Clean Controls

A four-part function to clean single-color controls in order to extract fluorophore signatures. Any part can be run independently:

• Stage 1: PeacoQC to eliminate flow artefacts. Not required in most cases.

• Stage 2: Trimming to eliminate extreme events. Not recommended for most use cases.

• Stage 3: Autofluorescence noise removal using PCA unmixing on matching unstained (cells only).

• Stage 4: Brightest event selection from positive, universal negative from matching negative, and downsampling to speed up RLM spectra optimization.

Usage

clean.controls(
  flow.control,
  asp,
  time.clean = FALSE,
  trim = FALSE,
  trim.factor = NULL,
  af.remove = TRUE,
  universal.negative = TRUE,
  downsample = TRUE,
  negative.n = asp$negative.n,
  positive.n = asp$positive.n,
  scatter.match = TRUE,
  scrub = FALSE,
  intermediate.figures = FALSE,
  main.figures = TRUE,
  parallel = FALSE,
  verbose = TRUE,
  threads = NULL
)

Arguments

flow.control

A list prepared using define.flow.control, containing the data and essential information about the cytometer and data structure.

asp

The AutoSpectral parameter list, prepared using get.autospectral.param.

time.clean

Logical, default is FALSE. Whether to run PeacoQC to remove time-based inconsistencies in the controls.

trim

Logical, default is FALSE. Whether to remove extreme events (positive and negative) from controls.

trim.factor

Numeric. Default is asp$rlm.trim.factor. Required if trim = TRUE.

af.remove

Logical, default is TRUE. Whether to remove intrusive autofluorescence contamination from cell controls using PCA-based identification and gating. Requires universal negatives to be defined in the control file and in flow.control.

universal.negative

Logical, default is TRUE. Whether to use a universal negative sample as the negative for spectral extraction. Requires universal negatives to be defined in the control file and in flow.control.

downsample

Logical, default is TRUE. Whether to reduce cell and bead control events to speed up processing.

negative.n

Integer. Number of events to include in the downsampled negative population. Default is asp$negative.n.

positive.n

Integer. Number of events to include in the downsampled positive population. Default is asp$positive.n.

scatter.match

Logical, default is TRUE. Whether to select negative events based on scatter profiles matching the positive events. Defines a region of FSC and SSC based on the distribution of selected positive events.

scrub

Logical, if TRUE allows for re-cleaning of already cleaned data, provided there are clean data in flow.control.

intermediate.figures

Logical, if TRUE returns additional figures to show the inner workings of the cleaning, including definition of low-AF cell gates on the PCA-unmixed unstained and spectral ribbon plots of the AF exclusion from the unstained. Default is FALSE to speed up processing.

main.figures

Logical, if TRUE creates the main figures to show the impact of intrusive autofluorescent event removal and scatter-matching for the negatives.

parallel

Logical, default is FALSE, in which case parallel processing will not be used. Parallel processing will likely be faster when many small files are read in. If the data is larger, parallel processing may not accelerate the process much.

verbose

Logical, default is TRUE. Set to FALSE to suppress messages.

threads

Numeric, number of threads to use for parallel processing. Default is NULL which will revert to asp$worker.process.n if parallel=TRUE.

Value

Returns a modified flow.control with the original data intact. New, cleaned data and corresponding factors are stored in new slots.