greg: Aggregate single-cell expression profiles
In pwirapati/acdx: Aggregated Cell Differential Expression

Description Usage Arguments Details Value Examples

View source: R/greg.R

Produce aggregate expression profiles from raw counts

1
2
3

  greg( x, cell_sample, cell_type, cell_names=NULL,
    u_0 = NULL, s2_0 = 1/12, 
    sep="\t", verbose=1 )

`x`	Raw counts in sparse dgCMatrix format, with genes as rows and cells as columns. Alternatively, if it is a single string, it is interpreted as local file name containing text table (which can be compressed).
`cell_sample`	Sample identifiers of the cells.
`cell_type`	Cell type labels of the cells.
`cell_names`	Cell identifiers to be matched by the names in `x` (should have exactly the same length as `cell_sample`). If `NULL`, then the data and annotations are assumed to have the same length and in the same order. If the names is missing in `x` but `cell_names` is supplied, then it is considered an error. The both names are present, only the intersection are used, silently dropping the mismatches, and the ordering is not important.
`u_0`	Constant shift from the origin
`s2_0`	Baseline variance, with the default 1/12 equals to that of the uniform distribution in an interval of width one.
`sep`	Column separator character; relevant only when `x` is a file of text table.
`verbose`	Output progress messages when performing the aggregation

The summary for each gene in an aggregate is the mean of raw counts, accompanied by an estimate of the variance of the mean (i.e., squared standard error). This variance is estimated by the sample variance plus s2_0, divided by the number of cells minus one.

If the cell count is zero or one, the variance is infinite (the mean is set to zero or the single-cell value, respectively). This aggregate will be ignored in subsequent analyses, although they are still formally represented by the data structure.

A list of:

`N`	A matrix of samples (rows) by cell types (columns), containing the number of cells in each aggregate.
`y`	The aggregate array with four dimensions, corresponding to the summaries, samples, genes and cell types.