tcR (version 1.3)

bootstrap.tcr: Bootstrap for data frames in package tcR.

Description

Resample rows (i.e., clones) in the given data frame and apply the given function to them.

Usage

bootstrap.tcr(.data, .fun = entropy.seg, .n = 1000, .size = nrow(.data),
  .sim = c("uniform", "percentage"), .postfun = function(x) {     unlist(x)
  }, .verbose = T, .prop.col = "Read.proportion", ...)

Arguments

.data

Data frame.

.fun

Function applied to each sample.

.n

Number of iterations (i.e., size of a resulting distribution).

.size

Size of samples. For .sim == "uniform" stands for number of rows to take. For .sim == "percentage" stands for number of barcodes / read counts to take.

.sim

A character string indicating the type of simulation required. Possible values are "uniform" or "percentage". See "Details" for more details of type of simulation.

.postfun

Function applied to the resulting list: list of results from each processed sample.

.verbose

If T than show progress bar.

.prop.col

Column with proportions for each clonotype.

...

Further values passed to .fun.

Value

• Either result from .postfun or list of length .n with values of .fun.

Details

Argument .sim can take two possible values: "uniform" (for uniform distribution), when each row can be taken with equal probability, and "perccentage" when each row can be taken with probability equal to its "Read.proportion" column.

Examples

# Apply entropy.seg function to samples of size 20000 from immdata$B data frame for 100 iterations.
bootstrap.tcr(immdata[[2]], .fun = entropy.seg, .n = 100, .size = 20000, .sim = 'uniform')