reGenotyper: Detecting mislabeled samples, recovering the optimal...
In reGenotyper: Detecting Mislabeled Samples in Genetic Data

Description Usage Arguments Value Author(s) References See Also Examples

Main function to detect mislabeled samples using perturbation strategy

1
2
3

reGenotyper(phenotype, genotype, fileName = "test", thres = 0.9, optGT = TRUE,
            optGTplot = FALSE,optGT.thres = 0, permu = FALSE, n.permu = 10, 
            wls.score.permu = NULL, process = TRUE, t.thres = 1.5, GT.ref=NULL)

`phenotype`	phenotype data: a nTrait-by-nSample matrix
`genotype`	genotype data: a nMarker-by-nSample matrix with two allels being 0 and 1 (or A and B) or three allels being 0, 0.5 and 1 (or, A, H, and B), where 0.5 (or H) represents heterozygous allele.
`fileName`	output file name. If NULL (default) it produces files starting with "test"
`thres`	probability threshold to decide if a sample is mislabled based on permutation result (Default=0.9).
`optGT`	recovered optimal genotype from the given phenotype
`optGTplot`	If TRUE it produces a plot of the genotype with two colors: green and red color indicate the original genotype of a sample (column) at certain marker (row) is correct or correct, respectively.
`optGT.thres`	threshold to decide if thr original genotype is correct
`permu`	If TRUE permutation is performed to estimate the likelihood of each sample being mislabled.
`n.permu`	The number of permutation to be performed. `n.permu`=1000 is usually recommended for a reliable estimate but it can take long time.
`wls.score.permu`	A vector with element being WLS score from permutation which can be obtained using function permutation: e.g. wls.score.permu <- permutation(phenotype,genotype,n.permu=1000,process=TRUE,fileName="test",t.thres=3)
`process`	If TRUE, it prints which step has been finished. Default = TRUE.
`t.thres`	threshold for deciding significant QTLs (t.test) that will be used to detecting mislabled samples
`GT.ref`	reference gentoype data from a large collection of strains. This is used to search for best mached gentoype for identified mislabeled samples. Default= NULL. If GT.ref is NULL, the orginal input genotype data willl beused to seach for best matched genotype for identified mislabeled samples.

An object of class wls. A list with elements:

`wls.score`	a vector with length being the number of samples; each element gives the score for the sample being mislabeled
`wls.names`	the names of sample that being detected as mislabeled using the Z score method
`gt.opt`	recovered the optimal genotype based on the given phenotype data
`wls.pValue`	p value for each sample using permutation, only when `permu`=TRUE
`wls.score.permu`	a vector with the length of n.permu. Each element represents the score of a randomly selected sample with permutated genotype, only when `permu`=TRUE.
`thres`	threshold used probability threshold to decide if a sample is mislabled based on permutation result

Yang Li <yang.li@rug.nl>

Li Y. et al, reGenotyper: detecting mislabeled samples in genetic data (submitted)

optimalGT, permutation,
tMatFunction,genotype,
phenotype

  library(reGenotyper)
  #load example genotype and phenotype data
  data(genotype)
  data(phenotype)
  ### For this test dataset 5 permutations is enough. In real case at least few hundreds 
  ### of permutations are needed.
  wlsObject <- reGenotyper(phenotype, genotype, fileName = "test", thres = 0.9, optGT = TRUE, 
  optGTplot = FALSE,   optGT.thres = 0,  permu = TRUE, n.permu = 5, wls.score.permu = NULL, 
  process = TRUE, t.thres = 1.5, GT.ref=NULL)
  ###Inspecting the output
  wlsObject
  plot(wlsObject)
  ### previous line takes around 30s to execute, you can also load the result:
  data(wlsObject)