X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=R%2FDNA.R;h=e7c0e3a85d74a51518b2a12b1a6593614877d6dc;hb=a03a8c554a6fde0dc4313688e3248bfae2e521e4;hp=b5f472be1ed6f53115220335dee71e00747f2052;hpb=6dfbab243973c0c3fa2e6d02b190aefbe5a67280;p=ape.git diff --git a/R/DNA.R b/R/DNA.R index b5f472b..e7c0e3a 100644 --- a/R/DNA.R +++ b/R/DNA.R @@ -1,12 +1,34 @@ -## DNA.R (2008-01-19) +## DNA.R (2010-03-16) ## Manipulations and Comparisons of DNA Sequences -## Copyright 2002-2008 Emmanuel Paradis +## Copyright 2002-2010 Emmanuel Paradis ## This file is part of the R-package `ape'. ## See the file ../COPYING for licensing issues. +del.gaps <- function(x) +{ + deleteGaps <- function(x) { + i <- which(x == 4) + if (length(i)) x[-i] else x + } + + if (!inherits(x, "DNAbin")) x <- as.DNAbin(x) + if (is.matrix(x)) { + n <- dim(x)[1] + y <- vector("list", n) + for (i in 1:n) y[[i]] <- x[i, ] + names(y) <- rownames(x) + x <- y + rm(y) + } + if (!is.list(x)) return(deleteGaps(x)) + x <- lapply(x, deleteGaps) + class(x) <- "DNAbin" + x +} + as.alignment <- function(x) { if (is.list(x)) n <- length(x) @@ -29,6 +51,7 @@ as.alignment <- function(x) "[.DNAbin" <- function(x, i, j, drop = TRUE) { + oc <- oldClass(x) class(x) <- NULL if (is.matrix(x)) { if (nargs() == 2 && !missing(i)) ans <- x[i] @@ -42,12 +65,12 @@ as.alignment <- function(x) if (missing(i)) i <- 1:length(x) ans <- x[i] } - structure(ans, class = "DNAbin") + class(ans) <- oc + ans } as.matrix.DNAbin <- function(x, ...) { - if (is.matrix(x)) return(x) if (is.list(x)) { if (length(unique(unlist(lapply(x, length)))) != 1) stop("DNA sequences in list not of the same length.") @@ -65,39 +88,72 @@ rbind.DNAbin <- function(...) ### works only with matrices for the moment { obj <- list(...) - nobj <- length(obj) - if (nobj == 1) stop("only one matrix to bind.") - NC <- ncol(obj[[1]]) - for (i in 2:nobj) - if(ncol(obj[[i]]) != NC) + n <- length(obj) + if (n == 1) return(obj[[1]]) + for (i in 1:n) + if (!is.matrix(obj[[1]])) + stop("the 'rbind' method for \"DNAbin\" accepts only matrices") + NC <- unlist(lapply(obj, ncol)) + if (length(unique(NC)) > 1) stop("matrices do not have the same number of columns.") - for (i in 1:nobj) class(obj[[i]]) <- NULL - ans <- obj[[1]] - for (i in 2:nobj) ans <- rbind(ans, obj[[i]]) - structure(ans, class = "DNAbin") + for (i in 1:n) class(obj[[i]]) <- NULL + for (i in 2:n) obj[[1]] <- rbind(obj[[1]], obj[[i]]) + structure(obj[[1]], class = "DNAbin") } -cbind.DNAbin <- function(..., check.names = TRUE) +cbind.DNAbin <- + function(..., check.names = TRUE, fill.with.gaps = FALSE, + quiet = FALSE) ### works only with matrices for the moment { obj <- list(...) - nobj <- length(obj) - if (nobj == 1) stop("only one matrix to bind.") - NR <- nrow(obj[[1]]) - for (i in 2:nobj) - if(nrow(obj[[i]]) != NR) - stop("matrices do not have the same number of rows.") - for (i in 1:nobj) class(obj[[i]]) <- NULL - nms <- rownames(obj[[1]]) + n <- length(obj) + if (n == 1) return(obj[[1]]) + for (i in 1:n) + if (!is.matrix(obj[[1]])) + stop("the 'cbind' method for \"DNAbin\" accepts only matrices") + NR <- unlist(lapply(obj, nrow)) + for (i in 1:n) class(obj[[i]]) <- NULL if (check.names) { - for (i in 2:nobj) - if (all(rownames(obj[[i]]) %in% nms)) - obj[[i]] <- obj[[i]][nms, ] - else stop("rownames do not match among matrices.") + nms <- unlist(lapply(obj, rownames)) + if (fill.with.gaps) { + NC <- unlist(lapply(obj, ncol)) + nms <- unique(nms) + ans <- matrix(as.raw(4), length(nms), sum(NC)) + rownames(ans) <- nms + from <- 1 + for (i in 1:n) { + to <- from + NC[i] - 1 + tmp <- rownames(obj[[i]]) + nmsi <- tmp[tmp %in% nms] + ans[nmsi, from:to] <- obj[[i]][nmsi, , drop = FALSE] + from <- to + 1 + } + } else { + tab <- table(nms) + ubi <- tab == n + nms <- names(tab)[which(ubi)] + ans <- obj[[1]][nms, , drop = FALSE] + for (i in 2:n) + ans <- cbind(ans, obj[[i]][nms, , drop = FALSE]) + if (!quiet && !all(ubi)) + warning("some rows were dropped.") + } + } else { + if (length(unique(NR)) > 1) + stop("matrices do not have the same number of rows.") + ans <- matrix(unlist(obj), NR) + rownames(ans) <- rownames(obj[[1]]) } - ans <- matrix(unlist(obj), NR) - rownames(ans) <- nms - structure(ans, class = "DNAbin") + class(ans) <- "DNAbin" + ans +} + +c.DNAbin <- function(..., recursive = FALSE) +{ + if (!all(unlist(lapply(list(...), is.list)))) + stop("the 'c' method for \"DNAbin\" accepts only lists") + structure(NextMethod("c"), class = "DNAbin") } print.DNAbin <- function(x, ...) @@ -120,8 +176,16 @@ summary.DNAbin <- function(object, printlen = 6, digits = 3, ...) cat("Label:", nms, "\n\n") } else { cat(n, "DNA sequences in binary format stored in a list.\n\n") - cat("Summary of sequence lengths:\n") - print(summary(unlist(lapply(object, length)))) + tmp <- unlist(lapply(object, length)) + mini <- min(tmp) + maxi <- max(tmp) + if (mini == maxi) + cat("All sequences of same length:", maxi, "\n") + else { + cat("Mean sequence length:", round(mean(tmp), 3), "\n") + cat(" Shortest sequence:", mini, "\n") + cat(" Longest sequence:", maxi, "\n") + } TAIL <- "\n\n" if (printlen < n) { nms <- nms[1:printlen] @@ -205,45 +269,31 @@ as.character.DNAbin <- function(x, ...) if (is.list(x)) lapply(x, f) else f(x) } -base.freq <- function(x) +base.freq <- function(x, freq = FALSE) { if (is.list(x)) x <- unlist(x) n <- length(x) - BF <- .C("BaseProportion", as.raw(x), as.integer(n), - double(4), PACKAGE = "ape")[[3]] + BF <- .C("BaseProportion", x, n, double(4), freq, + DUP = FALSE, NAOK = TRUE, PACKAGE = "ape")[[3]] names(BF) <- letters[c(1, 3, 7, 20)] BF } -GC.content <- function(x) -{ - BF <- base.freq(x) - sum(BF[2:3]) -} +GC.content <- function(x) sum(base.freq(x)[2:3]) seg.sites <- function(x) { - n <- dim(x) - s <- n[2] - n <- n[1] - ans <- .C("SegSites", x, as.integer(n), as.integer(s), - integer(s), PACKAGE = "ape") - which(as.logical(ans[[4]])) -} - -nuc.div <- function(x, variance = FALSE, pairwise.deletion = FALSE) -{ - if (pairwise.deletion && variance) - warning("cannot compute the variance of nucleotidic diversity\nwith pairwise deletion: try 'pairwise.deletion = FALSE' instead.") if (is.list(x)) x <- as.matrix(x) - n <- dim(x)[1] - ans <- sum(dist.dna(x, "raw", pairwise.deletion = pairwise.deletion))/ - (n*(n - 1)/2) - if (variance) { - var <- (n + 1)*ans/(3*(n + 1)*s) + 2*(n^2 + n + 3)*ans/(9*n*(n - 1)) - ans <- c(ans, var) + if (is.vector(x)) n <- 1 + else { # 'x' is a matrix + n <- dim(x) + s <- n[2] + n <- n[1] } - ans + if (n == 1) return(integer(0)) + ans <- .C("SegSites", x, n, s, integer(s), + DUP = FALSE, NAOK = TRUE, PACKAGE = "ape") + which(as.logical(ans[[4]])) } dist.dna <- function(x, model = "K80", variance = FALSE, gamma = FALSE, @@ -251,8 +301,11 @@ dist.dna <- function(x, model = "K80", variance = FALSE, gamma = FALSE, as.matrix = FALSE) { MODELS <- c("RAW", "JC69", "K80", "F81", "K81", "F84", "T92", "TN93", - "GG95", "LOGDET", "BH87", "PARALIN") - imod <- which(MODELS == toupper(model)) + "GG95", "LOGDET", "BH87", "PARALIN", "N") + imod <- pmatch(toupper(model), MODELS) + if (is.na(imod)) + stop(paste("'model' must be one of:", + paste("\"", MODELS, "\"", sep = "", collapse = " "))) if (imod == 11 && variance) { warning("computing variance temporarily not available for model BH87.") variance <- FALSE @@ -268,9 +321,8 @@ dist.dna <- function(x, model = "K80", variance = FALSE, gamma = FALSE, n <- n[1] BF <- if (is.null(base.freq)) base.freq(x) else base.freq if (!pairwise.deletion) { - keep <- .C("GlobalDeletionDNA", x, as.integer(n), - as.integer(s), as.integer(rep(1, s)), - PACKAGE = "ape")[[4]] + keep <- .C("GlobalDeletionDNA", x, n, s, + rep(1L, s), PACKAGE = "ape")[[4]] x <- x[, as.logical(keep)] s <- dim(x)[2] } @@ -278,10 +330,10 @@ dist.dna <- function(x, model = "K80", variance = FALSE, gamma = FALSE, var <- if (variance) double(Ndist) else 0 if (!gamma) gamma <- alpha <- 0 else alpha <- gamma <- 1 - d <- .C("dist_dna", x, as.integer(n), as.integer(s), - as.integer(imod), double(Ndist), BF, + d <- .C("dist_dna", x, n, s, imod, double(Ndist), BF, as.integer(pairwise.deletion), as.integer(variance), - var, as.integer(gamma), alpha, PACKAGE = "ape") + var, as.integer(gamma), alpha, DUP = FALSE, NAOK = TRUE, + PACKAGE = "ape") if (variance) var <- d[[9]] d <- d[[5]] if (imod == 11) {