-## dist.topo.R (2010-05-25)
+## dist.topo.R (2011-07-13)
## Topological Distances, Tree Bipartitions,
## Consensus Trees, and Bootstrapping Phylogenies
-## Copyright 2005-2010 Emmanuel Paradis
+## Copyright 2005-2011 Emmanuel Paradis
## This file is part of the R-package `ape'.
## See the file ../COPYING for licensing issues.
dist.topo <- function(x, y, method = "PH85")
{
if (method == "score" && (is.null(x$edge.length) || is.null(y$edge.length)))
- stop("trees must have branch lengths for Billera et al.'s distance.")
+ stop("trees must have branch lengths for branch score distance.")
nx <- length(x$tip.label)
x <- unroot(x)
y <- unroot(y)
if (any(table(ref) != 1))
stop("some tip labels are duplicated in tree no. 1")
n <- length(ref)
- for (i in 2:length(x)) {
- label <- x[[i]]$tip.label
- if (!identical(label, ref)) {
- if (length(label) != length(ref))
- stop(paste("tree no.", i, "has a different number of tips"))
- ilab <- match(label, ref)
- ## can use tabulate here because 'ilab' contains integers
- if (any(is.na(ilab)))
- stop(paste("tree no.", i, "has different tip labels"))
+ Ntree <- length(x)
+ if (Ntree > 1) {
+ for (i in 2:Ntree) {
+ label <- x[[i]]$tip.label
+ if (!identical(label, ref)) {
+ if (length(label) != length(ref))
+ stop(paste("tree no.", i, "has a different number of tips"))
+ ilab <- match(label, ref)
+ ## can use tabulate here because 'ilab' contains integers
+ if (any(is.na(ilab)))
+ stop(paste("tree no.", i, "has different tip labels"))
### <FIXME> the test below does not seem useful anymore
### if (any(tabulate(ilab) > 1))
### stop(paste("some tip labels are duplicated in tree no.", i))
### </FIXME>
- ie <- match(1:n, x[[i]]$edge[, 2])
- x[[i]]$edge[ie, 2] <- ilab
+ ie <- match(1:n, x[[i]]$edge[, 2])
+ x[[i]]$edge[ie, 2] <- ilab
+ }
+ x[[i]]$tip.label <- NULL
}
- x[[i]]$tip.label <- NULL
}
x[[1]]$tip.label <- NULL
attr(x, "TipLabel") <- ref
if (length(obj) == 1 && class(obj[[1]]) != "phylo")
obj <- obj[[1]]
## <FIXME>
- ## class(obj) <- NULL # needed?
+ ## class(obj) <- NULL # needed? apparently not, see below (2010-11-18)
## </FIXME>
ntree <- length(obj)
if (ntree == 1) check.labels <- FALSE
- if (check.labels) obj <- .compressTipLabel(obj)
+ if (check.labels) obj <- .compressTipLabel(obj) # fix by Klaus Schliep (2011-02-21)
for (i in 1:ntree) storage.mode(obj[[i]]$Nnode) <- "integer"
## <FIXME>
## The 1st must have tip labels
## Maybe simply pass the number of tips to the C code??
- if (!is.null(attr(obj, "TipLabel")))
- for (i in 1:ntree) obj[[i]]$tip.label <- attr(obj, "TipLabel")
+ obj <- .uncompressTipLabel(obj) # fix a bug (2010-11-18)
## </FIXME>
clades <- .Call("prop_part", obj, ntree, TRUE, PACKAGE = "ape")
attr(clades, "number") <- attr(clades, "number")[1:length(clades)]
n
}
-boot.phylo <- function(phy, x, FUN, B = 100, block = 1, trees = FALSE)
+boot.phylo <- function(phy, x, FUN, B = 100, block = 1,
+ trees = FALSE, quiet = FALSE)
{
if (is.list(x) && !is.data.frame(x)) {
if (inherits(x, "DNAbin")) x <- as.matrix(x)
}
}
boot.tree <- vector("list", B)
+ if (!quiet) progbar <- utils::txtProgressBar(style = 3) # suggestion by Alastair Potts
for (i in 1:B) {
if (block > 1) {
y <- seq(block, ncol(x), block)
boot.samp[y - j] <- boot.i - j
} else boot.samp <- sample(ncol(x), replace = TRUE)
boot.tree[[i]] <- FUN(x[, boot.samp])
+ if (!quiet) utils::setTxtProgressBar(progbar, i/B)
}
+ if (!quiet) close(progbar)
for (i in 1:B) storage.mode(boot.tree[[i]]$Nnode) <- "integer"
storage.mode(phy$Nnode) <- "integer"
- ans <- attr(.Call("prop_part", c(list(phy), boot.tree),
- B + 1, FALSE, PACKAGE = "ape"), "number") - 1
- if (trees) ans <- list(BP = ans, trees = boot.tree)
+ ans <- prop.clades(phy, boot.tree)
+ ##ans <- attr(.Call("prop_part", c(list(phy), boot.tree),
+ ## B + 1, FALSE, PACKAGE = "ape"), "number") - 1
+ if (trees) {
+ class(boot.tree) <- "multiPhylo"
+ ans <- list(BP = ans, trees = boot.tree)
+ }
ans
}
## Get all observed partitions and their frequencies:
pp <- prop.part(obj, check.labels = FALSE)
## Drop the partitions whose frequency is less than 'p':
+ if (p == 0.5) p <- 0.5000001 # avoid incompatible splits
pp <- pp[attr(pp, "number") >= p * ntree]
## Get the order of the remaining partitions by decreasing size:
ind <- sort(unlist(lapply(pp, length)), decreasing = TRUE,