quest: Question differential expression patterns using min-max...

In pwirapati/acdx: Aggregated Cell Differential Expression

Description

Rank genes according to difference between the minimum of a set of parameter and the maximum of another sets.

Usage

  quest(
    param,
    showcomb = F,
    default="L",
    H=NULL,
    N=NULL,
    L=NULL,
    sep=",",
    na.rm=FALSE,
    include=NULL,
    exclude=NULL,
    method="minmax",
    multest_func=bremt,
    ...
    )
  questmm(...)

Arguments

param	A parameter object that is a part of acdx output, typically fit$beta, but can be other parameters or transformed values thereof.
showcomb	If TRUE, just return the list of possible combinations of coefficients and cell types, without doing any analysis.
default	The default class for comparions, which is "L" (the "low" group
H	A set of regular expression patterns for the "high" group
N	A set of regular expression patterns for the "neutral" group
L	A set of regular expression patterns for the "low" group
sep	String separator for coeffcient-celltype combinations. Can be changed if comma is already used as part of the names.
na.rm	Skip coefficients if they are NA. The default is to make the interest score NA for any missing coefficient.
include	Vector of gene names to be included. If NULL (default), all are included. If specified, it excludes all but those in the set.
exclude	A vector of gene names to be excluded. If NULL (default), none is excluded (from the result of applying include=).
method	Interest functional method.
multest_func	Multiple-testing procedure callback function. It should take a matrix of interest scores with genes as rows and bootstraps as column.
...	Arguments to multest_func

Details

questmm calculates, for each gene and each bootstrap replicate, a score defined as the difference (log-ratio) between the minimum values of the "high" group and maximum value of the "low" group.

The groups of coefficients can span across multiple cell types, and they can be arbitrarily chosen.

The resulting scores are passed on a multiple-testing correction method bremt, which collapse the bootstrap replicates into significance test statistics.

See Also

bremt

pwirapati/acdx documentation built on Jan. 11, 2021, 12:31 a.m.