beta.dosage {hierfstat}

Estimates pairwise kinships and individual inbreeding coefficients from dosage data

Description

Estimates pairwise kinships (coancestries) and individual inbreeding coefficient using Weir and Goudet (2017) beta estimator.

Usage

beta.dosage(dos,inb=TRUE,Mb=FALSE,MATCHING=FALSE)

Arguments

dos

A matrix of 0, 1 and 2s with loci (SNPs) in columns and individuals in rows. Missing values are allowed

inb

whether individual inbreeding coefficient should be estimated (rather than self-coancestries)

Mb

whether to output the mean matching

MATCHING

if MATCHING=FALSE, dos is a (ni x nl) dosage matrix; if MATCHING=TRUE, dos is a (ni x ni) matrix of matching proportions, as obtained from a call to the matching function

Details

This function is written for dosage data, i.e., how many doses of an allele (0, 1 or 2) an individual carries. It should be use for bi-allelic markers only (e.g. SNPs), although you might "force" a k multiallelic locus to k biallelic loci (see fstat2dos).

Matching proportions can be obtained by the following equation: M=\beta*(1-Mb)+Mb

By default (inb=TRUE) the inbreeding coefficient is returned on the main diagonal. With inb=FALSE, self coancestries are reported.

Twice the betas with self-coancestries on the diagonal gives the Genomic Relationship Matrix (GRM)

Following a suggestion from Olivier Hardy, missing data are removed from the estimation procedure, rather than imputed (this is taken care off automatically)

Value

if Mb=FALSE, a matrix of pairwise kinships and inbreeding coefficients (if inb=TRUE) or self-coancestries (inb=FALSE); if Mb=TRUE, a list with elements inb (whether inbreeding coefficients rather than kinships should be returned on the main diagonal), MB (the average off-diagonal matching) and betas the kinships or inbreeding coefficients.

Author(s)

Jerome Goudet [e-mail]

References

Weir, BS and Goudet J. 2017 A Unified Characterization of Population Structure and Relatedness. Genetics (2017) 206:2085

Goudet, J., Kay, T. and Weir BS. 2018 How to estimate kinship. Molecular Ecology 27:4121.

Examples

## Not run: 
 dos<-matrix(sample(0:2,size=10000,replace=TRUE),ncol=100)
 beta.dosage(dos,inb=TRUE)
 
 #matrix of kinship/inbreeding coeff
 data(gtrunchier)
 beta.dosage(fstat2dos(gtrunchier[,-c(1:2)]))
 
 #individual inbreeding coefficients
 dat<-sim.genot(size=100,nbloc=100,nbal=20,mig=0.01,f=c(0,0.3,0.7))
 hist(diag(beta.dosage(fstat2dos(dat[,-1]))),breaks=-10:100/100,main="",xlab="",ylab="")
 abline(v=c(0.0,0.3,0.7),col="red")
 #only 20 loci
 hist(diag(beta.dosage(fstat2dos(dat[,2:21]))),breaks=-5:20/20,main="",xlab="",ylab="")
 abline(v=c(0.0,0.3,0.7),col="red")
 

## End(Not run)