gl.filter.pa.r

#' Filters loci that contain private (and fixed alleles) between two populations
#'
#' This script is meant to be used prior to \code{gl.nhybrids} to maximise the
#' information content of the SNPs used to identify hybrids (currently
#'  newhybrids does allow only 200 SNPs). The idea is to use first all loci that
#'   have fixed alleles between the potential source populations and then 'fill
#'    up' to 200 loci using loci that have private alleles between those. The
#'    functions filters for those loci (if invers is set to TRUE, the opposite
#'    is returned (all loci that are not fixed and have no private alleles - not
#'     sure why yet, but maybe useful.)
#'
#' @param x Name of the genlight object containing the SNP data [required].
#' @param pop1 Name of the first parental population (in quotes) [required].
#' @param pop2 Name of the second parental population (in quotes) [required].
#' @param invers Switch to filter for all loci that have no private alleles and
#' are not fixed [FALSE].
#' @param verbose Verbosity: 0, silent or fatal errors; 1, begin and end; 2,
#' progress log; 3, progress and results summary; 5, full report
#' [default 2, unless specified using gl.set.verbosity].
#' @return The reduced genlight dataset, containing now only fixed and private
#' alleles.
#' @export
#' @family filter functions
#' @author Authors: Bernd Gruber & Ella Kelly (University of Melbourne);
#' Custodian: Luis Mijangos -- Post to
#' \url{https://groups.google.com/d/forum/dartr}
#' @examples
#' result <- gl.filter.pa(testset.gl, pop1=pop(testset.gl)[1], 
#' pop2=pop(testset.gl)[2],verbose=3)

gl.filter.pa <- function(x,
                         pop1,
                         pop2,
                         invers = FALSE,
                         verbose = NULL) {
    # SET VERBOSITY
    verbose <- gl.check.verbosity(verbose)
    
    # FLAG SCRIPT START
    funname <- match.call()[[1]]
    utils.flag.start(func = funname,
                     build = "Jody",
                     verbose = verbose)
    
    # CHECK DATATYPE
    datatype <- utils.check.datatype(x, verbose = verbose)
    
    # FUNCTION SPECIFIC ERROR CHECKING
    
    # DO THE JOB
    
    pops <- seppop(x)
    p1 <- as.matrix(pops[[pop1]])
    p2 <- as.matrix(pops[[pop2]])
    p1alf <- colMeans(p1, na.rm = T) / 2
    p2alf <- colMeans(p2, na.rm = T) / 2
    # private alleles for pop 1
    priv1 <-
        c(names(p1alf)[p2alf == 0 &
                           p1alf != 0], names(p1alf)[p2alf == 1 &
                                                         p1alf != 1])  
    # private alleles for pop 2
    priv2 <-
        c(names(p2alf)[p1alf == 0 &
                           p2alf != 0], names(p2alf)[p1alf == 1 &
                                                         p2alf != 1])  
    pfLoci <- unique(c(priv1, priv2))  # put all together
    index <- locNames(x) %in% pfLoci
    if (invers){
        index <- !index
    }

      x2 <- x[, index]
      x2@other$loc.metrics <- x@other$loc.metrics[index, ]

    # ADD TO HISTORY
    nh <- length(x2@other$history)
    x2@other$history[[nh + 1]] <- match.call()
    
    # FLAG SCRIPT END
    
    if (verbose > 0) {
        cat(report("Completed:", funname, "\n"))
    }
    
    return(x2)
}