twoLocusIBD: Estimating two-locus IBD coefficients

In IBDsim: Simulation of Chromosomal Regions Shared by Family Members

Description

Estimates by simulation the two-locus IBD coefficients of a non-inbred pairwise relationship.

Usage

twoLocusIBD(x, ind1, ind2, rho=NULL, cM=NULL, Nsim, Xchrom=FALSE, verbose=TRUE, ...)

Arguments

x	A pedigree in the form of a linkdat object.
ind1, ind2	Numeric ID labels of the two individuals.
rho	NULL, or a number in the interval [0, 0.5]: the recombination fraction between the two loci. If non-NULL, it is converted to centiMorgan using Haldanes map function: cM = -50*log(1-2*rho).
cM	NULL, or a non-negative number: the distance in centiMorgan between the two loci.
Nsim	The number of simulations to be performed.
Xchrom	A logical indicating if the loci are X-linked (if TRUE) or autosomal (FALSE).
verbose	A logical.
...	Further arguments to be passed on to IBDsim.

Details

For a given pair of autosomal loci, the two-locus IBD coefficients of a non-inbred pairwise relationship are defined by

κ[i,j] = Pr(sharing i alleles IBD at locus 1, and j alleles IBD at locus 2),

where 0 ≤ i,j ≤ 2. For X-chromosomal loci, and if at least one of the individuals is male, κ[i,j] is defined only for 0 ≤ i,j ≤ 1.

While the single-locus IBD coefficients κ=(κ[0], κ[1], κ[2]) (see oneLocusIBD) depend only on the genealogy relating the two individuals, the two-locus coefficients also depend on the genetic distance between the loci. In particular, if the loci are completely linked (rho=0; cM=0) we have κ[i,j] = κ[i] if i=j and 0 otherwise. If the loci are unlinked (rho=0.5; cM=Inf) we have κ[i,j] = κ[i]*κ[j]. (See examples.)

Value

A numerical matrix, where the entry in row a and column b is the estimate of κ[a-1, b-1] defined above.

Author(s)

Magnus Dehli Vigeland

See Also

oneLocusIBD, oneLocusJacquard, twoLocusJacquard

Examples

### Example 1: Full siblings
x <- nuclearPed(2) 
Nsim <- 100 # Should be increased substantially

# One-locus kappa estimates (autosomal and X):
k_est = oneLocusIBD(x, ind1=3, ind2=4, Nsim=Nsim, seed=123)
k_est_X = oneLocusIBD(x, ind1=3, ind2=4, Nsim=Nsim, Xchrom=TRUE, seed=123)

### Two-locus IBD estimation
# Completely linked, autosomal
rho = 0 
k2_linked = twoLocusIBD(x, ind1=3, ind2=4, rho=rho, Nsim=Nsim, seed=123)
stopifnot(identical(diag(k2_linked), k_est))

# Completely unlinked, autosomal
rho = 0.5
k2_unlinked = twoLocusIBD(x, ind1=3, ind2=4, rho=rho, Nsim=Nsim, seed=123)
stopifnot(identical(k2_unlinked, outer(k_est, k_est)))

# Recombination rate 10%, autosomal
rho <- 0.1 
cM <- -50*log(1-2*rho)
r1 <- twoLocusIBD(x, ind1=3, ind2=4, rho=rho, Nsim=Nsim, seed=17)
r2 <- twoLocusIBD(x, ind1=3, ind2=4, cM=cM, Nsim=Nsim, seed=17)
stopifnot(identical(r1, r2))

### Two-locus IBD on X
# Completely linked, X chromosome
rho = 0 
k2_linked_X = twoLocusIBD(x, ind1=3, ind2=4, rho=rho, Nsim=Nsim, Xchrom=TRUE, seed=123)
stopifnot(identical(diag(k2_linked_X), k_est_X))

# Completely unlinked, X chromosome
rho = 0.5
k2_unlinked_X = twoLocusIBD(x, ind1=3, ind2=4, rho=rho, Nsim=Nsim, Xchrom=TRUE, seed=123)
stopifnot(identical(k2_unlinked_X, outer(k_est_X, k_est_X)))

# Recombination rate 10%, X chromosome
rho <- 0.1 
cM <- -50*log(1-2*rho)
r1_X <- twoLocusIBD(x, ind1=3, ind2=4, rho=rho, Nsim=Nsim, Xchrom=TRUE, seed=123)
r2_X <- twoLocusIBD(x, ind1=3, ind2=4, cM=cM, Nsim=Nsim, Xchrom=TRUE, seed=123)
stopifnot(identical(r1_X, r2_X))


### Example 2: Comparing half sib, grandparent and uncle
# These are indistinguishable with unlinked loci, see e.g. 
# pages 182-183 in Egeland, Kling and Mostad (2016).
# Each simulations followed by exact counterpart.
x <- addSon(addSon(nuclearPed(2, sex=1:2), 4), 5)
plot(x)
rho <- 0.25
Nsim <- 10 # Should be increased to at least 10000
twoLocusIBD(x, 1, 6, rho=rho, Nsim=Nsim, verbose=FALSE)[2,2]; .5*(1-rho)
twoLocusIBD(x, 8, 6, rho=rho, Nsim=Nsim, verbose=FALSE)[2,2]; .5*(rho^2+(1-rho)^2)
twoLocusIBD(x, 3, 6, rho=rho, Nsim=Nsim, verbose=FALSE)[2,2]; .5*((1-rho)*(rho^2+(1-rho)^2) + rho/2)

### Example 3: X chromosome, granddaughter vs maternal grandfather.
y <- addDaughter(nuclearPed(1, sex=2), 3)
plot(y)
rho <- 0.25
Nsim <-10
twoLocusIBD(y, 1, 5, rho=rho, Nsim=Nsim, Xchrom=TRUE)

# Exact
matrix(c(1-rho, rho, rho, 1-rho)/2, ncol=2) 

IBDsim documentation built on May 2, 2019, 6:01 a.m.