Beads unsupervised classification — BeadsUnsup • denoisingCTF

This function uses clustering to classify the events into beads or cells. Beads are expected to be a group of very little dispersion, so the classification is assessed by looking into the Coefficient of Variation of the rows classified as beads. If this is > CV_max, the classification failed and the df is converted to 0. Otherwise, df with a new class column is outputted. The summary stats are written in CSV files either case.

BeadsUnsup(
  df,
  method = c("GMM", "k_means"),
  beads_ch = beads_ch,
  n_clusters = 2,
  CV_max = 0.05,
  filename = filename,
  output_path = output_path,
  FCSdesc = FCSdesc
)

Arguments

df	An object of class `data.frame`.
method	A character vector with the name of the clustering algorithm to be used. Options are Gaussian Mixture Models 'GMM' or K-means 'k_means'.
beads_ch	A numeric vector with the column indexes for the beads channels.
n_clusters	A numeric vector indicating the number of clusters to be targeted.
CV_max	A numeric vector indicating the maximum value for the coefficient of variation (CV) to evaluate beads cluster.
filename	A character vector with the name of the file that originated the data.frame.
output_path	A character vector with full path names.
FCSdesc	FCS column descriptions (extracted from `flowFrame` object)

Value

If CV< CV_max returns a data.frame with new class column for beads or cells. Otherwise, returns a vector = 0.