removeOutlierByPCs: Remove population outliers

In transbioZI/Gimpute: Gimpute: An efficient genetic data processing and imputation pipeline

Remove population outliers

Description

Remove population outliers by using principle component analysis.

Usage

removeOutlierByPCs(
  plink,
  gcta,
  inputPrefix,
  nThread = 20,
  cutoff = NULL,
  cutoffSign,
  inputPC4subjFile,
  outputPC4outlierFile,
  outputPCplotFile,
  outputPrefix
)

Arguments

plink	an executable program in either the current working directory or somewhere in the command path.
gcta	an executable program in either the current working directory or somewhere in the command path.
inputPrefix	the prefix of the input PLINK binary files.
nThread	the number of threads used for computation. The default is 20.
cutoff	the cutoff that distinguishes the outliers from ordinary population using PCA. If it is null, then there are no outliers or outliers are not required to be removed. The default is NULL.
cutoffSign	the cutoff sign: 'greater' or 'smaller' that determines if the outliers should be greater or smaller than the cutoff value.
inputPC4subjFile	the pure text file that stores all the subject IDs and their corresponding eigenvalues of the first two principle components.
outputPC4outlierFile	the pure text file that stores the outlier IDs and their corresponding eigenvalues of the first two principle components.
outputPCplotFile	the plot file for visualizing the first two principle components of all subjects without population outliers.
outputPrefix	the prefix of the output PLINK binary files.

Details

This function is used for removing population outliers. If the outliers are necessary to be removed, then one uses the eigenvalues from the first principle component as a criterion to find out the outliers by assigning an appropriate cutoff.

Value

1.) The output PLINK binary files after outlier removal. 2.) The output pure text file (if any) for storing removed outlier IDs and their corresponding PCs. 3.) The plot file (if any) for visualizing the first two principle components after outlier removal.

Author(s)

Junfang Chen

Examples

 
## In the current working directory
bedFile <- system.file("extdata", "QCdata.bed", package="Gimpute")
bimFile <- system.file("extdata", "QCdata.bim", package="Gimpute") 
famFile <- system.file("extdata", "QCdata.fam", package="Gimpute")
system(paste0("scp ", bedFile, bimFile, famFile, " ."))  
inputPrefix <- "QCdata"  
cutoff <- NULL ## no outlier to be removed
cutoffSign <- "greater" ## not used if cutoff == NULL
inputPC4subjFile <- "2_13_eigenvalAfterQC.txt"
outputPC4outlierFile <- "2_13_eigenval4outliers.txt"
outputPCplotFile <- "2_13_removedOutliers.png"
outputPrefix <- "2_13_removedOutliers" 
## Not run: Requires an executable program PLINK and GCTA, e.g.
## plink <- "/home/tools/plink"
## gcta <- "/home/tools/gcta64"
## removeOutlierByPCs(plink, gcta, inputPrefix, nThread=20, 
##                    cutoff, cutoffSign, inputPC4subjFile, 
##                    outputPC4outlierFile, outputPCplotFile, outputPrefix)

transbioZI/Gimpute documentation built on April 10, 2022, 4:20 a.m.