POS	73	152	48.0
CD45RA_POS	133	152	87.5
EOMES_POS	119	152	78.3
Denom = CD4_NEG_CD8_NEG_CD45RA_NEG
LAG3_NEG	17	21	81.0
CCR7_NEG	17	21	81.0
EOMES_NEG	7	21	33.3
LAG3_POS	4	21	19.0
CCR7_POS	4	21	19.0
EOMES_POS	14	21	66.7
Denom = CD4_NEG_CD8_POS_CD45RA_NEG
LAG3_NEG	344	413	83.3
CCR7_NEG	329	413	79.7
EOMES_NEG	125	413	30.3
LAG3_POS	69	413	16.7
CCR7_POS	84	413	20.3
EOMES_POS	288	413	69.7
Denom = CD4_POS_CD8_NEG_CD45RA_NEG
LAG3_NEG	1700	1830	92.9
CCR7_NEG	927	1830	50.7
EOMES_NEG	1461	1830	79.8
LAG3_POS	130	1830	7.1
CCR7_POS	903	1830	49.3
EOMES_POS	369	1830	20.2
Denom = CD4_POS_CD8_POS_CD45RA_NEG
LAG3_NEG	26	35	74.3
CCR7_NEG	19	35	54.3
EOMES_NEG	21	35	60.0
LAG3_POS	9	35	25.7
CCR7_POS	16	35	45.7
EOMES_POS	14	35	40.0
Denom = CD4_NEG_CD8_NEG_EOMES_NEG
LAG3_NEG	58	59	98.3
CCR7_NEG	20	59	33.9
CD45RA_NEG	7	59	11.9
LAG3_POS	1	59	1.7
CCR7_POS	39	59	66.1
CD45RA_POS	52	59	88.1
Denom = CD4_NEG_CD8_POS_EOMES_NEG
LAG3_NEG	587	648	90.6
CCR7_NEG	258	648	39.8
CD45RA_NEG	125	648	19.3
LAG3_POS	61	648	9.4
CCR7_POS	390	648	60.2
CD45RA_POS	523	648	80.7
Denom = CD4_POS_CD8_NEG_EOMES_NEG
LAG3_NEG	7627	7800	97.8
CCR7_NEG	811	7800	10.4
CD45RA_NEG	1461	7800	18.7
LAG3_POS	173	7800	2.2
CCR7_POS	6989	7800	89.6
CD45RA_POS	6339	7800	81.3
Denom = CD4_POS_CD8_POS_EOMES_NEG
LAG3_NEG	50	71	70.4
CCR7_NEG	33	71	46.5
CD45RA_NEG	21	71	29.6
LAG3_POS	21	71	29.6
CCR7_POS	38	71	53.5
CD45RA_POS	50	71	70.4
Denom = CD4_NEG_CD8_NEG_LAG3_POS
CCR7_NEG	73	75	97.3
CD45RA_NEG	4	75	5.3
EOMES_NEG	1	75	1.3
CCR7_POS	2	75	2.7
CD45RA_POS	71	75	94.7
EOMES_POS	74	75	98.7
Denom = CD4_NEG_CD8_POS_LAG3_POS
CCR7_NEG	534	591	90.4
CD45RA_NEG	69	591	11.7
EOMES_NEG	61	591	10.3
CCR7_POS	57	591	9.6
CD45RA_POS	522	591	88.3
EOMES_POS	530	591	89.7
Denom = CD4_POS_CD8_NEG_LAG3_POS
CCR7_NEG	174	326	53.4
CD45RA_NEG	130	326	39.9
EOMES_NEG	173	326	53.1
CCR7_POS	152	326	46.6
CD45RA_POS	196	326	60.1
EOMES_POS	153	326	46.9
Denom = CD4_POS_CD8_POS_LAG3_POS
CCR7_NEG	73	137	53.3
CD45RA_NEG	9	137	6.6
EOMES_NEG	21	137	15.3
CCR7_POS	64	137	46.7
CD45RA_POS	128	137	93.4
EOMES_POS	116	137	84.7
Denom = CD4_NEG_CD8_NEG_CCR7_POS
LAG3_NEG	49	51	96.1
CD45RA_NEG	4	51	7.8
EOMES_NEG	39	51	76.5
LAG3_POS	2	51	3.9
CD45RA_POS	47	51	92.2
EOMES_POS	12	51	23.5
Denom = CD4_NEG_CD8_POS_CCR7_POS
LAG3_NEG	546	603	90.5
CD45RA_NEG	84	603	13.9
EOMES_NEG	390	603	64.7
LAG3_POS	57	603	9.5
CD45RA_POS	519	603	86.1
EOMES_POS	213	603	35.3
Denom = CD4_POS_CD8_NEG_CCR7_POS
LAG3_NEG	7068	7220	97.9
CD45RA_NEG	903	7220	12.5
EOMES_NEG	6989	7220	96.8
LAG3_POS	152	7220	2.1
CD45RA_POS	6317	7220	87.5
EOMES_POS	231	7220	3.2
Denom = CD4_POS_CD8_POS_CCR7_POS
LAG3_NEG	60	124	48.4
CD45RA_NEG	16	124	12.9
EOMES_NEG	38	124	30.6
LAG3_POS	64	124	51.6
CD45RA_POS	108	124	87.1
EOMES_POS	86	124	69.4
Denom = CD4_NEG_CD8_NEG_CD45RA_POS
LAG3_NEG	179	250	71.6
CCR7_NEG	203	250	81.2
EOMES_NEG	52	250	20.8
LAG3_POS	71	250	28.4
CCR7_POS	47	250	18.8
EOMES_POS	198	250	79.2
Denom = CD4_NEG_CD8_POS_CD45RA_POS
LAG3_NEG	1367	1889	72.4
CCR7_NEG	1370	1889	72.5
EOMES_NEG	523	1889	27.7
LAG3_POS	522	1889	27.6
CCR7_POS	519	1889	27.5
EOMES_POS	1366	1889	72.3
Denom = CD4_POS_CD8_NEG_CD45RA_POS
LAG3_NEG	6589	6785	97.1
CCR7_NEG	468	6785	6.9
EOMES_NEG	6339	6785	93.4
LAG3_POS	196	6785	2.9
CCR7_POS	6317	6785	93.1
EOMES_POS	446	6785	6.6
Denom = CD4_POS_CD8_POS_CD45RA_POS
LAG3_NEG	113	241	46.9
CCR7_NEG	133	241	55.2
EOMES_NEG	50	241	20.7
LAG3_POS	128	241	53.1
CCR7_POS	108	241	44.8
EOMES_POS	191	241	79.3
Denom = CD4_NEG_CD8_NEG_EOMES_POS
LAG3_NEG	138	212	65.1
CCR7_NEG	200	212	94.3
CD45RA_NEG	14	212	6.6
LAG3_POS	74	212	34.9
CCR7_POS	12	212	5.7
CD45RA_POS	198	212	93.4
Denom = CD4_NEG_CD8_POS_EOMES_POS
LAG3_NEG	1124	1654	68.0
CCR7_NEG	1441	1654	87.1
CD45RA_NEG	288	1654	17.4
LAG3_POS	530	1654	32.0
CCR7_POS	213	1654	12.9
CD45RA_POS	1366	1654	82.6
Denom = CD4_POS_CD8_NEG_EOMES_POS
LAG3_NEG	662	815	81.2
CCR7_NEG	584	815	71.7
CD45RA_NEG	369	815	45.3
LAG3_POS	153	815	18.8
CCR7_POS	231	815	28.3
CD45RA_POS	446	815	54.7
Denom = CD4_POS_CD8_POS_EOMES_POS
LAG3_NEG	89	205	43.4
CCR7_NEG	119	205	58.0
CD45RA_NEG	14	205	6.8
LAG3_POS	116	205	56.6
CCR7_POS	86	205	42.0
CD45RA_POS	191	205	93.2

Expand example for expand_num = TRUE and expand_denom = TRUE , and keep_indicators = FALSE

example_perc4 <-
  # Should only count the CD3+ cells
  dplyr::filter(example_intensity_gated, cd3_pos == 1) |>
  getPerc(
    intens_dat = _,
    num_marker = example_markers,
    denom_marker = c("CD4", "CD8"),
    expand_num = TRUE,
    expand_denom = TRUE,
    keep_indicators = FALSE
  )

# For display only, group based on the denominators and
# simplify the names to be numerators
example_perc4 |>
  tidyr::separate_wider_delim(subpopulation,
    delim = "_OF_",
    names = c("num", "denom"),
    cols_remove = FALSE
  ) |>
  dplyr::mutate(denom = paste("Denom = ", denom)) |>
  dplyr::group_by(denom) |>
  dplyr::select(-subpopulation) |>
  gt::gt() |>
  gt::fmt_number(
    columns = "perc",
    decimals = 1
  )

num	n_num	n_denom	perc
Denom = CD4_NEG_CD8_NEG
LAG3_NEG	196	271	72.3
CCR7_NEG	220	271	81.2
CD45RA_NEG	21	271	7.7
EOMES_NEG	59	271	21.8
LAG3_POS	75	271	27.7
CCR7_POS	51	271	18.8
CD45RA_POS	250	271	92.3
EOMES_POS	212	271	78.2
LAG3_NEG_CCR7_NEG	147	271	54.2
LAG3_NEG_CD45RA_NEG	17	271	6.3
LAG3_NEG_EOMES_NEG	58	271	21.4
LAG3_NEG_CCR7_POS	49	271	18.1
LAG3_NEG_CD45RA_POS	179	271	66.1
LAG3_NEG_EOMES_POS	138	271	50.9
CCR7_NEG_CD45RA_NEG	17	271	6.3
CCR7_NEG_EOMES_NEG	20	271	7.4
CCR7_NEG_LAG3_POS	73	271	26.9
CCR7_NEG_CD45RA_POS	203	271	74.9
CCR7_NEG_EOMES_POS	200	271	73.8
CD45RA_NEG_EOMES_NEG	7	271	2.6
CD45RA_NEG_LAG3_POS	4	271	1.5
CD45RA_NEG_CCR7_POS	4	271	1.5
CD45RA_NEG_EOMES_POS	14	271	5.2
EOMES_NEG_LAG3_POS	1	271	0.4
EOMES_NEG_CCR7_POS	39	271	14.4
EOMES_NEG_CD45RA_POS	52	271	19.2
LAG3_POS_CCR7_POS	2	271	0.7
LAG3_POS_CD45RA_POS	71	271	26.2
LAG3_POS_EOMES_POS	74	271	27.3
CCR7_POS_CD45RA_POS	47	271	17.3
CCR7_POS_EOMES_POS	12	271	4.4
CD45RA_POS_EOMES_POS	198	271	73.1
Denom = CD4_NEG_CD8_POS
LAG3_NEG	1711	2302	74.3
CCR7_NEG	1699	2302	73.8
CD45RA_NEG	413	2302	17.9
EOMES_NEG	648	2302	28.1
LAG3_POS	591	2302	25.7
CCR7_POS	603	2302	26.2
CD45RA_POS	1889	2302	82.1
EOMES_POS	1654	2302	71.9
LAG3_NEG_CCR7_NEG	1165	2302	50.6
LAG3_NEG_CD45RA_NEG	344	2302	14.9
LAG3_NEG_EOMES_NEG	587	2302	25.5
LAG3_NEG_CCR7_POS	546	2302	23.7
LAG3_NEG_CD45RA_POS	1367	2302	59.4
LAG3_NEG_EOMES_POS	1124	2302	48.8
CCR7_NEG_CD45RA_NEG	329	2302	14.3
CCR7_NEG_EOMES_NEG	258	2302	11.2
CCR7_NEG_LAG3_POS	534	2302	23.2
CCR7_NEG_CD45RA_POS	1370	2302	59.5
CCR7_NEG_EOMES_POS	1441	2302	62.6
CD45RA_NEG_EOMES_NEG	125	2302	5.4
CD45RA_NEG_LAG3_POS	69	2302	3.0
CD45RA_NEG_CCR7_POS	84	2302	3.6
CD45RA_NEG_EOMES_POS	288	2302	12.5
EOMES_NEG_LAG3_POS	61	2302	2.6
EOMES_NEG_CCR7_POS	390	2302	16.9
EOMES_NEG_CD45RA_POS	523	2302	22.7
LAG3_POS_CCR7_POS	57	2302	2.5
LAG3_POS_CD45RA_POS	522	2302	22.7
LAG3_POS_EOMES_POS	530	2302	23.0
CCR7_POS_CD45RA_POS	519	2302	22.5
CCR7_POS_EOMES_POS	213	2302	9.3
CD45RA_POS_EOMES_POS	1366	2302	59.3
Denom = CD4_POS_CD8_NEG
LAG3_NEG	8289	8615	96.2
CCR7_NEG	1395	8615	16.2
CD45RA_NEG	1830	8615	21.2
EOMES_NEG	7800	8615	90.5
LAG3_POS	326	8615	3.8
CCR7_POS	7220	8615	83.8
CD45RA_POS	6785	8615	78.8
EOMES_POS	815	8615	9.5
LAG3_NEG_CCR7_NEG	1221	8615	14.2
LAG3_NEG_CD45RA_NEG	1700	8615	19.7
LAG3_NEG_EOMES_NEG	7627	8615	88.5
LAG3_NEG_CCR7_POS	7068	8615	82.0
LAG3_NEG_CD45RA_POS	6589	8615	76.5
LAG3_NEG_EOMES_POS	662	8615	7.7
CCR7_NEG_CD45RA_NEG	927	8615	10.8
CCR7_NEG_EOMES_NEG	811	8615	9.4
CCR7_NEG_LAG3_POS	174	8615	2.0
CCR7_NEG_CD45RA_POS	468	8615	5.4
CCR7_NEG_EOMES_POS	584	8615	6.8
CD45RA_NEG_EOMES_NEG	1461	8615	17.0
CD45RA_NEG_LAG3_POS	130	8615	1.5
CD45RA_NEG_CCR7_POS	903	8615	10.5
CD45RA_NEG_EOMES_POS	369	8615	4.3
EOMES_NEG_LAG3_POS	173	8615	2.0
EOMES_NEG_CCR7_POS	6989	8615	81.1
EOMES_NEG_CD45RA_POS	6339	8615	73.6
LAG3_POS_CCR7_POS	152	8615	1.8
LAG3_POS_CD45RA_POS	196	8615	2.3
LAG3_POS_EOMES_POS	153	8615	1.8
CCR7_POS_CD45RA_POS	6317	8615	73.3
CCR7_POS_EOMES_POS	231	8615	2.7
CD45RA_POS_EOMES_POS	446	8615	5.2
Denom = CD4_POS_CD8_POS
LAG3_NEG	139	276	50.4
CCR7_NEG	152	276	55.1
CD45RA_NEG	35	276	12.7
EOMES_NEG	71	276	25.7
LAG3_POS	137	276	49.6
CCR7_POS	124	276	44.9
CD45RA_POS	241	276	87.3
EOMES_POS	205	276	74.3
LAG3_NEG_CCR7_NEG	79	276	28.6
LAG3_NEG_CD45RA_NEG	26	276	9.4
LAG3_NEG_EOMES_NEG	50	276	18.1
LAG3_NEG_CCR7_POS	60	276	21.7
LAG3_NEG_CD45RA_POS	113	276	40.9
LAG3_NEG_EOMES_POS	89	276	32.2
CCR7_NEG_CD45RA_NEG	19	276	6.9
CCR7_NEG_EOMES_NEG	33	276	12.0
CCR7_NEG_LAG3_POS	73	276	26.4
CCR7_NEG_CD45RA_POS	133	276	48.2
CCR7_NEG_EOMES_POS	119	276	43.1
CD45RA_NEG_EOMES_NEG	21	276	7.6
CD45RA_NEG_LAG3_POS	9	276	3.3
CD45RA_NEG_CCR7_POS	16	276	5.8
CD45RA_NEG_EOMES_POS	14	276	5.1
EOMES_NEG_LAG3_POS	21	276	7.6
EOMES_NEG_CCR7_POS	38	276	13.8
EOMES_NEG_CD45RA_POS	50	276	18.1
LAG3_POS_CCR7_POS	64	276	23.2
LAG3_POS_CD45RA_POS	128	276	46.4
LAG3_POS_EOMES_POS	116	276	42.0
CCR7_POS_CD45RA_POS	108	276	39.1
CCR7_POS_EOMES_POS	86	276	31.2
CD45RA_POS_EOMES_POS	191	276	69.2
Denom = CD4_NEG_CD8_NEG_LAG3_NEG
CCR7_NEG	147	196	75.0
CD45RA_NEG	17	196	8.7
EOMES_NEG	58	196	29.6
CCR7_POS	49	196	25.0
CD45RA_POS	179	196	91.3
EOMES_POS	138	196	70.4
CCR7_NEG_CD45RA_NEG	15	196	7.7
CCR7_NEG_EOMES_NEG	19	196	9.7
CCR7_NEG_CD45RA_POS	132	196	67.3
CCR7_NEG_EOMES_POS	128	196	65.3
CD45RA_NEG_EOMES_NEG	7	196	3.6
CD45RA_NEG_CCR7_POS	2	196	1.0
CD45RA_NEG_EOMES_POS	10	196	5.1
EOMES_NEG_CCR7_POS	39	196	19.9
EOMES_NEG_CD45RA_POS	51	196	26.0
CCR7_POS_CD45RA_POS	47	196	24.0
CCR7_POS_EOMES_POS	10	196	5.1
CD45RA_POS_EOMES_POS	128	196	65.3
Denom = CD4_NEG_CD8_POS_LAG3_NEG
CCR7_NEG	1165	1711	68.1
CD45RA_NEG	344	1711	20.1
EOMES_NEG	587	1711	34.3
CCR7_POS	546	1711	31.9
CD45RA_POS	1367	1711	79.9
EOMES_POS	1124	1711	65.7
CCR7_NEG_CD45RA_NEG	269	1711	15.7
CCR7_NEG_EOMES_NEG	203	1711	11.9
CCR7_NEG_CD45RA_POS	896	1711	52.4
CCR7_NEG_EOMES_POS	962	1711	56.2
CD45RA_NEG_EOMES_NEG	115	1711	6.7
CD45RA_NEG_CCR7_POS	75	1711	4.4
CD45RA_NEG_EOMES_POS	229	1711	13.4
EOMES_NEG_CCR7_POS	384	1711	22.4
EOMES_NEG_CD45RA_POS	472	1711	27.6
CCR7_POS_CD45RA_POS	471	1711	27.5
CCR7_POS_EOMES_POS	162	1711	9.5
CD45RA_POS_EOMES_POS	895	1711	52.3
Denom = CD4_POS_CD8_NEG_LAG3_NEG
CCR7_NEG	1221	8289	14.7
CD45RA_NEG	1700	8289	20.5
EOMES_NEG	7627	8289	92.0
CCR7_POS	7068	8289	85.3
CD45RA_POS	6589	8289	79.5
EOMES_POS	662	8289	8.0
CCR7_NEG_CD45RA_NEG	837	8289	10.1
CCR7_NEG_EOMES_NEG	736	8289	8.9
CCR7_NEG_CD45RA_POS	384	8289	4.6
CCR7_NEG_EOMES_POS	485	8289	5.9
CD45RA_NEG_EOMES_NEG	1371	8289	16.5
CD45RA_NEG_CCR7_POS	863	8289	10.4
CD45RA_NEG_EOMES_POS	329	8289	4.0
EOMES_NEG_CCR7_POS	6891	8289	83.1
EOMES_NEG_CD45RA_POS	6256	8289	75.5
CCR7_POS_CD45RA_POS	6205	8289	74.9
CCR7_POS_EOMES_POS	177	8289	2.1
CD45RA_POS_EOMES_POS	333	8289	4.0
Denom = CD4_POS_CD8_POS_LAG3_NEG
CCR7_NEG	79	139	56.8
CD45RA_NEG	26	139	18.7
EOMES_NEG	50	139	36.0
CCR7_POS	60	139	43.2
CD45RA_POS	113	139	81.3
EOMES_POS	89	139	64.0
CCR7_NEG_CD45RA_NEG	11	139	7.9
CCR7_NEG_EOMES_NEG	22	139	15.8
CCR7_NEG_CD45RA_POS	68	139	48.9
CCR7_NEG_EOMES_POS	57	139	41.0
CD45RA_NEG_EOMES_NEG	19	139	13.7
CD45RA_NEG_CCR7_POS	15	139	10.8
CD45RA_NEG_EOMES_POS	7	139	5.0
EOMES_NEG_CCR7_POS	28	139	20.1
EOMES_NEG_CD45RA_POS	31	139	22.3
CCR7_POS_CD45RA_POS	45	139	32.4
CCR7_POS_EOMES_POS	32	139	23.0
CD45RA_POS_EOMES_POS	82	139	59.0
Denom = CD4_NEG_CD8_NEG_CCR7_NEG
LAG3_NEG	147	220	66.8
CD45RA_NEG	17	220	7.7
EOMES_NEG	20	220	9.1
LAG3_POS	73	220	33.2
CD45RA_POS	203	220	92.3
EOMES_POS	200	220	90.9
LAG3_NEG_CD45RA_NEG	15	220	6.8
LAG3_NEG_EOMES_NEG	19	220	8.6
LAG3_NEG_CD45RA_POS	132	220	60.0
LAG3_NEG_EOMES_POS	128	220	58.2
CD45RA_NEG_EOMES_NEG	6	220	2.7
CD45RA_NEG_LAG3_POS	2	220	0.9
CD45RA_NEG_EOMES_POS	11	220	5.0
EOMES_NEG_LAG3_POS	1	220	0.5
EOMES_NEG_CD45RA_POS	14	220	6.4
LAG3_POS_CD45RA_POS	71	220	32.3
LAG3_POS_EOMES_POS	72	220	32.7
CD45RA_POS_EOMES_POS	189	220	85.9
Denom = CD4_NEG_CD8_POS_CCR7_NEG
LAG3_NEG	1165	1699	68.6
CD45RA_NEG	329	1699	19.4
EOMES_NEG	258	1699	15.2
LAG3_POS	534	1699	31.4
CD45RA_POS	1370	1699	80.6
EOMES_POS	1441	1699	84.8
LAG3_NEG_CD45RA_NEG	269	1699	15.8
LAG3_NEG_EOMES_NEG	203	1699	11.9
LAG3_NEG_CD45RA_POS	896	1699	52.7
LAG3_NEG_EOMES_POS	962	1699	56.6
CD45RA_NEG_EOMES_NEG	81	1699	4.8
CD45RA_NEG_LAG3_POS	60	1699	3.5
CD45RA_NEG_EOMES_POS	248	1699	14.6
EOMES_NEG_LAG3_POS	55	1699	3.2
EOMES_NEG_CD45RA_POS	177	1699	10.4
LAG3_POS_CD45RA_POS	474	1699	27.9
LAG3_POS_EOMES_POS	479	1699	28.2
CD45RA_POS_EOMES_POS	1193	1699	70.2
Denom = CD4_POS_CD8_NEG_CCR7_NEG
LAG3_NEG	1221	1395	87.5
CD45RA_NEG	927	1395	66.5
EOMES_NEG	811	1395	58.1
LAG3_POS	174	1395	12.5
CD45RA_POS	468	1395	33.5
EOMES_POS	584	1395	41.9
LAG3_NEG_CD45RA_NEG	837	1395	60.0
LAG3_NEG_EOMES_NEG	736	1395	52.8
LAG3_NEG_CD45RA_POS	384	1395	27.5
LAG3_NEG_EOMES_POS	485	1395	34.8
CD45RA_NEG_EOMES_NEG	642	1395	46.0
CD45RA_NEG_LAG3_POS	90	1395	6.5
CD45RA_NEG_EOMES_POS	285	1395	20.4
EOMES_NEG_LAG3_POS	75	1395	5.4
EOMES_NEG_CD45RA_POS	169	1395	12.1
LAG3_POS_CD45RA_POS	84	1395	6.0
LAG3_POS_EOMES_POS	99	1395	7.1
CD45RA_POS_EOMES_POS	299	1395	21.4
Denom = CD4_POS_CD8_POS_CCR7_NEG
LAG3_NEG	79	152	52.0
CD45RA_NEG	19	152	12.5
EOMES_NEG	33	152	21.7
LAG3_POS	73	152	48.0
CD45RA_POS	133	152	87.5
EOMES_POS	119	152	78.3
LAG3_NEG_CD45RA_NEG	11	152	7.2
LAG3_NEG_EOMES_NEG	22	152	14.5
LAG3_NEG_CD45RA_POS	68	152	44.7
LAG3_NEG_EOMES_POS	57	152	37.5
CD45RA_NEG_EOMES_NEG	9	152	5.9
CD45RA_NEG_LAG3_POS	8	152	5.3
CD45RA_NEG_EOMES_POS	10	152	6.6
EOMES_NEG_LAG3_POS	11	152	7.2
EOMES_NEG_CD45RA_POS	24	152	15.8
LAG3_POS_CD45RA_POS	65	152	42.8
LAG3_POS_EOMES_POS	62	152	40.8
CD45RA_POS_EOMES_POS	109	152	71.7
Denom = CD4_NEG_CD8_NEG_CD45RA_NEG
LAG3_NEG	17	21	81.0
CCR7_NEG	17	21	81.0
EOMES_NEG	7	21	33.3
LAG3_POS	4	21	19.0
CCR7_POS	4	21	19.0
EOMES_POS	14	21	66.7
LAG3_NEG_CCR7_NEG	15	21	71.4
LAG3_NEG_EOMES_NEG	7	21	33.3
LAG3_NEG_CCR7_POS	2	21	9.5
LAG3_NEG_EOMES_POS	10	21	47.6
CCR7_NEG_EOMES_NEG	6	21	28.6
CCR7_NEG_LAG3_POS	2	21	9.5
CCR7_NEG_EOMES_POS	11	21	52.4
EOMES_NEG_LAG3_POS	0	21	0.0
EOMES_NEG_CCR7_POS	1	21	4.8
LAG3_POS_CCR7_POS	2	21	9.5
LAG3_POS_EOMES_POS	4	21	19.0
CCR7_POS_EOMES_POS	3	21	14.3
Denom = CD4_NEG_CD8_POS_CD45RA_NEG
LAG3_NEG	344	413	83.3
CCR7_NEG	329	413	79.7
EOMES_NEG	125	413	30.3
LAG3_POS	69	413	16.7
CCR7_POS	84	413	20.3
EOMES_POS	288	413	69.7
LAG3_NEG_CCR7_NEG	269	413	65.1
LAG3_NEG_EOMES_NEG	115	413	27.8
LAG3_NEG_CCR7_POS	75	413	18.2
LAG3_NEG_EOMES_POS	229	413	55.4
CCR7_NEG_EOMES_NEG	81	413	19.6
CCR7_NEG_LAG3_POS	60	413	14.5
CCR7_NEG_EOMES_POS	248	413	60.0
EOMES_NEG_LAG3_POS	10	413	2.4
EOMES_NEG_CCR7_POS	44	413	10.7
LAG3_POS_CCR7_POS	9	413	2.2
LAG3_POS_EOMES_POS	59	413	14.3
CCR7_POS_EOMES_POS	40	413	9.7
Denom = CD4_POS_CD8_NEG_CD45RA_NEG
LAG3_NEG	1700	1830	92.9
CCR7_NEG	927	1830	50.7
EOMES_NEG	1461	1830	79.8
LAG3_POS	130	1830	7.1
CCR7_POS	903	1830	49.3
EOMES_POS	369	1830	20.2
LAG3_NEG_CCR7_NEG	837	1830	45.7
LAG3_NEG_EOMES_NEG	1371	1830	74.9
LAG3_NEG_CCR7_POS	863	1830	47.2
LAG3_NEG_EOMES_POS	329	1830	18.0
CCR7_NEG_EOMES_NEG	642	1830	35.1
CCR7_NEG_LAG3_POS	90	1830	4.9
CCR7_NEG_EOMES_POS	285	1830	15.6
EOMES_NEG_LAG3_POS	90	1830	4.9
EOMES_NEG_CCR7_POS	819	1830	44.8
LAG3_POS_CCR7_POS	40	1830	2.2
LAG3_POS_EOMES_POS	40	1830	2.2
CCR7_POS_EOMES_POS	84	1830	4.6
Denom = CD4_POS_CD8_POS_CD45RA_NEG
LAG3_NEG	26	35	74.3
CCR7_NEG	19	35	54.3
EOMES_NEG	21	35	60.0
LAG3_POS	9	35	25.7
CCR7_POS	16	35	45.7
EOMES_POS	14	35	40.0
LAG3_NEG_CCR7_NEG	11	35	31.4
LAG3_NEG_EOMES_NEG	19	35	54.3
LAG3_NEG_CCR7_POS	15	35	42.9
LAG3_NEG_EOMES_POS	7	35	20.0
CCR7_NEG_EOMES_NEG	9	35	25.7
CCR7_NEG_LAG3_POS	8	35	22.9
CCR7_NEG_EOMES_POS	10	35	28.6
EOMES_NEG_LAG3_POS	2	35	5.7
EOMES_NEG_CCR7_POS	12	35	34.3
LAG3_POS_CCR7_POS	1	35	2.9
LAG3_POS_EOMES_POS	7	35	20.0
CCR7_POS_EOMES_POS	4	35	11.4
Denom = CD4_NEG_CD8_NEG_EOMES_NEG
LAG3_NEG	58	59	98.3
CCR7_NEG	20	59	33.9
CD45RA_NEG	7	59	11.9
LAG3_POS	1	59	1.7
CCR7_POS	39	59	66.1
CD45RA_POS	52	59	88.1
LAG3_NEG_CCR7_NEG	19	59	32.2
LAG3_NEG_CD45RA_NEG	7	59	11.9
LAG3_NEG_CCR7_POS	39	59	66.1
LAG3_NEG_CD45RA_POS	51	59	86.4
CCR7_NEG_CD45RA_NEG	6	59	10.2
CCR7_NEG_LAG3_POS	1	59	1.7
CCR7_NEG_CD45RA_POS	14	59	23.7
CD45RA_NEG_LAG3_POS	0	59	0.0
CD45RA_NEG_CCR7_POS	1	59	1.7
LAG3_POS_CCR7_POS	0	59	0.0
LAG3_POS_CD45RA_POS	1	59	1.7
CCR7_POS_CD45RA_POS	38	59	64.4
Denom = CD4_NEG_CD8_POS_EOMES_NEG
LAG3_NEG	587	648	90.6
CCR7_NEG	258	648	39.8
CD45RA_NEG	125	648	19.3
LAG3_POS	61	648	9.4
CCR7_POS	390	648	60.2
CD45RA_POS	523	648	80.7
LAG3_NEG_CCR7_NEG	203	648	31.3
LAG3_NEG_CD45RA_NEG	115	648	17.7
LAG3_NEG_CCR7_POS	384	648	59.3
LAG3_NEG_CD45RA_POS	472	648	72.8
CCR7_NEG_CD45RA_NEG	81	648	12.5
CCR7_NEG_LAG3_POS	55	648	8.5
CCR7_NEG_CD45RA_POS	177	648	27.3
CD45RA_NEG_LAG3_POS	10	648	1.5
CD45RA_NEG_CCR7_POS	44	648	6.8
LAG3_POS_CCR7_POS	6	648	0.9
LAG3_POS_CD45RA_POS	51	648	7.9
CCR7_POS_CD45RA_POS	346	648	53.4
Denom = CD4_POS_CD8_NEG_EOMES_NEG
LAG3_NEG	7627	7800	97.8
CCR7_NEG	811	7800	10.4
CD45RA_NEG	1461	7800	18.7
LAG3_POS	173	7800	2.2
CCR7_POS	6989	7800	89.6
CD45RA_POS	6339	7800	81.3
LAG3_NEG_CCR7_NEG	736	7800	9.4
LAG3_NEG_CD45RA_NEG	1371	7800	17.6
LAG3_NEG_CCR7_POS	6891	7800	88.3
LAG3_NEG_CD45RA_POS	6256	7800	80.2
CCR7_NEG_CD45RA_NEG	642	7800	8.2
CCR7_NEG_LAG3_POS	75	7800	1.0
CCR7_NEG_CD45RA_POS	169	7800	2.2
CD45RA_NEG_LAG3_POS	90	7800	1.2
CD45RA_NEG_CCR7_POS	819	7800	10.5
LAG3_POS_CCR7_POS	98	7800	1.3
LAG3_POS_CD45RA_POS	83	7800	1.1
CCR7_POS_CD45RA_POS	6170	7800	79.1
Denom = CD4_POS_CD8_POS_EOMES_NEG
LAG3_NEG	50	71	70.4
CCR7_NEG	33	71	46.5
CD45RA_NEG	21	71	29.6
LAG3_POS	21	71	29.6
CCR7_POS	38	71	53.5
CD45RA_POS	50	71	70.4
LAG3_NEG_CCR7_NEG	22	71	31.0
LAG3_NEG_CD45RA_NEG	19	71	26.8
LAG3_NEG_CCR7_POS	28	71	39.4
LAG3_NEG_CD45RA_POS	31	71	43.7
CCR7_NEG_CD45RA_NEG	9	71	12.7
CCR7_NEG_LAG3_POS	11	71	15.5
CCR7_NEG_CD45RA_POS	24	71	33.8
CD45RA_NEG_LAG3_POS	2	71	2.8
CD45RA_NEG_CCR7_POS	12	71	16.9
LAG3_POS_CCR7_POS	10	71	14.1
LAG3_POS_CD45RA_POS	19	71	26.8
CCR7_POS_CD45RA_POS	26	71	36.6
Denom = CD4_NEG_CD8_NEG_LAG3_POS
CCR7_NEG	73	75	97.3
CD45RA_NEG	4	75	5.3
EOMES_NEG	1	75	1.3
CCR7_POS	2	75	2.7
CD45RA_POS	71	75	94.7
EOMES_POS	74	75	98.7
CCR7_NEG_CD45RA_NEG	2	75	2.7
CCR7_NEG_EOMES_NEG	1	75	1.3
CCR7_NEG_CD45RA_POS	71	75	94.7
CCR7_NEG_EOMES_POS	72	75	96.0
CD45RA_NEG_EOMES_NEG	0	75	0.0
CD45RA_NEG_CCR7_POS	2	75	2.7
CD45RA_NEG_EOMES_POS	4	75	5.3
EOMES_NEG_CCR7_POS	0	75	0.0
EOMES_NEG_CD45RA_POS	1	75	1.3
CCR7_POS_CD45RA_POS	0	75	0.0
CCR7_POS_EOMES_POS	2	75	2.7
CD45RA_POS_EOMES_POS	70	75	93.3
Denom = CD4_NEG_CD8_POS_LAG3_POS
CCR7_NEG	534	591	90.4
CD45RA_NEG	69	591	11.7
EOMES_NEG	61	591	10.3
CCR7_POS	57	591	9.6
CD45RA_POS	522	591	88.3
EOMES_POS	530	591	89.7
CCR7_NEG_CD45RA_NEG	60	591	10.2
CCR7_NEG_EOMES_NEG	55	591	9.3
CCR7_NEG_CD45RA_POS	474	591	80.2
CCR7_NEG_EOMES_POS	479	591	81.0
CD45RA_NEG_EOMES_NEG	10	591	1.7
CD45RA_NEG_CCR7_POS	9	591	1.5
CD45RA_NEG_EOMES_POS	59	591	10.0
EOMES_NEG_CCR7_POS	6	591	1.0
EOMES_NEG_CD45RA_POS	51	591	8.6
CCR7_POS_CD45RA_POS	48	591	8.1
CCR7_POS_EOMES_POS	51	591	8.6
CD45RA_POS_EOMES_POS	471	591	79.7
Denom = CD4_POS_CD8_NEG_LAG3_POS
CCR7_NEG	174	326	53.4
CD45RA_NEG	130	326	39.9
EOMES_NEG	173	326	53.1
CCR7_POS	152	326	46.6
CD45RA_POS	196	326	60.1
EOMES_POS	153	326	46.9
CCR7_NEG_CD45RA_NEG	90	326	27.6
CCR7_NEG_EOMES_NEG	75	326	23.0
CCR7_NEG_CD45RA_POS	84	326	25.8
CCR7_NEG_EOMES_POS	99	326	30.4
CD45RA_NEG_EOMES_NEG	90	326	27.6
CD45RA_NEG_CCR7_POS	40	326	12.3
CD45RA_NEG_EOMES_POS	40	326	12.3
EOMES_NEG_CCR7_POS	98	326	30.1
EOMES_NEG_CD45RA_POS	83	326	25.5
CCR7_POS_CD45RA_POS	112	326	34.4
CCR7_POS_EOMES_POS	54	326	16.6
CD45RA_POS_EOMES_POS	113	326	34.7
Denom = CD4_POS_CD8_POS_LAG3_POS
CCR7_NEG	73	137	53.3
CD45RA_NEG	9	137	6.6
EOMES_NEG	21	137	15.3
CCR7_POS	64	137	46.7
CD45RA_POS	128	137	93.4
EOMES_POS	116	137	84.7
CCR7_NEG_CD45RA_NEG	8	137	5.8
CCR7_NEG_EOMES_NEG	11	137	8.0
CCR7_NEG_CD45RA_POS	65	137	47.4
CCR7_NEG_EOMES_POS	62	137	45.3
CD45RA_NEG_EOMES_NEG	2	137	1.5
CD45RA_NEG_CCR7_POS	1	137	0.7
CD45RA_NEG_EOMES_POS	7	137	5.1
EOMES_NEG_CCR7_POS	10	137	7.3
EOMES_NEG_CD45RA_POS	19	137	13.9
CCR7_POS_CD45RA_POS	63	137	46.0
CCR7_POS_EOMES_POS	54	137	39.4
CD45RA_POS_EOMES_POS	109	137	79.6
Denom = CD4_NEG_CD8_NEG_CCR7_POS
LAG3_NEG	49	51	96.1
CD45RA_NEG	4	51	7.8
EOMES_NEG	39	51	76.5
LAG3_POS	2	51	3.9
CD45RA_POS	47	51	92.2
EOMES_POS	12	51	23.5
LAG3_NEG_CD45RA_NEG	2	51	3.9
LAG3_NEG_EOMES_NEG	39	51	76.5
LAG3_NEG_CD45RA_POS	47	51	92.2
LAG3_NEG_EOMES_POS	10	51	19.6
CD45RA_NEG_EOMES_NEG	1	51	2.0
CD45RA_NEG_LAG3_POS	2	51	3.9
CD45RA_NEG_EOMES_POS	3	51	5.9
EOMES_NEG_LAG3_POS	0	51	0.0
EOMES_NEG_CD45RA_POS	38	51	74.5
LAG3_POS_CD45RA_POS	0	51	0.0
LAG3_POS_EOMES_POS	2	51	3.9
CD45RA_POS_EOMES_POS	9	51	17.6
Denom = CD4_NEG_CD8_POS_CCR7_POS
LAG3_NEG	546	603	90.5
CD45RA_NEG	84	603	13.9
EOMES_NEG	390	603	64.7
LAG3_POS	57	603	9.5
CD45RA_POS	519	603	86.1
EOMES_POS	213	603	35.3
LAG3_NEG_CD45RA_NEG	75	603	12.4
LAG3_NEG_EOMES_NEG	384	603	63.7
LAG3_NEG_CD45RA_POS	471	603	78.1
LAG3_NEG_EOMES_POS	162	603	26.9
CD45RA_NEG_EOMES_NEG	44	603	7.3
CD45RA_NEG_LAG3_POS	9	603	1.5
CD45RA_NEG_EOMES_POS	40	603	6.6
EOMES_NEG_LAG3_POS	6	603	1.0
EOMES_NEG_CD45RA_POS	346	603	57.4
LAG3_POS_CD45RA_POS	48	603	8.0
LAG3_POS_EOMES_POS	51	603	8.5
CD45RA_POS_EOMES_POS	173	603	28.7
Denom = CD4_POS_CD8_NEG_CCR7_POS
LAG3_NEG	7068	7220	97.9
CD45RA_NEG	903	7220	12.5
EOMES_NEG	6989	7220	96.8
LAG3_POS	152	7220	2.1
CD45RA_POS	6317	7220	87.5
EOMES_POS	231	7220	3.2
LAG3_NEG_CD45RA_NEG	863	7220	12.0
LAG3_NEG_EOMES_NEG	6891	7220	95.4
LAG3_NEG_CD45RA_POS	6205	7220	85.9
LAG3_NEG_EOMES_POS	177	7220	2.5
CD45RA_NEG_EOMES_NEG	819	7220	11.3
CD45RA_NEG_LAG3_POS	40	7220	0.6
CD45RA_NEG_EOMES_POS	84	7220	1.2
EOMES_NEG_LAG3_POS	98	7220	1.4
EOMES_NEG_CD45RA_POS	6170	7220	85.5
LAG3_POS_CD45RA_POS	112	7220	1.6
LAG3_POS_EOMES_POS	54	7220	0.7
CD45RA_POS_EOMES_POS	147	7220	2.0
Denom = CD4_POS_CD8_POS_CCR7_POS
LAG3_NEG	60	124	48.4
CD45RA_NEG	16	124	12.9
EOMES_NEG	38	124	30.6
LAG3_POS	64	124	51.6
CD45RA_POS	108	124	87.1
EOMES_POS	86	124	69.4
LAG3_NEG_CD45RA_NEG	15	124	12.1
LAG3_NEG_EOMES_NEG	28	124	22.6
LAG3_NEG_CD45RA_POS	45	124	36.3
LAG3_NEG_EOMES_POS	32	124	25.8
CD45RA_NEG_EOMES_NEG	12	124	9.7
CD45RA_NEG_LAG3_POS	1	124	0.8
CD45RA_NEG_EOMES_POS	4	124	3.2
EOMES_NEG_LAG3_POS	10	124	8.1
EOMES_NEG_CD45RA_POS	26	124	21.0
LAG3_POS_CD45RA_POS	63	124	50.8
LAG3_POS_EOMES_POS	54	124	43.5
CD45RA_POS_EOMES_POS	82	124	66.1
Denom = CD4_NEG_CD8_NEG_CD45RA_POS
LAG3_NEG	179	250	71.6
CCR7_NEG	203	250	81.2
EOMES_NEG	52	250	20.8
LAG3_POS	71	250	28.4
CCR7_POS	47	250	18.8
EOMES_POS	198	250	79.2
LAG3_NEG_CCR7_NEG	132	250	52.8
LAG3_NEG_EOMES_NEG	51	250	20.4
LAG3_NEG_CCR7_POS	47	250	18.8
LAG3_NEG_EOMES_POS	128	250	51.2
CCR7_NEG_EOMES_NEG	14	250	5.6
CCR7_NEG_LAG3_POS	71	250	28.4
CCR7_NEG_EOMES_POS	189	250	75.6
EOMES_NEG_LAG3_POS	1	250	0.4
EOMES_NEG_CCR7_POS	38	250	15.2
LAG3_POS_CCR7_POS	0	250	0.0
LAG3_POS_EOMES_POS	70	250	28.0
CCR7_POS_EOMES_POS	9	250	3.6
Denom = CD4_NEG_CD8_POS_CD45RA_POS
LAG3_NEG	1367	1889	72.4
CCR7_NEG	1370	1889	72.5
EOMES_NEG	523	1889	27.7
LAG3_POS	522	1889	27.6
CCR7_POS	519	1889	27.5
EOMES_POS	1366	1889	72.3
LAG3_NEG_CCR7_NEG	896	1889	47.4
LAG3_NEG_EOMES_NEG	472	1889	25.0
LAG3_NEG_CCR7_POS	471	1889	24.9
LAG3_NEG_EOMES_POS	895	1889	47.4
CCR7_NEG_EOMES_NEG	177	1889	9.4
CCR7_NEG_LAG3_POS	474	1889	25.1
CCR7_NEG_EOMES_POS	1193	1889	63.2
EOMES_NEG_LAG3_POS	51	1889	2.7
EOMES_NEG_CCR7_POS	346	1889	18.3
LAG3_POS_CCR7_POS	48	1889	2.5
LAG3_POS_EOMES_POS	471	1889	24.9
CCR7_POS_EOMES_POS	173	1889	9.2
Denom = CD4_POS_CD8_NEG_CD45RA_POS
LAG3_NEG	6589	6785	97.1
CCR7_NEG	468	6785	6.9
EOMES_NEG	6339	6785	93.4
LAG3_POS	196	6785	2.9
CCR7_POS	6317	6785	93.1
EOMES_POS	446	6785	6.6
LAG3_NEG_CCR7_NEG	384	6785	5.7
LAG3_NEG_EOMES_NEG	6256	6785	92.2
LAG3_NEG_CCR7_POS	6205	6785	91.5
LAG3_NEG_EOMES_POS	333	6785	4.9
CCR7_NEG_EOMES_NEG	169	6785	2.5
CCR7_NEG_LAG3_POS	84	6785	1.2
CCR7_NEG_EOMES_POS	299	6785	4.4
EOMES_NEG_LAG3_POS	83	6785	1.2
EOMES_NEG_CCR7_POS	6170	6785	90.9
LAG3_POS_CCR7_POS	112	6785	1.7
LAG3_POS_EOMES_POS	113	6785	1.7
CCR7_POS_EOMES_POS	147	6785	2.2
Denom = CD4_POS_CD8_POS_CD45RA_POS
LAG3_NEG	113	241	46.9
CCR7_NEG	133	241	55.2
EOMES_NEG	50	241	20.7
LAG3_POS	128	241	53.1
CCR7_POS	108	241	44.8
EOMES_POS	191	241	79.3
LAG3_NEG_CCR7_NEG	68	241	28.2
LAG3_NEG_EOMES_NEG	31	241	12.9
LAG3_NEG_CCR7_POS	45	241	18.7
LAG3_NEG_EOMES_POS	82	241	34.0
CCR7_NEG_EOMES_NEG	24	241	10.0
CCR7_NEG_LAG3_POS	65	241	27.0
CCR7_NEG_EOMES_POS	109	241	45.2
EOMES_NEG_LAG3_POS	19	241	7.9
EOMES_NEG_CCR7_POS	26	241	10.8
LAG3_POS_CCR7_POS	63	241	26.1
LAG3_POS_EOMES_POS	109	241	45.2
CCR7_POS_EOMES_POS	82	241	34.0
Denom = CD4_NEG_CD8_NEG_EOMES_POS
LAG3_NEG	138	212	65.1
CCR7_NEG	200	212	94.3
CD45RA_NEG	14	212	6.6
LAG3_POS	74	212	34.9
CCR7_POS	12	212	5.7
CD45RA_POS	198	212	93.4
LAG3_NEG_CCR7_NEG	128	212	60.4
LAG3_NEG_CD45RA_NEG	10	212	4.7
LAG3_NEG_CCR7_POS	10	212	4.7
LAG3_NEG_CD45RA_POS	128	212	60.4
CCR7_NEG_CD45RA_NEG	11	212	5.2
CCR7_NEG_LAG3_POS	72	212	34.0
CCR7_NEG_CD45RA_POS	189	212	89.2
CD45RA_NEG_LAG3_POS	4	212	1.9
CD45RA_NEG_CCR7_POS	3	212	1.4
LAG3_POS_CCR7_POS	2	212	0.9
LAG3_POS_CD45RA_POS	70	212	33.0
CCR7_POS_CD45RA_POS	9	212	4.2
Denom = CD4_NEG_CD8_POS_EOMES_POS
LAG3_NEG	1124	1654	68.0
CCR7_NEG	1441	1654	87.1
CD45RA_NEG	288	1654	17.4
LAG3_POS	530	1654	32.0
CCR7_POS	213	1654	12.9
CD45RA_POS	1366	1654	82.6
LAG3_NEG_CCR7_NEG	962	1654	58.2
LAG3_NEG_CD45RA_NEG	229	1654	13.8
LAG3_NEG_CCR7_POS	162	1654	9.8
LAG3_NEG_CD45RA_POS	895	1654	54.1
CCR7_NEG_CD45RA_NEG	248	1654	15.0
CCR7_NEG_LAG3_POS	479	1654	29.0
CCR7_NEG_CD45RA_POS	1193	1654	72.1
CD45RA_NEG_LAG3_POS	59	1654	3.6
CD45RA_NEG_CCR7_POS	40	1654	2.4
LAG3_POS_CCR7_POS	51	1654	3.1
LAG3_POS_CD45RA_POS	471	1654	28.5
CCR7_POS_CD45RA_POS	173	1654	10.5
Denom = CD4_POS_CD8_NEG_EOMES_POS
LAG3_NEG	662	815	81.2
CCR7_NEG	584	815	71.7
CD45RA_NEG	369	815	45.3
LAG3_POS	153	815	18.8
CCR7_POS	231	815	28.3
CD45RA_POS	446	815	54.7
LAG3_NEG_CCR7_NEG	485	815	59.5
LAG3_NEG_CD45RA_NEG	329	815	40.4
LAG3_NEG_CCR7_POS	177	815	21.7
LAG3_NEG_CD45RA_POS	333	815	40.9
CCR7_NEG_CD45RA_NEG	285	815	35.0
CCR7_NEG_LAG3_POS	99	815	12.1
CCR7_NEG_CD45RA_POS	299	815	36.7
CD45RA_NEG_LAG3_POS	40	815	4.9
CD45RA_NEG_CCR7_POS	84	815	10.3
LAG3_POS_CCR7_POS	54	815	6.6
LAG3_POS_CD45RA_POS	113	815	13.9
CCR7_POS_CD45RA_POS	147	815	18.0
Denom = CD4_POS_CD8_POS_EOMES_POS
LAG3_NEG	89	205	43.4
CCR7_NEG	119	205	58.0
CD45RA_NEG	14	205	6.8
LAG3_POS	116	205	56.6
CCR7_POS	86	205	42.0
CD45RA_POS	191	205	93.2
LAG3_NEG_CCR7_NEG	57	205	27.8
LAG3_NEG_CD45RA_NEG	7	205	3.4
LAG3_NEG_CCR7_POS	32	205	15.6
LAG3_NEG_CD45RA_POS	82	205	40.0
CCR7_NEG_CD45RA_NEG	10	205	4.9
CCR7_NEG_LAG3_POS	62	205	30.2
CCR7_NEG_CD45RA_POS	109	205	53.2
CD45RA_NEG_LAG3_POS	7	205	3.4
CD45RA_NEG_CCR7_POS	4	205	2.0
LAG3_POS_CCR7_POS	54	205	26.3
LAG3_POS_CD45RA_POS	109	205	53.2
CCR7_POS_CD45RA_POS	82	205	40.0

Optional: Adding density gates back to GatingSet

Let’s add the gate for LAG3 of CD4+ and CD8+. This is a good visualization to see all sequential gating steps applied to the sample.

# Grab gate as a numeric
current_gate <-
  dens_gates |>
  dplyr::filter(cd4_pos_cd8_pos == "cd4_neg_cd8_pos") |>
  dplyr::pull(LAG3)

# Apply using gs_add_gating-method and
# We want a boundary gate
openCyto::gs_add_gating_method(
  gs_qc,
  alias = "lag3_cd8",
  pop = "+",
  parent = "cd4-cd8+",
  dims = "LAG3",
  gating_method = "boundary",
  gating_args = list(min = current_gate, max = Inf)
)

current_gate <-
  dens_gates |>
  dplyr::filter(cd4_pos_cd8_pos == "cd4_pos_cd8_neg") |>
  dplyr::pull(LAG3)

openCyto::gs_add_gating_method(
  gs_qc,
  alias = "lag3_cd4",
  pop = "+",
  parent = "cd4+cd8-",
  dims = "LAG3",
  gating_method = "boundary",
  gating_args = list(min = current_gate, max = Inf)
)

ggcyto::autoplot(gs_qc[[1]])

After running {staRgate} to gate flow cytometry data, it is recommended to perform some quality checks (QC) on the gate placements to ensure they are reasonable. We suggest to use ridgeplots in addition to the ggcyto::autoplot function to visualize the density distributions per marker across samples. When examining a large batch of samples, downsampling, such as to a random sample of 10k CD3+ cells, will make the QC process more manageable. In addition, random spot checks of a few samples would also be helpful QC to detect any edge cases.

Currently in this tutorial, we do not extend to the QC steps and suggest to lean on other examples for how to put together a ridgeplot for example.

In the near future, we hope to incorporate some examples for the additional QC steps as well, stay tuned!

Session info

sessionInfo()
#> R version 4.4.2 (2024-10-31)
#> Platform: x86_64-pc-linux-gnu
#> Running under: Ubuntu 22.04.5 LTS
#> 
#> Matrix products: default
#> BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3 
#> LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.20.so;  LAPACK version 3.10.0
#> 
#> locale:
#>  [1] LC_CTYPE=C.UTF-8       LC_NUMERIC=C           LC_TIME=C.UTF-8       
#>  [4] LC_COLLATE=C.UTF-8     LC_MONETARY=C.UTF-8    LC_MESSAGES=C.UTF-8   
#>  [7] LC_PAPER=C.UTF-8       LC_NAME=C              LC_ADDRESS=C          
#> [10] LC_TELEPHONE=C         LC_MEASUREMENT=C.UTF-8 LC_IDENTIFICATION=C   
#> 
#> time zone: UTC
#> tzcode source: system (glibc)
#> 
#> attached base packages:
#> [1] stats     graphics  grDevices utils     datasets  methods   base     
#> 
#> other attached packages:
#> [1] ggplot2_3.5.1        flowCore_2.18.0      flowWorkspace_4.18.0
#> [4] openCyto_2.18.0      staRgate_0.99.3      BiocStyle_2.34.0    
#> 
#> loaded via a namespace (and not attached):
#>  [1] gtable_0.3.6        xfun_0.49           bslib_0.8.0        
#>  [4] htmlwidgets_1.6.4   lattice_0.22-6      Biobase_2.66.0     
#>  [7] vctrs_0.6.5         tools_4.4.2         generics_0.1.3     
#> [10] stats4_4.4.2        parallel_4.4.2      tibble_3.2.1       
#> [13] fansi_1.0.6         pkgconfig_2.0.3     data.table_1.16.2  
#> [16] RColorBrewer_1.1-3  desc_1.4.3          S4Vectors_0.44.0   
#> [19] gt_0.11.1           graph_1.84.0        lifecycle_1.0.4    
#> [22] farver_2.1.2        compiler_4.4.2      stringr_1.5.1      
#> [25] textshaping_0.4.0   munsell_0.5.1       janitor_2.2.0      
#> [28] snakecase_0.11.1    htmltools_0.5.8.1   sass_0.4.9         
#> [31] yaml_2.3.10         hexbin_1.28.5       pillar_1.9.0       
#> [34] pkgdown_2.1.1       jquerylib_0.1.4     tidyr_1.3.1        
#> [37] cachem_1.1.0        RProtoBufLib_2.18.0 commonmark_1.9.2   
#> [40] tidyselect_1.2.1    digest_0.6.37       stringi_1.8.4      
#> [43] dplyr_1.1.4         purrr_1.0.2         bookdown_0.41      
#> [46] labeling_0.4.3      flowClust_3.44.0    fastmap_1.2.0      
#> [49] grid_4.4.2          colorspace_2.1-1    cli_3.6.3          
#> [52] magrittr_2.0.3      ncdfFlow_2.52.0     RBGL_1.82.0        
#> [55] XML_3.99-0.17       utf8_1.2.4          withr_3.0.2        
#> [58] scales_1.3.0        lubridate_1.9.3     timechange_0.3.0   
#> [61] rmarkdown_2.29      matrixStats_1.4.1   gridExtra_2.3      
#> [64] cytolib_2.18.0      ragg_1.3.3          evaluate_1.0.1     
#> [67] knitr_1.49          markdown_1.13       rlang_1.1.4        
#> [70] Rcpp_1.0.13-1       glue_1.8.0          xml2_1.3.6         
#> [73] Rgraphviz_2.50.0    BiocManager_1.30.25 BiocGenerics_0.52.0
#> [76] jsonlite_1.8.9      plyr_1.8.9          R6_2.5.1           
#> [79] ggcyto_1.34.0       systemfonts_1.1.0   fs_1.6.5           
#> [82] zlibbioc_1.52.0

References

G. Finak, J. Frelinger, W. Jiang, E. W. Newell, J. Ramey, M. M. Davis, S. A. Kalams, S. C. De Rosa and R. Gottardo, “OpenCyto: An Open Source Infrastructure for Scalable, Robust, Reproducible, and Automated, End-to-End Flow Cytometry Data Analysis,” PLoS Computational Biology, vol. 10, p. e1003806, August 2014.

G. Finak and M. Jiang, “flowWorkspace: Infrastructure for representing and interacting with gated and ungated cytometry data sets.,” 2023.

G. Finak, W. Jiang and R. Gottardo, “CytoML for cross-platform cytometry data sharing,” Cytometry Part A, vol. 93, pp. -7, 2018.

G. Monaco, H. Chen, M. Poidinger, J. Chen, J. P. de Magalhães and A. Larbi, “flowAI: automatic and interactive anomaly discerning tools for flow cytometry data,” Bioinformatics, vol. 32, p. 2473–2480, April 2016. P. Van, W. Jiang, R. Gottardo and G. Finak, “ggcyto: Next-generation open-source visualization software for cytometry,” Bioinformatics, 2018.