R/genepop_newhybrids.R

In rystanley/genepopedit: Simple and flexible manipulation of genomic data

# Genepop -> New Hybrids
#' @title Convert Genepop to New Hybrids format.
#' @description Function to convert Genepop to NewHybrids
#' @param genepop the genepop data to be manipulated. This can be either a file path
#' or a dataframe read in with tab separation, header=FALSE , quote="", and stringsAsFactors=FALSE.
#' This will be the standard genepop format with the first n+1 rows corresponding to the n loci names,
#' or a single comma delimited row of loci names followed by the locus data. Populations are
#' separated by "Pop". Each individual ID is linked to the locus data by " ,  " (space,space space) and is read in as
#' as a single row (character).
#' @param path the filepath and filename of output.
#' @rdname genepop_newhybrids
#' @importFrom data.table fread as.data.table
#' @importFrom utils write.table
#' @export


genepop_newhybrids <- function(genepop,path=NULL){

  #Check to see if genepop is a data.frame from the workspace and convert to data.table
  if(is.data.frame(genepop)){genepop <- as.data.table(genepop)}

  #Check to see if genepop is a file path or dataframe
  if(is.character(genepop)){
    genepop <- fread(genepop,
                                 header = FALSE, sep = "\t",
                                 stringsAsFactors = FALSE)
  }

  ## check if loci names are read in as one large character vector (1 row)
  header <- genepop[1,]
  if(length(gregexpr(',', header, fixed=F)[[1]])>1){
    lociheader <- strsplit(header,",")
    lociheader <- gsub(" ","",unlist(lociheader))
    #remove the first column of loci names
    genepop <- as.vector(genepop)
    genepop <- genepop[-1,]
    genepop <- c(lociheader,genepop)
    genepop <- as.data.table(genepop,stringsAsFactors = FALSE)
  }

  ## Stacks version information
  stacks.version <- genepop[1,] #this could be blank or any other source. First row is ignored by genepop

  #Remove first label of the stacks version
  genepop <- genepop[-1,]
  colnames(genepop) <- "data"

  #ID the rows which flag the Populations
  Pops  <-  which(genepop$data == "Pop" | genepop$data =="pop" | genepop$data == "POP")
  npops  <-  1:length(Pops)

  ## separate the data into the column headers and the rest
  ColumnData <- genepop$data[1:(Pops[1]-1)]
  ColumnData <- gsub("\r","",ColumnData)#remove any hidden carriage returns
  snpData <- genepop[Pops[1]:NROW(genepop),]

  #Get a datafile with just the snp data no pops
  tempPops <- which(snpData$data=="Pop"| snpData$data =="pop" | snpData$data == "POP") ## Changed because we allowed
  snpData <- snpData[-tempPops,]

  #separate the snpdata
  temp <- as.data.frame(do.call(rbind, strsplit(snpData$data," ")))

  #data format check
  if(unique(temp[,2])!="," | !length(which(temp[,3]==""))>1){
    stop("Genepop sampleID delimiter not in proper format. Ensure sampleIDs are separated from loci by ' ,  ' (space comma space space). Function stopped.",call. = FALSE)
  }

  temp2 <- temp[,4:length(temp)] #split characters by spaces

  #Contingency to see if R read in the top line as the "stacks version"
  if (length(temp2)!=length(ColumnData)){colnames(temp2) <- c(stacks.version,ColumnData)}
  if (length(temp2)==length(ColumnData)){colnames(temp2) <- ColumnData}
  if (length(temp2)!=length(ColumnData)){stacks.version="No STACKS version specified"}

  #stacks version character
  stacks.version <- as.character(stacks.version)

  #paste together the Loci as one long integer separated for each loci by a space
  Loci <- do.call(paste,c(temp2[,], sep=" "))

  #paste the loci and row individual count together.
  temp3=paste(1:nrow(temp),Loci,sep=" ")

#compile the output
Output=c(paste0("NumIndivs ",nrow(temp)),
         paste0("NumLoci ",length(temp2)),
         paste0("Digits ",nchar(as.character(temp2[1,1]))/2),
         "Format Lumped",
         paste0("LocusNames ",do.call(paste,c(as.list(colnames(temp2))))),
         temp3)

utils::write.table(Output,path,col.names=FALSE,row.names=FALSE,quote=FALSE)

}