Skip to contents

Two displays of a da_sensitivity() result: which features each run rejects ("tiles"), and how each feature's effect changes with one setting ("traces").

Usage

# S3 method for class 'ecodive_da_sensitivity'
plot(
  x,
  type = c("tiles", "traces"),
  covariate = NULL,
  parameter = NULL,
  n_features = 30,
  ...
)

Arguments

x

A da_sensitivity() result.

type

"tiles" (the default) or "traces".

covariate

For the model methods, which covariate to show: one of x$covariate. Default: the covariate of the term of interest (attr(x, "covariate")).

parameter

Which runs to show. For "tiles", the base run and the runs of one parameter ("pseudocount", "zeros", "scale_sd" or "crossed"); default: every run. For "traces", the setting to draw the effects against ("pseudocount", "zeros" or "scale_sd"); default: the first numeric one varied. With crossed = TRUE, the traces hold the other setting at its first value, as the subtitle says.

n_features

The most features to show. Default: 30.

...

Further graphical parameters for graphics::image() ("tiles") or graphics::plot() ("traces"), such as main.

Value

x, invisibly.

Details

Tiles. A row per feature (those rejected in at least one run: the most often rejected first, then by the size of their effect in the base run), a column per run (the base run first, then the runs of each setting, separated by vertical lines, with the setting named above). A cell is red when the run rejects the feature with a positive effect, blue with a negative one, dark grey when the method gives no effect (ANCOM), light grey when the run doesn't reject it, and white when the run has no result for it. A feature's name is bold when every run makes the same decision (summary$passed).

Traces. For one setting, a line per feature (those rejected in the base run, else in any run, the largest effects first) through its effect at each of the setting's values, with filled points where the run rejects it and hollow ones where it doesn't, and a dotted line at 0. Names are bold as in the tiles. Against scale_sd, this is the display of ALDEx2's plotGamma(). ANCOM has no effects to draw.

See also

Examples

    x <- rbind(
      c( 20, 110, 100, 101, 100, 103, 104),
      c( 33, 110, 120, 100, 101, 100, 102),
      c( 12, 110, 100, 110, 100,  50,  90),
      c(202, 201,   9,  10,  10,  11,  11),
      c(200, 202,  10,  10,  13,  10,  10),
      c(203, 201,  14,  10,  10,  13,  12) )
    colnames(x) <- paste0('b', 1:7)
    grp <- rep(c('treatment', 'placebo'), each = 3)

    res <- da_sensitivity(
      x, 'da_ttest', grp, pseudocounts = c(0.01, 0.1, 0.5, 1, 5),
      zeros = TRUE, cpus = 1 )
    plot(res)

    plot(res, 'traces')