1 ## dist.topo.R (2008-02-27)
3 ## Topological Distances, Tree Bipartitions,
4 ## Consensus Trees, and Bootstrapping Phylogenies
6 ## Copyright 2005-2008 Emmanuel Paradis
8 ## This file is part of the R-package `ape'.
9 ## See the file ../COPYING for licensing issues.
11 dist.topo <- function(x, y, method = "PH85")
13 if (method == "BHV01" && (is.null(x$edge.length) || is.null(y$edge.length)))
14 stop("trees must have branch lengths for Billera et al.'s distance.")
15 n <- length(x$tip.label)
16 bp1 <- .Call("bipartition", x$edge, n, x$Nnode, PACKAGE = "ape")
17 bp1 <- lapply(bp1, function(xx) sort(x$tip.label[xx]))
18 bp2 <- .Call("bipartition", y$edge, n, y$Nnode, PACKAGE = "ape")
19 bp2 <- lapply(bp2, function(xx) sort(y$tip.label[xx]))
22 if (method == "PH85") {
26 if (identical(all.equal(bp1[[i]], bp2[[j]]), TRUE)) {
32 dT <- if (q1 == q2) 2*(q1 - p) else 2*(min(q1, q2) - p) + abs(q1 - q2)
34 if (method == "BHV01") {
41 if (identical(bp1[[i]], bp2[[j]])) {
42 dT <- dT + abs(x$edge.length[which(x$edge[, 2] == n + i)] -
43 y$edge.length[which(y$edge[, 2] == n + j)])
44 found1 <- found2[j] <- TRUE
48 if (found1) found1 <- FALSE
49 else dT <- dT + x$edge.length[which(x$edge[, 2] == n + i)]
52 dT <- dT + sum(y$edge.length[y$edge[, 2] %in% (n + which(!found2))])
57 prop.part <- function(..., check.labels = FALSE)
60 if (length(obj) == 1 && class(obj[[1]]) != "phylo")
61 obj <- unlist(obj, recursive = FALSE)
64 for (i in 1:ntree) storage.mode(obj[[i]]$Nnode) <- "integer"
65 clades <- .Call("prop_part", obj, ntree, TRUE, PACKAGE = "ape")
66 attr(clades, "number") <- attr(clades, "number")[1:length(clades)]
67 attr(clades, "labels") <- obj[[1]]$tip.label
69 bp <- .Call("bipartition", obj[[1]]$edge, length(obj[[1]]$tip.label),
70 obj[[1]]$Nnode, PACKAGE = "ape")
71 clades <- lapply(bp, function(xx) sort(obj[[1]]$tip.label[xx]))
72 no <- rep(1, length(clades))
76 bp <- .Call("bipartition", obj[[k]]$edge,
77 length(obj[[k]]$tip.label), obj[[k]]$Nnode,
79 bp <- lapply(bp, function(xx) sort(obj[[k]]$tip.label[xx]))
80 for (i in 1:length(bp)) {
82 for (j in 1:length(clades)) {
83 if (identical(all.equal(bp[[i]], clades[[j]]), TRUE)) {
90 clades <- c(clades, bp[i])
96 attr(clades, "number") <- no
98 class(clades) <- "prop.part"
102 print.prop.part <- function(x, ...)
104 if (is.null(attr(x, "labels"))) {
105 for (i in 1:length(x)) {
106 cat("==>", attr(x, "number")[i], "time(s):")
107 print(x[[i]], quote = FALSE)
110 for (i in 1:length(attr(x, "labels")))
111 cat(i, ": ", attr(x, "labels")[i], "\n", sep = "")
113 for (i in 1:length(x)) {
114 cat("==>", attr(x, "number")[i], "time(s):")
115 print(x[[i]], quote = FALSE)
120 summary.prop.part <- function(object, ...) attr(object, "number")
122 plot.prop.part <- function(x, barcol = "blue", leftmar = 4, ...)
124 if (is.null(attr(x, "labels")))
125 stop("cannot plot this partition object; see ?prop.part for details.")
127 n <- length(attr(x, "labels"))
128 layout(matrix(1:2, 2, 1), heights = c(1, 3))
129 par(mar = c(0.1, leftmar, 0.1, 0.1))
130 plot(1:L, attr(x, "number"), type = "h", col = barcol, xlim = c(1, L),
131 xlab = "", ylab = "Number", xaxt = "n", bty = "n")
132 plot(0, type = "n", xlim = c(1, L), ylim = c(1, n),
133 xlab = "", ylab = "", xaxt = "n", yaxt = "n")
134 for (i in 1:L) points(rep(i, length(x[[i]])), x[[i]], ...)
135 mtext(attr(x, "labels"), side = 2, at = 1:n, las = 1)
138 prop.clades <- function(phy, ..., part = NULL)
142 if (length(obj) == 1 && class(obj[[1]]) != "phylo")
143 obj <- unlist(obj, recursive = FALSE)
144 part <- prop.part(obj, check.labels = TRUE)
146 bp <- .Call("bipartition", phy$edge, length(phy$tip.label),
147 phy$Nnode, PACKAGE = "ape")
148 if (!is.null(attr(part, "labels")))
149 for (i in 1:length(part))
150 part[[i]] <- sort(attr(part, "labels")[part[[i]]])
151 bp <- lapply(bp, function(xx) sort(phy$tip.label[xx]))
152 n <- numeric(phy$Nnode)
153 for (i in 1:phy$Nnode) {
154 for (j in 1:length(part)) {
155 if (identical(all.equal(bp[[i]], part[[j]]), TRUE)) {
156 n[i] <- attr(part, "number")[j]
165 boot.phylo <- function(phy, x, FUN, B = 100, block = 1)
168 if (class(x) == "DNAbin") x <- as.matrix(x)
174 x <- matrix(x, n, nL/n, byrow = TRUE)
178 boot.tree <- vector("list", B)
181 y <- seq(block, ncol(x), block)
182 boot.i <- sample(y, replace = TRUE)
183 boot.samp <- numeric(ncol(x))
184 boot.samp[y] <- boot.i
185 for (j in 1:(block - 1))
186 boot.samp[y - j] <- boot.i - j
187 } else boot.samp <- sample(ncol(x), replace = TRUE)
188 boot.tree[[i]] <- FUN(x[, boot.samp])
190 for (i in 1:B) storage.mode(boot.tree[[i]]$Nnode) <- "integer"
191 storage.mode(phy$Nnode) <- "integer"
192 attr(.Call("prop_part", c(list(phy), boot.tree), B + 1, FALSE,
193 PACKAGE = "ape"), "number") - 1
196 consensus <- function(..., p = 1)
199 if (length(obj) == 1 && class(obj[[1]]) != "phylo")
200 obj <- unlist(obj, recursive = FALSE)
202 ## Get all observed partitions and their frequencies:
203 pp <- prop.part(obj, check.labels = TRUE)
204 ## Drop the partitions whose frequency is less than 'p':
205 pp <- pp[attr(pp, "number") >= p * ntree]
206 ## Get the order of the remaining partitions by decreasing size:
207 ind <- rev(sort(unlist(lapply(pp, length)),
208 index.return = TRUE)$ix)
209 pp <- lapply(pp, function(xx) paste("IMPROBABLE_PREFIX", xx,
210 "IMPROBABLE_SUFFIX", sep = "_"))
211 STRING <- paste(pp[[1]], collapse = ",")
212 STRING <- paste("(", STRING, ");", sep = "")
214 ## 1. Delete all tips in the focus partition:
215 STRING <- unlist(strsplit(STRING, paste(pp[[i]], collapse = "|")))
216 ## 2. Put the partition in any of the created gaps:
217 STRING <- c(STRING[1],
218 paste("(", paste(pp[[i]], collapse = ","), ")", sep = ""),
220 ## 3. Stick back the Newick string:
221 STRING <- paste(STRING, collapse = "")
223 ## Remove the extra commas:
224 STRING <- gsub(",{2,}", ",", STRING)
225 STRING <- gsub("\\(,", "\\(", STRING)
226 STRING <- gsub(",\\)", "\\)", STRING)
227 STRING <- gsub("IMPROBABLE_PREFIX_", "", STRING)
228 STRING <- gsub("_IMPROBABLE_SUFFIX", "", STRING)
229 read.tree(text = STRING)