COMPASSContainer: Generate the Data Object used by COMPASS

In COMPASS: Combinatorial Polyfunctionality Analysis of Single Cells

Description

This function generates the data container suitable for use with COMPASS.

Usage

COMPASSContainer(
  data,
  counts,
  meta,
  individual_id,
  sample_id,
  countFilterThreshold = 0
)

Arguments

data	A list of matrices. Each matrix M_i is made up of N_i cells by K markers; for example, it could be the intensity information from an intracellular cytokine experiment. Each element of the list should be named; this name denotes which sample the cell intensities were measured from.
counts	A named integer vector of the cell counts(of the parent population) for each sample in data.
meta	A data.frame of metadata, describing the individuals in the experiment. Each row in meta should correspond to a row in data. There should be one row for each sample; i.e., one row for each element of data.
individual_id	The name of the vector in meta that denotes the individuals from which samples were drawn. In this sense an individual equates to a single subject, or person.
sample_id	The name of the vector in meta that denotes the samples. The sample_id identifies a combination of a subject with visit (if any), cell subset measured (e.g. CD4), and stimulation. This vector should contain all of the names in the data input.
countFilterThreshold	Numeric; if the number of parent cells is less than this threshold, we remove that file. Default is 0, which means filter is disabled.

Details

The names attributes for the data and counts objects passed should match.

Value

A COMPASSContainer returns a list made up of the same components as input the model, but checks and sanitizes the supplied data to ensure that it conforms to the expectations outlied above.

Examples

set.seed(123)
n <- 10 ## number of samples
k <- 3 ## number of markers

## generate some sample data
sid_vec <- paste0("sid_", 1:n) ## sample ids; unique names used to denote samples
iid_vec <- rep_len( paste0("iid_", 1:(n/2) ), n ) ## individual ids

## generate n matrices of 'cell intensities'
data <- replicate(n, {
  nrow <- round(runif(1) * 1E2 + 1000)
  ncol <- k
  vals <- rexp( nrow * ncol, runif(1, 1E-5, 1E-3) )
  vals[ vals < 2000 ] <- 0
  output <- matrix(vals, nrow, ncol)
  output <- output[ apply(output, 1, sum) > 0, ]
  colnames(output) <- paste0("M", 1:k)
  return(output)
})
names(data) <- sid_vec

## make a sample metadata data.frame
meta <- data.frame(
  sid=sid_vec,
  iid=iid_vec,
  trt=rep( c("Control", "Treatment"), each=5 )
)

## generate an example total counts
## recall that cells not expressing any marker are not included
## in the 'data' matrices
counts <- sapply(data, nrow) + round( rnorm(n, 1E3, 1E2) )
counts <- setNames( as.integer(counts), names(counts) )

## insert everything into a COMPASSContainer
CC <- COMPASSContainer(data, counts, meta, "iid", "sid")

Example output

COMPASS documentation built on Nov. 8, 2020, 8:05 p.m.