ManhattanPlot: Manhattan Plot

In pjnewcombe/R2BGLiMS: R interface to the java package BGLiMS (Bayesian Generalised Linear Model Selection).

Description

Produces a Manhattan plot of marginal predictor evidence from a Reversible Jump Results object.

Usage

ManhattanPlot(
  results = NULL,
  plot.quantity = "PosteriorProbability",
  results.vector = NULL,
  covariates.to.include = NULL,
  var.dictionary = NULL,
  plot.title = "Manhattan Plot",
  y.max = NULL,
  add.bonferroni = FALSE,
  top.hits = NULL,
  X.mat = NULL,
  point.cols = NULL,
  point.labs = NULL,
  include.var.names.on.x.axis = TRUE
)

Arguments

results	A R2BGLiMS_Results-class object frum running the R2BGLiMS command.
plot.quantity	Can be "PosteriorProbability" (default), "BayesFactor", "ABF", or "pvalue". The latter is implemented to allow easy comparison between analysis frameworks using the same plot style, and only works if a vector of p-values is supplied. ABF is for plotting Approximate Bayes Factors in favour of the null (i.e. small values correspond to evidence of association), as proposed in Wakefield, J. (2009). Genetic Epidemiology.
results.vector	Alternatively can simply pass a named vector of probabilties
covariates.to.include	Character vector indicating a subset of covariates to include in the plot.
var.dictionary	Character vector containing a mapping of covariate names to substitutions for the plot. The names of the elements are the variable names as in the original analysis (i.e. as named in the results object).
plot.title	Optional character string to name the plot with
y.max	If plotting Bayes Factors, this is an optional upper limit for the y-axis (e.g. to make comparison with a different set of results easier on the eye). Note that if p-values are being plotted, then this can be used to supply the maximum -log10(p-value) on the y-axis.
add.bonferroni	If plotting p-values this option draws a dashed red line at the Bonferroni threshold. (default is TRUE).
top.hits	If a list of "top hits" are provided here, along with the original X matrix below, then the points are coloured according to their pairwise correlation with the "top hits" (default NULL)
X.mat	Must be provided if top.hits are provided above - used to calculate pairwise correlations.
point.cols	Optional vector of colours to use for the points (as many and in the same order as the covariates)
point.labs	Optional vector of labels to use for the points (as many and in the same order as the covariates)
include.var.names.on.x.axis	Set to TRUE to avoid adding x tick mark labels (e.g. if there are many many variables)

Author(s)

Paul Newcombe

Examples

library(R2BGLiMS)

### --- Logistic regression with two model space prior components
# Load results from the biopsy logistic regression analysis (see R2MHRJ examples)
data(biopsyResults)
ManhattanPlot(biopsyResults)

pjnewcombe/R2BGLiMS documentation built on Feb. 10, 2020, 8:52 p.m.