R/import.ucsc.files.R

In drmjc/genomics: Genomics related functions.

#' import the knownGene file downloaded manually from UCSC genome browser.
#' 
#' By default it looks in the /biomirror/hgdownload.cse.ucsc.edu/tables for
#' a file called knownGene
#' #name chrom strand txStart txEnd cdsStart cdsEnd <snip>
#' NM_001011874 chr1 - 3174829 3630690 3176369 3630540 <snip>
#' AK035064 chr1 + 3709446 3783878 3709490 3783206 <snip>
#' NM_011283 chr1 - 4304522 4320770 4304977 4313203 <snip>
#' NM_011441 chr1 - 4451317 4456803 4452105 4453796 <snip>
#' <snip>
#' 
#' @param file path to knownGene file
#' @return see import.ucsc.file
#' 
#' @author Mark Cowley, 8 August 2006
#' @export
import.knowngene <- function(file="/biomirror/hgdownload.cse.ucsc.edu/tables/knownGene") {
	import.ucsc.file(file)
}


#' import the kgXref file downloaded manually from UCSC genome browser.
#' 
#' By default it looks in the /biomirror/hgdownload.cse.ucsc.edu/tables for
#' a file called kgXref.gz
#' #kgID mRNA spID spDisplayID geneSymbol <snip>
#' AB001750 AB001750 O35551 RABE1_MOUSE AB001750 <snip>
#' AB001926 AB001926 P48722-2 P48722-2 AB001926 <snip>
#' AB003147 AB003147 O54770 O54770_MOUSE AB003147 <snip>
#' AB003503 AB003503 O88180 O88180_MOUSE AB003503 <snip>
#' <snip>
#' 
#' @param file path to kgXref file
#' @return see import.ucsc.file
#' 
#' @author Mark Cowley, 8 August 2006
#' @export
#' @importFrom mjcbase file.gunzip
import.kgXref <- function(file="/biomirror/hgdownload.cse.ucsc.edu/tables/kgXref.gz") {
	tmp <- paste0(tempfile(pattern="ucsc"), ".gz")
	file.copy(file, tmp)
	file.gunzip(tmp)
	tmp <- sub("\\.gz", "", tmp)

	res <- import.ucsc.file(tmp)

	unlink(tmp)

	return( res )
}


#' import UCSC-style file
#' 
#' File from UCSC tend to have the column names in the first line preceeded
#' with a hash character.
#' 
#' This function imports the data and fixes the names to remove
#' this hash char.
#' 
#' @param file the path to a file
#' @return a \code{data.frame}
#' 
#' @author Mark Cowley, 8 August 2006
#' @export
import.ucsc.file <- function(file) {
	res <- read.delim(file, as.is=T, comment.char="")
	colnames(res)[1] <- "name"

	return( res )
}



#' Download knownGeneToLocusLink from UCSC Table Browser to: Import the
#' knownToLocusLink file, downloaded from the UCSC Table Browser. 
#' 
#' @param taxid the numerical taxonID
#' @param files a list with your file name in it.
#' 
#' @author Mark Cowley, 30/3/07
#' @export
import.knownGene2gene <- function(taxid=TAXID.MOUSE,
                                  files=list(mouse="/biomirror/ucsc/tables/mm8/knownToLocusLink.gz",
                                             human="/biomirror/ucsc/tables/hg18/knownToLocusLink.gz",
                                             rat="/biomirror/ucsc/tables/rn4/knownToLocusLink.gz")
                                  ) {

    if( taxid == TAXID.MOUSE )
        file <- files$"mouse"
    else if( taxid == TAXID.HUMAN )
        file <- files$"human"
    else if( taxid == TAXID.RAT )
        file <- files$"rat"
    else {
        if( (is.list(files) && length(files) == 1) && file.exists(files[1]) )
            file <- files[[1]]
        else if( is.character(files) && length(files) == 1 && file.exists(files[1]) )
            file <- files[1]
        else
            stop('unsupported taxid; try setting taxid and files=list("/path/to/your/taxid/knownToLocusLink.gz")')
    }


    tmp <- tempfile()
    file <- file.gunzip(file, tmp)
    res <- import.ucsc.file(tmp)
    unlink(tmp)
    colnames(res) <- c("knownGeneID", "GeneID")
    res <- res[order(res$knownGeneID),]
    # res <- rm.duplicate.rows(res, issorted=T)
    res <- res[!duplicated(res), ]

    rownames(res) <- 1:nrow(res)

    res
}