cohesiveness: Cohesiveness analysis on large omic datasets.

In fjcamlab/cohesiveness: Non-parametric feature selection for large omic datasets

Usage

cohesiveness(mx, cl, comb.p.val = "fisher",
                 method = "complete", trim = 0.01)

Arguments

mx	a numeric matrix containing the omic dataset, where rows are features and columns are samples.
cl	a character vector describing which samples are belonging to each category.
comb.p.val	a character indicating if "fisher" or "stouffer" method must be applied to summarize p-values.
method	a character indicating if "complete" or "trimmed" mean must be computed.
trim	fraction of quantile to be discarded from each tail of omic distributions, if "trimmed" mean is applied to.

Details

R function to perform a cohesiveness analysis.

Value

This function returns a 'data.frame', including the cohesiveness statistics per feature associated to each category of samples.

Author(s)

FJ Campos-Laborie ([e-mail])

See Also

plotCohesiveness

Examples

## Example of omic dataset based on normal distribution data
mx <- t(sapply(seq_len(1000), function(x) rnorm(n = 200, mean = 5, sd = 1)))

## Formatting the example matrix
rownames(mx) <- paste("g", seq_len(1000), sep = "")
colnames(mx) <- paste("s", seq_len(200), sep = "")

## Naming a vector of sample categories
classes <- rep(LETTERS[1:8], c(20, 40, 15, 40, 65, 5, 5, 10))
names(classes) <- colnames(mx)

## Running cohesiveness
resultsCohesiveness <- cohesiveness(mx, classes)

fjcamlab/cohesiveness documentation built on May 6, 2019, 12:09 p.m.