colStats: Row and Column Summary Statistics

In matter: A framework for rapid prototyping with file-based data structures

Description

These functions perform calculation of summary statistics over matrix rows and columns, for each level of a grouping variable (optionally), and with implicit row/column scaling and centering if desired.

Usage

## S4 method for signature 'ANY'
colStats(x, stat, groups,
    na.rm = FALSE, tform = identity,
    col.center = NULL, col.scale = NULL,
    row.center = NULL, row.scale = NULL,
    drop = TRUE, BPPARAM = bpparam(), ...)

## S4 method for signature 'ANY'
rowStats(x, stat, groups,
    na.rm = FALSE, tform = identity,
    col.center = NULL, col.scale = NULL,
    row.center = NULL, row.scale = NULL,
    drop = TRUE, BPPARAM = bpparam(), ...)

Arguments

x	A matrix on which to calculate summary statistics.
stat	The name of summary statistics to compute over the rows or columns of a matrix. Allowable values include: "min", "max", "prod", "sum", "mean", "var", "sd", "any", "all", and "nnzero".
groups	A factor or vector giving the grouping. If not provided, no grouping will be used.
na.rm	If TRUE, remove NA values before summarizing.
tform	A dimensionality-preserving transformation to be applied to the matrix (e.g., log() or sqrt()).
col.center	A vector of column centers to substract from each row. (Or a matrix with a column for each level of groups.)
col.scale	A vector of column scaling factors to divide from each row. (Or a matrix with a column for each level of groups.)
row.center	A vector of row centers to substract from each column. (Or a matrix with a column for each level of groups.)
row.scale	A vector of row centers to scaling factors to divide from each column. (Or a matrix with a column for each level of groups.)
drop	If only a single summary statistic is calculated, return the results as a vector (or matrix) rather than a list.
BPPARAM	An optional instance of BiocParallelParam. See documentation for bplapply.
...	Additional arguments.

Details

The summary statistics methods are calculated over chunks of the matrix using colstreamStats and rowstreamStats. For matter objects, the iteration is performed over the major dimension for IO efficiency.

Value

A list for each stat requested, where each element is either a vector (if no grouping variable is provided) or a matrix where each column corresponds to a different level of groups.

If drop=TRUE, and only a single statistic is requested, then the result will be unlisted and returned as a vector or matrix.

Author(s)

Kylie A. Bemis

See Also

colSums

Examples

register(SerialParam())

set.seed(1)

x <- matrix(runif(100^2), nrow=100, ncol=100)

groups <- as.factor(rep(letters[1:5], each=20))

colStats(x, "mean", groups=groups)

matter documentation built on Nov. 8, 2020, 6:15 p.m.