proBin: Probability binning - a metric for evaluating multivariate...

Description Usage Arguments Details Value Author(s) See Also Examples

View source: R/pbin.R

Description

This function divides the flowframe events into bins such that each bin contains the same number of events. The number of events falling into each bin can then be compared across the control and test samples using statistical methods such as the Chi-squared test.

Usage

1
proBin(m, minEvents=500,channels=NULL)

Arguments

m

An object of class flowFrame

minEvents

The minEvents The minimum number of events in each bin. (i.e. the termination criterion for the probability binning algorithm)

channels

A character vector for the Flourescence channels on which probability binning is to be performed. Defaults is NULL, in which case, all flourescence channels are used for probability binning.( Time information, if provided in the flowFrame is discarded)

Details

The flowSet is first filtered using a rectangleGate and the norm2Filter is subsequently fitted to the remaining subset.

Value

A list with items:

table

A data.frame that stores information regarding each node of the tree generated during each stage of the probability binning algorithm. Each row in the table represents a node, the first row representing the original flowFrame matrix.

The dataIndx column provides indexes for retrieving the matrices during each stage of the binning process from the enviroment data .

The parent field indicates the row number in the table that holds the parent information for the corresponding node.

The left and right columns indicates the row numbers in the table that stores information regarding the children of that particular node. The leaf nodes that hold the binned data can be identified by the nodes with the left of right values of zero( ie. no children nodes)

The visited column is used internally by the algorithm to check if a particular node has been visited during the computation process.

data

An enviroment in which the matrices generated during each stage of the probability binning process is stored. The matrices stored at the leaf nodes represent the binned events obtained after the stop criterion of minEvents has been achieved. These can be identified by the corresponding dataIndx fields provided by the rows in the table with the left or right column values of zero.

limits

A list containing the the boundaries of each hyperplane generated during probability binning

splitPars

A data.frame containing two columns splitCol - indicates the column number of the flowFrame , the split was performed.

splitMed - The median value which was used as the threshold for splitting the flowFrame

The splitCol and splitMed parameters are utilized by the plotBins and shadeBins functions in visualizing the differences between control and test sample cases.

Author(s)

Nishant Gopalakrishnan

See Also

plotBins, binByRef

Examples

1
2
3
library(flowCore)
data(GvHD)
res<-proBin(GvHD[[1]],200,channels=c("FSC-H","SSC-H","FL1-H","FL4-H")) 

Example output



flowStats documentation built on Nov. 8, 2020, 6:49 p.m.