treeGate: Use decision tree to find a group of cells that are...

In CytoDx: Robust prediction of clinical outcomes using cytometry data without cell gating

Description

A function that sse decision tree to find a group of cells that are associated with clinical outcome.

Usage

treeGate(P, x, ...)

Arguments

P	The predicted association of each cell with a clinical outcome.
x	The marker profile of each cell. Each row is a cell, each column is a marker. Must have length(P) rows.
...	Other parameters to be passed into the rpart function

Value

Returns a object created by rpart function. Also plots a graph of decision tree.

Examples

# Find the table containing fcs file names in CytoDx package
path=system.file("extdata",package="CytoDx")
# read the table
fcs_info <- read.csv(file.path(path,"fcs_info.csv"))
# Specify the path to the cytometry files
fn <- file.path(path,fcs_info$fcsName)
# Read cytometry files using fcs2DF function
train_data <- fcs2DF(fcsFiles=fn,
                    y=fcs_info$Label,
                    assay="FCM",
                    b=1/150,
                    excludeTransformParameters=
                      c("FSC-A","FSC-W","FSC-H","Time"))
# build the model
fit <- CytoDx.fit(x=as.matrix(train_data[,1:7]),
                y=train_data$y,
                xSample = train_data$xSample,
                reg=FALSE,
                family="binomial")
# check accuracy for training data
pred <- CytoDx.pred(fit,
                   xNew=as.matrix(train_data[,1:7]),
                   xSampleNew=train_data$xSample)

boxplot(pred$xNew.Pred.sample$y.Pred.s0~
          fcs_info$Label)

# Find the associated population using treeGate
TG <- treeGate(P = fit$train.Data.cell$y.Pred.s0,
              x= train_data[,1:7])

CytoDx documentation built on Nov. 8, 2020, 11 p.m.