R/read.dna.R

   1 ## read.dna.R (2013-01-31)
   2
   3 ##   Read DNA Sequences in a File
   4
   5 ## Copyright 2003-2013 Emmanuel Paradis
   6
   7 ## This file is part of the R-package `ape'.
   8 ## See the file ../COPYING for licensing issues.
   9
  10 read.FASTA <- function(file)
  11 {
  12     sz <- file.info(file)$size
  13     x <- readBin(file, "raw", sz)
  14     icr <- which(x == as.raw(0x0d)) # CR
  15     x <- x[-icr]
  16     res <- .Call("rawStreamToDNAbin", x, PACKAGE = "ape")
  17     names(res) <- sub("^ +", "", names(res)) # to permit phylosim
  18     class(res) <- "DNAbin"
  19     res
  20 }
  21
  22 read.dna <- function(file, format = "interleaved", skip = 0,
  23                      nlines = 0, comment.char = "#",
  24                      as.character = FALSE, as.matrix = NULL)
  25 {
  26     findFirstNucleotide <- function(x) {
  27         ## actually find the 1st non-blank character
  28         ## just in case: pat.base <- "[-AaCcGgTtUuMmRrWwSsYyKkVvHhDdBbNn?]{10}"
  29         tmp <- regexpr("[[:blank:]]+", x[1]) # consider only a single string
  30         tmp[1] + attr(tmp, "match.length")
  31     }
  32     getTaxaNames <- function(x) {
  33         x <- sub("^['\" ]+", "", x) # remove the leading quotes and spaces
  34         x <- sub("['\" ]+$", "", x) #   "     "  trailing  "     "    "
  35         x
  36     }
  37     getNucleotide <- function(x) {
  38         x <- gsub(" ", "", x)
  39         x <- strsplit(x, NULL)
  40         tolower(unlist(x))
  41     }
  42     formats <- c("interleaved", "sequential", "fasta", "clustal")
  43     format <- match.arg(format, formats)
  44     if (format == "fasta") {
  45         obj <- read.FASTA(file)
  46     } else {
  47         X <- scan(file = file, what = "", sep = "\n", quiet = TRUE,
  48                   skip = skip, nlines = nlines, comment.char = comment.char)
  49         if (format %in% formats[1:2]) {
  50             ## need to remove the possible leading spaces and/or tabs in the first line
  51             fl <- gsub("^[[:blank:]]+", "", X[1])
  52             fl <- as.numeric(unlist(strsplit(fl, "[[:blank:]]+")))
  53             if (length(fl) != 2 || any(is.na(fl)))
  54                 stop("the first line of the file must contain the dimensions of the data")
  55             n <- fl[1]
  56             s <- fl[2]
  57             obj <- matrix("", n, s)
  58             X <- X[-1]
  59         }
  60         switch(format,
  61                "interleaved" = {
  62                    start.seq <- findFirstNucleotide(X[1])
  63                    one2n <- 1:n
  64                    taxa <- getTaxaNames(substr(X[one2n], 1, start.seq - 1))
  65                    X[one2n] <- substr(X[one2n], start.seq, nchar(X[one2n]))
  66                    nl <- length(X)
  67                    for (i in one2n)
  68                        obj[i, ] <- getNucleotide(X[seq(i, nl, n)])
  69                },
  70                "sequential" = {
  71                    taxa <- character(n)
  72                    j <- 1L # line number
  73                    for (i in 1:n) {
  74                        start.seq <- findFirstNucleotide(X[j])
  75                        taxa[i] <- getTaxaNames(substr(X[j], 1, start.seq - 1))
  76                        sequ <- getNucleotide(substr(X[j], start.seq, nchar(X[j])))
  77                        j <- j + 1L
  78                        while (length(sequ) < s) {
  79                            sequ <- c(sequ, getNucleotide(X[j]))
  80                            j <- j + 1L
  81                        }
  82                        obj[i, ] <- sequ
  83                    }
  84                    taxa <- getTaxaNames(taxa)
  85                },
  86                "clustal" = {
  87                    X <- X[-1] # drop the line with "Clustal bla bla..."
  88                    ## find where the 1st sequence starts
  89                    start.seq <- findFirstNucleotide(X[1])
  90                    ## find the lines with *********....
  91                    nspaces <- paste("^ {", start.seq - 1, "}", sep = "", collapse = "")
  92                    stars <- grep(nspaces, X)
  93                    ## we now know how many sequences in the file:
  94                    n <- stars[1] - 1
  95                    taxa <- getTaxaNames(substr(X[1:n], 1, start.seq - 1))
  96                    ## need to remove the sequence names before getting the sequences:
  97                    X <- substr(X, start.seq, nchar(X))
  98                    nl <- length(X)
  99                    ## find the length of the 1st sequence:
 100                    tmp <- getNucleotide(X[seq(1, nl, n + 1)])
 101                    s <- length(tmp)
 102                    obj <- matrix("", n, s)
 103                    obj[1, ] <- tmp
 104                    for (i in 2:n)
 105                        obj[i, ] <- getNucleotide(X[seq(i, nl, n + 1)])
 106                })
 107     }
 108     if (format != "fasta") {
 109         rownames(obj) <- taxa
 110         if (!as.character) obj <- as.DNAbin(obj)
 111     } else {
 112         LENGTHS <- unique(unlist(lapply(obj, length)))
 113         allSameLength <- length(LENGTHS) == 1
 114         if (is.logical(as.matrix)) {
 115             if (as.matrix && !allSameLength)
 116                 stop("sequences in FASTA file not of the same length")
 117         } else {
 118             as.matrix <- allSameLength
 119         }
 120         if (as.matrix) {
 121             taxa <- names(obj)
 122             n <- length(obj)
 123             y <- matrix(as.raw(0), n, LENGTHS)
 124             for (i in seq_len(n)) y[i, ] <- obj[[i]]
 125             obj <- y
 126             rownames(obj) <- taxa
 127             class(obj) <- "DNAbin"
 128         }
 129         if (as.character) obj <- as.character(obj)
 130     }
 131     obj
 132 }