autospillwin

The autospillwin package implements the AutoSpill algorithm for calculating spillover coefficients, used to compensate or unmix flow cytometry data. Please note that this was developed for conventional flow cytometry. Calculation of spillover (spectral) matrices for unmixing spectral flow cytometry samples is dealt with better in AutoSpectral, which is not out yet.

This fork is intended to be Windows-compatible. For Mac, Unix or cluster usage, we recommend the original version.

For more details, please see:
Roca et al: AutoSpill is a principled framework that simplifies the analysis of multichromatic flow cytometry data Nature Communications 2021.05.17; doi:10.1038/s41467-021-23126-8 (2021). And Colibri Cytometry: How does AutoSpill work?

Installation

To install autospillwin from this GitHub repository, use the function install_github in the devtools package.

library( devtools )

install_github( "DrCytometer/autospillwin" )

Help

You can use the standard help in R.

library( autospillwin )

? get.marker.spillover
? refine.spillover

Examples

calculate_compensation_minimal.r

Runs a calculation of compensation with autospill, without creating any figures or tables.

Requires assigning proper values to the variables: control.dir directory with the set of single-color controls control.def.file csv file defining the names and channels of the single-color controls

The creation of the fcs_control_file is critical. See the example provided. You may also use the control_file_prep.r script to get started. Both are in the examples folder.

library( autospillwin )

# set parameters

asp <- get.autospill.param()

# read flow controls

control.dir <- "../fcs_control_data/"
control.def.file <- "../fcs_control_data/fcs_control.csv"

flow.control <- read.flow.control( control.dir, control.def.file, asp )

# gate events before calculating spillover

flow.gate <- gate.flow.data( flow.control, asp )

# get initial spillover matrices from untransformed data

marker.spillover.unco.untr <- get.marker.spillover( TRUE, flow.gate,
    flow.control, asp )

# refine spillover matrix iteratively

refine.spillover.result <- refine.spillover( marker.spillover.unco.untr,
    NULL, flow.gate, flow.control, asp )

For more example scripts, see the batch folder after installing the package.

The scripts calculate_compensation_paper.r and calculate_compensation_paper.sh can be used to reproduce the results of AutoSpill for single-color controls appearing in the paper above. For this, you will need to download the datasets (FCS files and auxiliary fcs_control.csv files) from FlowRepository: MM1 dataset, HS1 & HS2 datasets, and Be1 dataset.