Learn R Programming

outbreaker (version 1.1-6)

outbreaker graphics: Plot outbreaker's results

Description

These are the main functions used for generating graphics from the raw output of outbreaker and outbreaker.parallel.

  • plotChainsis used for plotting MCMCs
  • transGraphplots a graph of inferred ancestries
  • plotOutbreakattempts to synthetize the reconstruction of small outbreaks

Usage

plotChains(x, what="post", type=c("series","density"), burnin=0,
           dens.all=TRUE, col=funky(x$n.runs), lty=1, lwd=1,
           main=what, legend=TRUE, posi="bottomleft", ...)

transGraph(x, labels=NULL, burnin=x$burnin, threshold=0.2, col.pal=NULL, curved.edges=TRUE, annot=c("dist","support"), sep="/", ...)

plotOutbreak(x, burnin=x$burnin, thres.hide=0.2, col=NULL, col.pal=colorRampPalette(c("blue","lightgrey")), edge.col.pal=NULL, col.edge.by="prob", annot=c("dist","prob"), sep="/", cex.bubble=1, edge.max.dist = 10, lwd.arrow=2, xlim=NULL, ...)

Arguments

x
the output of outbreaker or outbreaker.parallel.
what
a character chains giving the name of the item to be plotted. See names(x$chains) for possible values. By default, log-posterior values are plotted
type
a character indicating if the chains should be plotted as time series ("series"), or as density ("density").
burnin
an integer indicating the number of MCMC steps to discard before plotting chains.
dens.all
a logical indicating if, in the case of multiple runs, the overall density of the different chains should be plotted in addition to individual densities.
col
a vector of colors to be used to plot different chains.
lty
a vector of integers specifying line types for the different chains.
lwd
same as lty, but for line width.
main
the title to be added to the plot.
labels
the labels to be used to name the nodes of the graph (cases).
threshold
the minimum support for ancestries to be plotted; 'support' is defined as the frequency of a given ancestor in the posterior distribution; defaults to 0.2.
thres.hide
a threshold of posterior support for displaying ancestries; ancestries with less than this frequency in the posterior are hidden.
col.pal,edge.col.pal
the color palette to be used for the edges (ancestries).
curved.edges
a logical indicating whether edges should be curved.
col.edge.by
a character string indicating which information should be used to color the edges ('dist': genetic distance; 'prob': support for the ancestry)
annot
a character indicating which information should be used to annotate the edges; this can be the distances between ancestors and descendents ("dist") and the posterior support for ancestries ("support"); if both are requested, fields will be con
sep
a character indicating the separator to be used when concatenating several types of annotation.
cex.bubble
a numeric value indicating the size factor for the bubbles representing the generation time distribution.
edge.max.dist
a number indicating the threshold distance bounding the color palette used for the edges; useful to avoid showing edges corresponding to distances larger than a given number.
lwd.arrow
a numeric value indicating the size factor for the arrows.
xlim
the limits of the X axis; if NULL, determined from the data.
legend
a logical indicating if a legend should be plotted for the different runs.
posi
a character string indicating the position of the legend (see ?legend).
...
further arguments to be passed to other functions.

Examples

Run this code
data(fakeOutbreak)
attach(fakeOutbreak)

## examine MCMC
plotChains(res)
plotChains(res,type="dens")
plotChains(res,type="dens", what="mu1", burnin=2e4)

## represent posterior ancestries
transGraph(res, annot="", main="Posterior ancestries")
transGraph(res, annot="", main="Posterior ancestries - support > 0.5",
   threshold=0.5)
if(require(adegenet)){
transGraph(res, annot="", main="Posterior ancestries - support > 0.01",
   threshold=0.01, col.pal=spectral)
}
## summary plot
plotOutbreak(res,cex.bubble=0.5, thres.hide=0.5,
   main="Outbreak reconstruction")


detach(fakeOutbreak)

Run the code above in your browser using DataLab