grm.pair: GRM for a pair of individuals.

In bowenwang7/r3: Realized Relatedness Estimation and Simulation

Description

grm.pair computes relatedness estimates between two individuals.

Usage

grm.pair(geno1, geno2, freq, method = "twostep", weights = NULL,
  init.est = NULL)

Arguments

geno1, geno2	numeric vector.
freq	numeric vector, values between 0 and 1.
method	string.
weights	numeric vector, values between 0 and 1.
init.est	numeric.

Details

geno1 and geno2 are vectors of counts of reference alleles. freq is the vector of reference allele frequencies.

The default method is "twostep", other options include "classic", "robust" and "general". When using the default "twostep" method, user can supply an initial estimate through init.est to bypass the first step. When "general" is selected, weights must also be specified. The difference between the two-step GRM, classic GRM and robust GRM is discussed in Wang et al. (2017).

Value

An estimate of realized relatedness.

References

Wang et al. (2017) Genetics 205:1063-1078, https://www.ncbi.nlm.nih.gov/pubmed/28100587.

See Also

grm.matrix

Examples

# simulate genotypes for a full sib pair
pedigree = as.character(rep(1, 4))
member = as.character(c(11, 12, 21, 22))
sex = as.numeric(c(1, 2, 1, 2))
father = as.character(c(NA, NA, 11, 11))
mother = as.character(c(NA, NA, 12, 12))
pedinfo = data.frame(pedigree, member, sex, father, mother, stringsAsFactors = FALSE)
set.seed(1)
inher = sim.recomb(pedinfo, 3500) # on a hypothetical chromosome

nsnp = 100000
marker = seq(0,3500,length.out=nsnp)
freq = runif(nsnp, 0.05, 0.95)
haplo = sim.haplotype(freq, 4)
geno = populate.snp(inher, haplo, marker, output.allele = FALSE)

# simulation truth
ibd.proportion(inher,3,4)

# different GRM estimates
grm.pair(geno[3,], geno[4,], freq, method = "twostep")
grm.pair(geno[3,], geno[4,], freq, method = "classic")
grm.pair(geno[3,], geno[4,], freq, method = "robust")
grm.pair(geno[3,], geno[4,], freq, method = "general", weights = sample(freq, nsnp)/sum(freq))

# compute the relatedness matrix
grm.matrix(geno, freq)
grm.matrix(geno, freq, method = "robust")

bowenwang7/r3 documentation built on May 17, 2019, 6:39 p.m.