PolyOriginR

PolyOriginR is a wrapper of PolyOrigin.jl (https://github.com/chaozhi/PolyOrigin.jl), for haplotye reconstruction in polyploid multiparental pouplations.

Installation

First download and install Julia available at https://julialang.org/. Then install the development version of PolyOriginR from GitHub with:

# install.packages("devtools")
devtools::install_github("chaozhi/PolyOriginR")

Example

library(PolyOriginR)
polyOriginR("geno.csv","ped.csv",juliapath="C:/path/to/bin",workdir="workdir")

Here julia_home is where julia.exe is located, and workdir is the directory for input and output files. See R script and input files in the "inst/example" folder.

Usage

polyOriginR(
  genofile,
  pedfile,
  juliapath = "",
  epsilon = 0.01,
  seqerr = 0.001,
  chrpairing_phase = 22,
  chrpairing = 44,
  chrsubset = NULL,
  snpthin = 1,
  nworker = 1,
  delsiglevel = 0.05,
  maxstuck = 5,
  maxiter = 30,
  minrun = 3,
  maxrun = 10,
  byparent = TRUE,
  refhapfile = "nothing",
  correctthreshold = 0.15,
  refinemap = FALSE,
  refineorder = FALSE,
  maxwinsize = 50,
  inittemperature = 4,
  coolingrate = 0.5,
  stripdis = 20,
  maxepsilon = 0.5,
  skeletonsize = 50,
  isphysmap = FALSE,
  recomrate = 1,
  isplot = FALSE,
  workdir = getwd(),
  outstem = "outstem",
  verbose = TRUE
)

Arguments

Argument	Description
genofile	Input genotypic data for parents and offspring
pedfile	Input breeding pedigree
juliapath	Path to julia.exe, Default: ''
epsilon	genotyping error probability, Default: 0.01
seqerr	sequencing read error probability for GBS data, Default: 0.001
chrpairing_phase	chromosome pairing in parental phasing, with 22 being only bivalent formations and 44 being bi- and quadri-valent formations, Default: 22
chrpairing	chromosome pairing in offspring decoding, with 22 being only bivalent formations and 44 being bivalent and quadrivalent formations, Default: 44
chrsubset	subset of chromosomes, c(1,10) denotes first and tenth chromosomes, and NULL for all chromosomes, Default: NULL
snpthin	subset of markers, take every snpthin-th markers, Default: 1
nworker	number of parallel workers for computing among chromosomes, nonparallel if nworker=1, Default: 1
delsiglevel	significance level for deleting markers, Default: 0.05
maxstuck	the max number of consecutive iterations that are rejected in a phasing run, Default: 5
maxiter	the max number of iterations in a phasing run, Default: 30
minrun	if the min number of phasing runs that are at the same local maximimum or have the same parental phases reaches minrun, phasing algorithm will stop before reaching the maxrun, Default: 3
maxrun	the max number of phasing runs, Default: 10
byparent	if TRUE, update parental phases parent by parent, Default: TRUE
refhapfile	reference haplotype file for setting absolute parental phases, Default: 'nothing'
correctthreshold	a candidate marker is selected for parental error correction if the fraction of offspring genotypic error >= correctthreshold, Default: 0.15
refinemap	if TRUE, refine marker map, Default: FALSE
refineorder	if TRUE, refine marker mordering, valid only if refinemap=TRUE, Default: FALSE
maxwinsize	max size of sliding windown in map refinning, Default: 50
inittemperature	initial temperature of simulated annealing in map refinning, Default: 4
coolingrate	cooling rate of annealing temperature in map refinning, Default: 0.5
stripdis	a chromosome end in map refinement is removed if it has a distance gap > stripdis(centiMorgan) and it contains less than 5% markers, Default: 20
maxepsilon	markers in map refinement are removed it they have error rates > maxepsilon, Default: 0.5
skeletonsize	the number of markers in the skeleton map that is used to reduce map length inflation by subsampling markers, Default: 50
isphysmap	TRUE, if input marker map in genofile is physical map, Default: FALSE
recomrate	Recombination rate in cM/Mpb, Default: 1
isplot	TURE, if plot haploprob, Default: FALSE
workdir	Work directory, Default: getwd()
outstem	Stem of output files, Default: 'outstem'
verbose	TURE, if print details, Default: TRUE

Details

PolyOriginR is implemented via PolyOriginCmd, which uses PolyOrigin.jl by a command line.

Value

Return 0 if success, and export output files.

Argument	Description
outstem.log	log file
outstem_doseprob.csv	posterior dosage probabilities for all offspring
outstem_parentphased.csv	same as input genofile except parents being phased
outstem_parentphased_corrected.csv	exported if there exist detected parental errors
outstem_polyancestry.csv	genoprob and estimation of chromosome pairing
outstem_genoprob.csv	a simplified version of outstem_polyancestry.csv
outstem_plots	a folder contains plots of haploprob if isplot = true

Here outstem_doseprob.csv, outstem_parentphased.csv, and outstem_parentphased_corrected.csv can be iteratively used as input genofile, so that the step of parental phasing will be skipped. The outstem_doseprob.csv can be used for correcting observed genotypes and imputing missing genotypes, using an appropriate threshould.

Citing PolyOrigin

If you use PolyOriginR in your analyses and publish your results, please cite the article:

Zheng C, Amadeu R, Munoz P, and Endelman J. 2020. Haplotype reconstruction in tetraploid multi-parental populations. In preparation.