csTopTable: Compute Cell-Specific Statistics

Description Usage Arguments Value Methods (by class) See Also

View source: R/csstats.R

Description

csTopTable is a generic function that returns cell type-level statistics on biological features, e.g., cell type-specific expression.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
csTopTable(x, ...)

## Default S3 method:
csTopTable(x, n = 100L, sort.by = 1L,
  decreasing = FALSE, select = NULL, subset = names(threshold),
  threshold = NULL, annotation = NULL, ..., merge = FALSE,
  nodups = FALSE)

## S3 method for class 'array'
csTopTable(x, ...)

## S3 method for class 'character'
csTopTable(x, ...)

## S3 method for class 'csSAMfit'
csTopTable(x, alternative = c("two.sided", "greater",
  "less", "min"), sort.by = "FDR", decreasing = FALSE, ...)

Arguments

x

data object, typically obtained by fitting a model that computes cell type-specific statistics, e.g., csSAMfit.

...

extra parameters passed to specific methods.

n

maximum number of top features to retain – in each cell type. If NULL or Inf then all selected features are returned. Otherwise the first n feature in each cell type are returned, after ordering according to arguments sort.by and decreasing.

sort.by

variable to use to order the features.

decreasing

logical that controls the feature ordering based on sort.by.

select

specifies the statistic to include in the result.

subset

indicates the subset of tissue/cell type to include in the result. Default is to return only the cell types for which a threshold value was provided. If NULL, as default, then no subsetting is performed.

threshold

numeric vector that indicates the threshold applied to select features based on their values of sort.by. Thresholds can be defined for specific cell types by using naming the elements of threshold with corresponding cell type names. The default threshold that is applied to all cell types is taken from the first unnamed element.

Depending on decreasing, thresholds are interpreted differently:

  • decreasing = FALSE keeps features with sort.by "<= x"

  • decreasing = TRUE keeps features with sort.by ">= x"

annotation

Bioconductor annotation package to use to annotate the features (rownames) using xbioc::geneInfo.

merge

logical that indicates if the result should be a a list of data.frame objects (merge = FALSE), or a single data.frame object with columns "Feature" and "Cell type" containing the feature identifiers and cell type respectively.

nodups

a logical that indicates if features appearing in multiple cell types should be removed. This is useful in combination with threshold to extract features that are found significant in a single cell type.

alternative

Alternative hypothesis to consider. Alternative "min" returns, for each features, the alternative achieving the minimum FDR.

Value

a list of data.frame objects if merge = FALSE, or a single data.frame object if merge = TRUE.

If merge=TRUE and select is of length one, then the result is a matrix object (feature x cell type) containing the selected value.

Methods (by class)

See Also

fdrCsSAM


shenorrLab/csSAM documentation built on May 26, 2017, 8:06 a.m.