1 ## dist.topo.R (2008-06-28)
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$edge[, 2] %in% (n + which(!found2))])
57 .compressTipLabel <- function(x)
59 ## 'x' is a list of objects of class "phylo" possibly with no class
60 if (!is.null(attr(x, "TipLabel"))) return(x)
61 ref <- x[[1]]$tip.label
62 if (any(table(ref) != 1))
63 stop("some tip labels are duplicated in tree no. 1")
65 for (i in 2:length(x)) {
66 if (identical(x[[i]]$tip.label, ref)) next
67 ilab <- match(x[[i]]$tip.label, ref)
68 ## can use tabulate here because 'ilab' contains integers
69 if (any(tabulate(ilab) > 1))
70 stop(paste("some tip labels are duplicated in tree no.", i))
72 stop(paste("tree no.", i, "has different tip labels"))
73 ie <- match(1:n, x[[i]]$edge[, 2])
74 x[[i]]$edge[ie, 2] <- ilab
76 for (i in 1:length(x)) x[[i]]$tip.label <- NULL
77 attr(x, "TipLabel") <- ref
81 prop.part <- function(..., check.labels = TRUE)
84 if (length(obj) == 1 && class(obj[[1]]) != "phylo")
87 ## class(obj) <- NULL # needed?
90 if (check.labels) obj <- .compressTipLabel(obj)
91 for (i in 1:ntree) storage.mode(obj[[i]]$Nnode) <- "integer"
93 ## The 1st must have tip labels
94 ## Maybe simply pass the number of tips to the C code??
95 if (!is.null(attr(obj, "TipLabel")))
96 for (i in 1:ntree) obj[[i]]$tip.label <- attr(obj, "TipLabel")
98 clades <- .Call("prop_part", obj, ntree, TRUE, PACKAGE = "ape")
99 attr(clades, "number") <- attr(clades, "number")[1:length(clades)]
100 attr(clades, "labels") <- obj[[1]]$tip.label
101 class(clades) <- "prop.part"
105 print.prop.part <- function(x, ...)
107 if (is.null(attr(x, "labels"))) {
108 for (i in 1:length(x)) {
109 cat("==>", attr(x, "number")[i], "time(s):")
110 print(x[[i]], quote = FALSE)
113 for (i in 1:length(attr(x, "labels")))
114 cat(i, ": ", attr(x, "labels")[i], "\n", sep = "")
116 for (i in 1:length(x)) {
117 cat("==>", attr(x, "number")[i], "time(s):")
118 print(x[[i]], quote = FALSE)
123 summary.prop.part <- function(object, ...) attr(object, "number")
125 plot.prop.part <- function(x, barcol = "blue", leftmar = 4, ...)
127 if (is.null(attr(x, "labels")))
128 stop("cannot plot this partition object; see ?prop.part for details.")
130 n <- length(attr(x, "labels"))
131 layout(matrix(1:2, 2, 1), heights = c(1, 3))
132 par(mar = c(0.1, leftmar, 0.1, 0.1))
133 plot(1:L, attr(x, "number"), type = "h", col = barcol, xlim = c(1, L),
134 xlab = "", ylab = "Frequency", xaxt = "n", bty = "n")
135 plot(0, type = "n", xlim = c(1, L), ylim = c(1, n),
136 xlab = "", ylab = "", xaxt = "n", yaxt = "n")
137 for (i in 1:L) points(rep(i, length(x[[i]])), x[[i]], ...)
138 mtext(attr(x, "labels"), side = 2, at = 1:n, las = 1)
141 prop.clades <- function(phy, ..., part = NULL)
145 if (length(obj) == 1 && class(obj[[1]]) != "phylo")
146 obj <- unlist(obj, recursive = FALSE)
147 part <- prop.part(obj, check.labels = TRUE)
149 bp <- .Call("bipartition", phy$edge, length(phy$tip.label),
150 phy$Nnode, PACKAGE = "ape")
151 if (!is.null(attr(part, "labels")))
152 for (i in 1:length(part))
153 part[[i]] <- sort(attr(part, "labels")[part[[i]]])
154 bp <- lapply(bp, function(xx) sort(phy$tip.label[xx]))
155 n <- numeric(phy$Nnode)
156 for (i in 1:phy$Nnode) {
157 for (j in 1:length(part)) {
158 if (identical(all.equal(bp[[i]], part[[j]]), TRUE)) {
159 n[i] <- attr(part, "number")[j]
168 boot.phylo <- function(phy, x, FUN, B = 100, block = 1, trees = FALSE)
171 if (class(x) == "DNAbin") x <- as.matrix(x)
177 x <- matrix(x, n, nL/n, byrow = TRUE)
181 boot.tree <- vector("list", B)
184 y <- seq(block, ncol(x), block)
185 boot.i <- sample(y, replace = TRUE)
186 boot.samp <- numeric(ncol(x))
187 boot.samp[y] <- boot.i
188 for (j in 1:(block - 1))
189 boot.samp[y - j] <- boot.i - j
190 } else boot.samp <- sample(ncol(x), replace = TRUE)
191 boot.tree[[i]] <- FUN(x[, boot.samp])
193 for (i in 1:B) storage.mode(boot.tree[[i]]$Nnode) <- "integer"
194 storage.mode(phy$Nnode) <- "integer"
195 ans <- attr(.Call("prop_part", c(list(phy), boot.tree),
196 B + 1, FALSE, PACKAGE = "ape"), "number") - 1
197 if (trees) ans <- list(BP = ans, trees = boot.tree)
201 consensus <- function(..., p = 1, check.labels = TRUE)
203 foo <- function(ic, node) {
205 ## node: node number in the final tree
208 for (j in (ic + 1):m) {
209 wh <- match(pp[[j]], pool)
210 if (!any(is.na(wh))) {
211 edge[pos, 1] <<- node
213 edge[pos, 2] <<- nextnode <<- nextnode + 1L
221 ind <- pos:(pos + size - 1)
222 edge[ind, 1] <<- node
223 edge[ind, 2] <<- pool
228 if (length(obj) == 1) {
229 ## better than unlist(obj, recursive = FALSE)
230 ## because "[[" keeps the class of 'obj':
232 if (class(obj) == "phylo") return(obj)
234 if (!is.null(attr(obj, "TipLabel")))
235 labels <- attr(obj, "TipLabel")
237 labels <- obj[[1]]$tip.label
238 if (check.labels) obj <- .compressTipLabel(obj)
241 ## Get all observed partitions and their frequencies:
242 pp <- prop.part(obj, check.labels = FALSE)
243 ## Drop the partitions whose frequency is less than 'p':
244 pp <- pp[attr(pp, "number") >= p * ntree]
245 ## Get the order of the remaining partitions by decreasing size:
246 ind <- sort(unlist(lapply(pp, length)), decreasing = TRUE,
247 index.return = TRUE)$ix
251 edge <- matrix(0L, n + m - 1, 2)
260 structure(list(edge = edge, tip.label = labels,
261 Nnode = m), class = "phylo")