PhenoGMM: Train GMM-fitted model to FCS data.

In rprops/Phenoflow_package: Advanced analysis of microbial flow cytometry data

Description

Train GMM-fitted model to FCS data.

Usage

PhenoGMM(
  fcs_x,
  param,
  downsample = 0,
  nG = 128,
  auto_nG = FALSE,
  nG_interval = 4,
  fcs_scale = FALSE,
  diagnostic_plot = FALSE
)

Arguments

fcs_x	flowSet object with input data on which the model should be built
param	parameters to be used in the mixture modeling.
downsample	Indicate to which sample size individual samples should be downsampled. By default no downsampling is performed
nG	Number of mixtures to use. Defaults to 128.
auto_nG	TRUE/FALSE. Option to choose best number of mixtures from 1:nG based on BIC. Defaults to FALSE which forces nG clusters.
nG_interval	if auto_nG = TRUE, specify the intervals from nG_interval:nG to calculate BIC for. Defaults to 4.
fcs_scale	Should data be scaled/normalized by row and column before running GMM? Defaults to FALSE.
diagnostic_plot	Specify whether a diagnostic plot should be made showing the cluster allocation of each cell in the specified parameter space.

Examples

data(flowData_transformed)
testGMM <- PhenoGMM(flowData_transformed, downsample = 1e3, 
nG = 30,
auto_nG = TRUE,
nG_interval = 10,
param = c("FL1-H", "FL3-H"))
testPred <- PhenoMaskGMM(flowData_transformed, gmm = testGMM)

rprops/Phenoflow_package documentation built on Sept. 22, 2020, 5:43 p.m.