forestplot-methods: Making a forest plot of the results of uWMW

In CenterForStatistics-UGent/unifiedWMWqPCR: Unified Wilcoxon-Mann Whitney Test for testing differential expression in qPCR data

Making a forest plot of the results of uWMW

Description

This function creates a forest plot indicating the (log) odds ratios, the (log) odds or the probabilities for the results of the unified Wilcoxon-Mann-Whitney test.

Usage

## S4 method for signature 'ANY'
forestplot(x, ...)

## S4 method for signature 'uwmwRes'
forestplot(
  x,
  estimate = c("logor", "logodds", "or", "odds", "p"),
  level = 0.95,
  ...
)

## S4 method for signature 'uwmwEstimate'
forestplot(x, ...)

forestplot.internal(
  x,
  annotate = TRUE,
  addfit = TRUE,
  xlim = NULL,
  alim = NULL,
  ylim = NULL,
  at = NULL,
  steps = 5,
  level = 0.95,
  digits = 2,
  refline = NULL,
  xlab = NULL,
  slab = NULL,
  mlab = NULL,
  ilab = NULL,
  ilab.xpos = NULL,
  ilab.pos = NULL,
  order = NULL,
  transf = FALSE,
  atransf = FALSE,
  targs = NULL,
  rows = NULL,
  efac = 1,
  pch = 15,
  psize = 1,
  col = "darkgrey",
  border = "darkgrey",
  cex = NULL,
  cex.lab = NULL,
  cex.axis = NULL,
  refcol = "red",
  predcol = refcol,
  ...
)

Arguments

x	An object of class uwmwRes or uwmwEstimate.
...	parameters passed down to the internal functions. These can be any of the following.
estimate	An optional character string defining which measure should be plotted. It can take the values logor, or, logodds, odds or p (for displaying the probability of differential expression, not the p-value!). Defaults to logor. Note that this argument is ignored when plotting an uwmwEstimate object.
level	Numerical value between 0 and 1 to specify the width of the confidence interval. Defaults to 0.95 (95% confidence interval).
annotate	A logical value indicating whether the plot needs to be annotated, i.e. whether the values for the chosen measure and confidence interval should be displayed on the right of the plot. Defaults to TRUE
addfit	A logical value indicating whether the reference measure should be plotted. See Details.
xlim	The horizontal limits of the plot region. If unspecified, the function tries to set the horizontal plot limits to some sensible values. Should not be used by the user.
alim	the actual x axis limits. If unspecified, an educated guess is taken by the function.
ylim	The vertical limits of the plot. If unspecified, the function does what it thinks is best. Should not be used by the user.
at	Position of the x axis tick marks and corresponding labels are placed. If unspecified, the function tries to position them at sensible values.
steps	An integer indicating the number of tick marks on the X axis. Ignored when at is specified. Defaults to 5
digits	integer specyfing the number of decimal places for tick mark labels and annotations. Can also be a vector of two integers. In that case, the first value specifies the number of decimal places for the annotations, the second for the x axis labels.
refline	numerical value indicating where a reference line should be drawn. An NA value will prevent the line from being drawn. See Details.
xlab	title for the x axis. If unspecified, the function tries to figure out the fitting title.
slab	optional vector with names for the displayed genes.
mlab	optional character string giving a label to the intercept estimate. If unspecified, this is created in the function if necessary.
ilab	optional vector or matrix with character strings providing additional information that can be plotted next to the genes.
ilab.xpos	Vector of numerical values specifying the x axis positions of the character vectors given via ilab. This has to be specified when ilab is specified.
ilab.pos	integer from 1 to 4 specifying the alignment of the character vector(s) given via ilab (2 is right aligned, 4 is left aligned). Default is to center the labels.
order	optional character string, character vector or numerical vector specifying how the genes should be ordered.
transf	optional argument specifying the name of a function that should be used to transform the observed effect sizes, summary estimates, fitted values and confidence interval bounds (e.g., transf=exp). Defaults to FALSE, which means that no transformation is used.
atransf	optional argument specifying the name of a function that should be used to transform the x-axis labels and annotations (e.g., transf=exp). Defaults to FALSE, which means that no transformation is used.
targs	optional arguments needed by the function specified via transf or atransf.
rows	optional vector specifying the horizontal position for the plotted results. If unspecified, the layout happens automatically. See Details and Examples for more information.
efac	numerical value specifying the vertical expansion of the arrows, summary estimate symbols and Ci limits. Normally the default of 1 should work just fine.
pch	plotting symbol used for the observed effect sizes. By default, it's a filled square.
psize	optional vector with the point sizes for the observed effects. If set to NULL, the point sizes are drawn proportional to the value of the log of the test statistic. Defaults to 1.
col	character string specifying the color used for the individual estimates.
border	character string specifying the color used for the border of the individual estimates.
cex	optional numerical value for expansion of text and symbols. See also par.
cex.lab	Optional numerical value for expansion of the axis title.
cex.axis	Optional numerical value for expansion of the x axis labels.
refcol	Character string specyfying the color of the reference line. Defaults to red.
predcol	character string specifying the color of the estimated reference value. Ignored if estimate is logor or or.

Details

The function has methods for uwmwRes and uwmwEstimate objects. When called for an uwmwRes object, the requested estimate is first calculated using getEstimate and the result is passed on to the next method.

Note that in either case, it is not possible to use the function on a subset of either type of object. The subsetting functions for uwmwRes and uwmwEstimate objects return matrices, and hence necessary information on the reference value is lost. To plot a subset of your data, use the order argument as shown in the examples.

Adding a reference value to the plot only makes sense when plotting the log(odds), odds or probabilities. If log(OR) or OR are plotted, addfit is set to FALSE.

The default settings plot a reference line at a location depending on the plotted estimate. For log(OR), the line is plotted at refline = 0. For OR, the line is plotted at refline = 1.

Value

NULL invisibly

Warning

Although the internal function is shown here (merely for illustration of the arguments and defaults), the user shouldn't be calling this one directly. The function is not exported.

Note

Thanks to the work of W.Viechtbauer, forestplot provides many possibilities for tweaking and customizing the plots. Many of the arguments work the same as in the function forest.rma (metafor package). You can always check the help file of forest.rma for more illustrations on the different arguments.

This function is currently implemented using an internal function that expects an uwmwEstimate object. In a future version, the internal function will be rewritten to be more generic. This will enable the definition of methods for other classes without need to change the internal function itself.

Author(s)

This code is adapted by Joris Meys from the function forest.rma (metafor package). The original function is written by W. Viechtbauer.

Examples

data(NBmat)
NBtest <- uWMW(NBmat, groups=NBgroups)
sigid <- which(pval(NBtest) < 0.05)
forestplot(NBtest,"logodds",order=sigid)

nameid <- c("hsa-mir-30a-3p","hsa-mir-30a-5p")
forestplot(NBtest,"p",order=nameid,addfit=FALSE,
           refline=NA,main="Comparison 30a")

forestplot(NBtest,"p",order=nameid,addfit=FALSE,
           refline=0.5,main="Comparison 30a",
           alim=c(0,1),xlim=c(-1,2),at=c(0,0.5,1))

CenterForStatistics-UGent/unifiedWMWqPCR documentation built on Oct. 28, 2024, 2:34 p.m.