1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
|
\name{ancr}
\alias{ancr}
\alias{hide.hidden}
\alias{plot.ancr}
\title{Compute marginal or joint ancestral state estimates}
\usage{
ancr(object, ...)
hide.hidden(object, ...)
\method{plot}{ancr}(x, args.plotTree=list(...), args.nodelabels=list(...), ...)
}
\arguments{
\item{object}{fitted model (presently object of class \code{"bounded_bm"}, \code{"fitgammaMk"}, \code{"fitHRM"}, \code{"fitMk"}, \code{"fitpolyMk"}, \code{"fitThresh"}, or \code{"fitmultiBM"}), or a set of models in the form of a table from \code{anova} comparison. For \code{hide.hidden}, an object of class \code{"ancr"}.}
\item{...}{optional arguments. Depending on the input object class, may include \code{type} (whether to undertake marginal or joint estimation, defaults to \code{type="marginal"}) and \code{local} (whether to do global or local estimation, following Pagel 1999, defaults to \code{local=FALSE}).}
\item{x}{in the case of \code{plot.ancr}, an object of class \code{"ancr"}.}
\item{args.plotTree}{arguments to be passed to \code{\link{plotTree}}, in a list.}
\item{args.nodelabels}{arguments to be passed to \code{\link[ape]{nodelabels}}, in a list.}
}
\description{
By default, \code{ancr} computes marginal ancestral states, also known as empirical Bayes posterior probabilities, conditioning on the fitted (or set) model of \code{object}.
Can also perform \emph{joint} ancestral state estimation, if the optional argument \code{type} is set to \code{type="joint"}.
\code{hide.hidden} merges hidden states (if any).
}
\details{
If the optional argument \code{tips=TRUE}, then the matrix returned contains empirical Bayes posterior probabilities (marginal scaled likelihoods) for both tips \emph{and} internal nodes. Otherwise (the default) only node marginal states are returned.
If the input object is a set of models (in the form of an \code{anova} table), then \code{ancr} will compute model-averaged marginal ancestral states (for \code{type="marginal"}, unless the optional argument \code{weighted=FALSE}, in which case only the best-supported model is used.
}
\value{
An object of class \code{"ancr"}.
For most input obect types, in the case of \code{type="marginal"}, this object consists of a matrix of marginal (empirical Bayes) probabilities and a likelihood. In the case of \code{type="joint"}, the object contains a set of most-likely internal node states stored in a data frame.
}
\references{
Pagel, M. (1999) The Maximum Likelihood approach to reconstructing ancestral character states of discrete characters on phylogenies. \emph{Systematic Biology}, \bold{3}, 612-622.
Revell, L. J. (2024) phytools 2.0: an updated R ecosystem for phylogenetic comparative methods (and other things). \emph{PeerJ}, \bold{12}, e16505.
}
\author{Liam Revell \email{liam.revell@umb.edu}}
\seealso{
\code{\link{fitMk}}
}
\examples{
## load tree and data from Revell & Collar (2009)
data(sunfish.tree)
data(sunfish.data)
## extract discrete character (feeding mode)
fmode<-setNames(sunfish.data$feeding.mode,
rownames(sunfish.data))
## fit ARD model
ard_fmode<-fitMk(sunfish.tree,fmode,model="ARD",
pi="fitzjohn")
## compute ancestral states
anc_fmode<-ancr(ard_fmode)
## plot the results
par(mfrow=c(2,1))
cols<-setNames(c("blue","red"),levels(fmode))
plot(anc_fmode,
args.plotTree=list(lwd=2,direction="upwards",
mar=c(0.1,1.1,2.1,1.1),fsize=0.8),
args.nodelabels=list(piecol=cols),
args.tiplabels=list(cex=0.3),
legend="bottomright")
mtext("a) marginal states under ARD model",adj=0)
## fit ER model
er_fmode<-fitMk(sunfish.tree,fmode,model="ER",
pi="fitzjohn")
## compare models
aov_fmode<-anova(er_fmode,ard_fmode)
## compute model-averaged ancestral states
anc_fmode_model.averaged<-ancr(aov_fmode)
plot(anc_fmode_model.averaged,
args.plotTree=list(lwd=2,direction="upwards",
mar=c(0.1,1.1,2.1,1.1),fsize=0.8),
args.nodelabels=list(piecol=cols),
args.tiplabels=list(cex=0.3),
legend="bottomright")
mtext("b) marginal states model-averaging ER & ARD models",
adj=0)
## reset par to default
par(mar=c(5.1,4.1,4.1,2.1),mfrow=c(1,1))
}
\keyword{ancestral states}
\keyword{phylogenetics}
\keyword{comparative method}
\keyword{maximum likelihood}
|
