X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=R%2Fread.dna.R;h=9e46cba67bbc81b5eed7624232e5c8185519e5c2;hb=be6a044200152fd83d0b72f348012a0adc2593c5;hp=f9513eab3ec9b8ff1a76b59e7cd7d55c64d1b6f2;hpb=ba59c6508990bce5fa5071c0fb346e39d2b2a325;p=ape.git diff --git a/R/read.dna.R b/R/read.dna.R index f9513ea..9e46cba 100644 --- a/R/read.dna.R +++ b/R/read.dna.R @@ -1,103 +1,127 @@ -## read.dna.R (2008-03-28) +## read.dna.R (2012-12-27) ## Read DNA Sequences in a File -## Copyright 2003-2008 Emmanuel Paradis +## Copyright 2003-2012 Emmanuel Paradis ## This file is part of the R-package `ape'. ## See the file ../COPYING for licensing issues. +read.FASTA <- function(file) +{ + sz <- file.info(file)$size + x <- readBin(file, "raw", sz) + if (Sys.info()[1] == "Windows") { + icr <- which(x == as.raw(0x0d)) # CR + x <- x[-icr] + } + res <- .Call("rawStreamToDNAbin", x, PACKAGE = "ape") + class(res) <- "DNAbin" + res +} + read.dna <- function(file, format = "interleaved", skip = 0, - nlines = 0, comment.char = "#", seq.names = NULL, - as.character = FALSE) + nlines = 0, comment.char = "#", + as.character = FALSE, as.matrix = NULL) { + findFirstNucleotide <- function(x) { + ## actually find the 1st non-blank character + ## just in case: pat.base <- "[-AaCcGgTtUuMmRrWwSsYyKkVvHhDdBbNn?]{10}" + tmp <- regexpr("[[:blank:]]+", x[1]) # consider only a single string + tmp[1] + attr(tmp, "match.length") + } getTaxaNames <- function(x) { - x <- sub("^ +", "", x) # remove the leading spaces - x <- sub(" +$", "", x) # remove the trailing spaces - x <- sub("^['\"]", "", x) # remove the leading quotes - x <- sub("['\"]$", "", x) # remove the trailing quotes + x <- sub("^['\" ]+", "", x) # remove the leading quotes and spaces + x <- sub("['\" ]+$", "", x) # " " trailing " " " x } - format <- match.arg(format, c("interleaved", "sequential", "fasta")) - phylip <- if (format %in% c("interleaved", "sequential")) TRUE else FALSE - X <- scan(file = file, what = character(), sep = "\n", quiet = TRUE, - skip = skip, nlines = nlines, comment.char = comment.char) - if (phylip) { - fl <- X[1] - oop <- options(warn = -1) - ## need to remove the possible leading spaces in the first line - fl.num <- as.numeric(unlist(strsplit(gsub("^ +", "", fl), " +"))) - options(oop) - if (all(is.na(fl.num))) - stop("the first line of the file must contain the dimensions of the data") - if (length(fl.num) != 2) - stop("the first line of the file must contain TWO numbers") - else { - n <- fl.num[1] - s <- fl.num[2] - } - X <- X[-1] - obj <- vector("character", n*s) - dim(obj) <- c(n, s) - } - if (format == "interleaved") { - fl <- X[1] - fl <- unlist(strsplit(fl, NULL)) - bases <- grep("[-AaCcGgTtUuMmRrWwSsYyKkVvHhDdBbNn?]", fl) - z <- diff(bases) - for (i in 1:length(z)) if (all(z[i:(i + 8)] == 1)) break - start.seq <- bases[i] - if (is.null(seq.names)) - seq.names <- getTaxaNames(substr(X[1:n], 1, start.seq - 1)) - X[1:n] <- substr(X[1:n], start.seq, nchar(X[1:n])) - X <- gsub(" ", "", X) - nl <- length(X) - for (i in 1:n) - obj[i, ] <- unlist(strsplit(X[seq(i, nl, n)], NULL)) - } - if (format == "sequential") { - fl <- X[1] - taxa <- character(n) - j <- 1 - for (i in 1:n) { - bases <- grep("[-AaCcGgTtUuMmRrWwSsYyKkVvHhDdBbNn?]", - unlist(strsplit(X[j], NULL))) - z <- diff(bases) - for (k in 1:length(z)) if (all(z[k:(k + 8)] == 1)) break - start.seq <- bases[k] - taxa[i] <- substr(X[j], 1, start.seq - 1) - sequ <- substr(X[j], start.seq, nchar(X[j])) - sequ <- gsub(" ", "", sequ) - j <- j + 1 - while (nchar(sequ) < s) { - sequ <- paste(sequ, gsub(" " , "", X[j]), sep = "") - j <- j + 1 - } - obj[i, ] <- unlist(strsplit(sequ, NULL)) - } - if (is.null(seq.names)) seq.names <- getTaxaNames(taxa) + getNucleotide <- function(x) { + x <- gsub(" ", "", x) + x <- strsplit(x, NULL) + tolower(unlist(x)) } + formats <- c("interleaved", "sequential", "fasta", "clustal") + format <- match.arg(format, formats) if (format == "fasta") { - start <- grep("^ {0,}>", X) - taxa <- X[start] - n <- length(taxa) - obj <- vector("list", n) - if (is.null(seq.names)) { - taxa <- sub("^ {0,}>", "", taxa) # remove the hook and the spaces before - seq.names <- getTaxaNames(taxa) + obj <- read.FASTA(file) + } else { + X <- scan(file = file, what = "", sep = "\n", quiet = TRUE, + skip = skip, nlines = nlines, comment.char = comment.char) + + if (format %in% formats[1:2]) { + ## need to remove the possible leading spaces and/or tabs in the first line + fl <- gsub("^[[:blank:]]+", "", X[1]) + fl <- as.numeric(unlist(strsplit(fl, "[[:blank:]]+"))) + if (length(fl) != 2 || any(is.na(fl))) + stop("the first line of the file must contain the dimensions of the data") + n <- fl[1] + s <- fl[2] + obj <- matrix("", n, s) + X <- X[-1] } - start <- c(start, length(X) + 1) # this avoids the following to crash when `i = n' - for (i in 1:n) - obj[[i]] <- unlist(strsplit(gsub(" ", "", - X[(start[i] + 1):(start[i + 1] - 1)]), - NULL)) - } - if (phylip) { - rownames(obj) <- seq.names - obj <- tolower(obj) + switch(format, + "interleaved" = { + start.seq <- findFirstNucleotide(X[1]) + one2n <- 1:n + taxa <- getTaxaNames(substr(X[one2n], 1, start.seq - 1)) + X[one2n] <- substr(X[one2n], start.seq, nchar(X[one2n])) + nl <- length(X) + for (i in one2n) + obj[i, ] <- getNucleotide(X[seq(i, nl, n)]) + }, + "sequential" = { + taxa <- character(n) + j <- 1L # line number + for (i in 1:n) { + start.seq <- findFirstNucleotide(X[j]) + taxa[i] <- getTaxaNames(substr(X[j], 1, start.seq - 1)) + sequ <- getNucleotide(substr(X[j], start.seq, nchar(X[j]))) + j <- j + 1L + while (length(sequ) < s) { + sequ <- c(sequ, getNucleotide(X[j])) + j <- j + 1L + } + obj[i, ] <- sequ + } + taxa <- getTaxaNames(taxa) + }, + "clustal" = { + X <- X[-1] # drop the line with "Clustal bla bla..." + ## find where the 1st sequence starts + start.seq <- findFirstNucleotide(X[1]) + ## find the lines with *********.... + nspaces <- paste("^ {", start.seq - 1, "}", sep = "", collapse = "") + stars <- grep(nspaces, X) + ## we now know how many sequences in the file: + n <- stars[1] - 1 + taxa <- getTaxaNames(substr(X[1:n], 1, start.seq - 1)) + ## need to remove the sequence names before getting the sequences: + X <- substr(X, start.seq, nchar(X)) + nl <- length(X) + ## find the length of the 1st sequence: + tmp <- getNucleotide(X[seq(1, nl, n + 1)]) + s <- length(tmp) + obj <- matrix("", n, s) + obj[1, ] <- tmp + for (i in 2:n) + obj[i, ] <- getNucleotide(X[seq(i, nl, n + 1)]) + }) + + if (format != "fasta") { + rownames(obj) <- taxa } else { - names(obj) <- seq.names - obj <- lapply(obj, tolower) + LENGTHS <- unique(unlist(lapply(obj, length))) + allSameLength <- length(LENGTHS) == 1 + if (is.logical(as.matrix)) { + if (as.matrix && !allSameLength) + stop("sequences in FASTA file not of the same length") + } else { + as.matrix <- allSameLength + } + if (as.matrix) { + obj <- matrix(unlist(obj), ncol = LENGTHS, byrow = TRUE) + rownames(obj) <- taxa + } } if (!as.character) obj <- as.DNAbin(obj) obj