plot.COMPASSResult: Plot a COMPASSResult

Description Usage Arguments Value Examples

View source: R/plotMeanGamma.R

Description

This function can be used to visualize the mean probability of response; that is, the probability that there is a difference in response between samples subjected to the 'treatment' condition, and samples subjected to the 'control' condition.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
## S3 method for class 'COMPASSResult'
plot(
  x,
  y,
  subset = NULL,
  threshold = 0.01,
  minimum_dof = 1,
  maximum_dof = Inf,
  must_express = NULL,
  row_annotation,
  palette = colorRampPalette(brewer.pal(10, "Purples"))(20),
  show_rownames = FALSE,
  show_colnames = FALSE,
  measure = NULL,
  order_by = FunctionalityScore,
  order_by_max_functionality = TRUE,
  markers = NULL,
  ...
)

Arguments

x

An object of class COMPASSResult.

y

This argument gets passed to row_annotation, if row_annotation is missing. It can be used to group rows (individuals) by different conditions as defined in the metadata.

subset

An R expression, evaluated within the metadata, used to determine which individuals should be kept.

threshold

A numeric threshold for filtering under-expressed categories. Any categories with mean score < threshold are removed.

minimum_dof

The minimum degree of functionality for the categories to be plotted.

maximum_dof

The maximum degree of functionality for the categories to be plotted.

must_express

A character vector of markers that should be included in each subset plotted. For example, must_express=c("TNFa & IFNg") says we include only subsets that are positive for both TNFa or IFNg, while must_express=c("TNFa", "IFNg") says we should keep subsets which are positive for either TNFa or IFNg.

row_annotation

A vector of names, pulled from the metadata, to be used for row annotation.

palette

The colour palette to be used.

show_rownames

Boolean; if TRUE we display row names (ie, the individual ids).

show_colnames

Boolean; if TRUE we display column names (ie, the column name associated with a cytokine; typically not needed)

measure

Optional. By default, we produce a heatmap of the mean gammas produced in a model fit. We can override this by supplying a matrix of suitable dimension as well; these can be generated with the Posterior* functions – see Posterior for examples.

order_by

Order rows within a group. This should be a function; e.g. FunctionalityScore, mean, median, and so on. Set this to NULL to preserve the original ordering of the data.

order_by_max_functionality

Order columns by functionality within each degree subset. to TRUE.

markers

specifies a subset of markers to plot. default is NULL, which means all markers.

...

Optional arguments passed to pheatmap.

Value

The plot as a grid object (grob). It can be redrawn with e.g. grid::grid.draw().

Examples

1
2
3
4
5
## visualize the mean probability of reponse
plot(CR)

## visualize the proportion of cells belonging to a category
plot(CR, measure=PosteriorPs(CR))

RGLab/COMPASS documentation built on Feb. 11, 2021, 3:23 p.m.