ManhattanPlot.Rd
Produces a Manhattan plot of marginal predictor evidence from a Reversible Jump Results object.
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
)
A R2BGLiMS R2BGLiMS_Results-class object frum running the R2BGLiMS R2BGLiMS command.
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.
Alternatively can simply pass a named vector of probabilties
Character vector indicating a subset of covariates to include in the plot.
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).
Optional character string to name the plot with
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.
If plotting p-values this option draws a dashed red line at the Bonferroni threshold. (default is TRUE).
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)
Must be provided if top.hits are provided above - used to calculate pairwise correlations.
Optional vector of colours to use for the points (as many and in the same order as the covariates)
Optional vector of labels to use for the points (as many and in the same order as the covariates)
Set to TRUE to avoid adding x tick mark labels (e.g. if there are many many variables)