Identify topologically distinct rate shift configurations
that were sampled with BAMM, and assign each sample in the
posterior to one of the distinct shift configurations.
distinctShiftConfigurations(ephy, expectedNumberOfShifts, threshold, ...)An object of class bammshifts. This is a list with the
following components:
marg.probs: A list of the marginal probability of a shift occurring at each node of the phylogeny for each distinct rate shift configuration.
marginal_odd_ratio: Marginal posterior-to-prior odds ratios for one or more rate shifts an a given branch.
shifts: A list of the set of shift nodes for each distinct rate configuration.
samplesets: A list of sample indices that reduce to each of the unique shift sets.
frequency: A vector of frequencies of each distinct shift configuration.
coreshifts: A vector of node numbers corresponding to the
core shifts. All of these nodes have a marginal odds ratio of
at least threshold supporting a rate shift.
threshold: A single numeric value giving the marginal posterior:prior odds ratio threshold used during enumeration of distinct shift configurations.
Results are sorted by frequency:
$frequency[1] gives the most common shift configuration sampled.
$shifts[[1]] gives the corresponding node indices for that configuration.
$samplesets[[1]] gives the indices of samples with this configuration.
An object of class bammdata.
The expected number of rate shifts under the prior.
Threshold value for marginal posterior-to-prior odds ratios, used to identify branches with elevated shift probabilities relative to prior (core vs non-core shifts).
Other arguments to distinctShiftConfigurations (possibly deprecated args).
Dan Rabosky
See Rabosky et al (2014) and the BAMM project website for
details on the nature of distinct shift configurations and especially
the distinction between "core" and "non-core" rate shifts. Note that
branches with elevated marginal posterior probabilities relative to
the prior (marginal odds ratios) cannot be claimed to have
"significant" evidence for a rate shift on the basis of this evidence
alone.
plot.bammshifts, credibleShiftSet
data(whales, events.whales)
ed <- getEventData(whales, events.whales, burnin=0.25, nsamples=500)
sc <- distinctShiftConfigurations(ed, expectedNumberOfShifts = 1,
threshold = 5)
plot(sc, ed, rank=1)
Run the code above in your browser using DataLab