Heritability estimation based on genomic relationship matrix using JAGS

Usage

h2.jags(
  y,
  x,
  G,
  eps = 1e-04,
  sigma.p = 0,
  sigma.r = 1,
  parms = c("b", "p", "r", "h2"),
  ...
)

Arguments

y: outcome vector.
x: covariate matrix.
G: genomic relationship matrix.
eps: a positive diagonal perturbation to G.
sigma.p: initial parameter values.
sigma.r: initial parameter values.
parms: monitored parmeters.
...: parameters passed to jags, e.g., n.chains, n.burnin, n.iter.

Value

The returned value is a fitted model from jags().

Details

This function performs Bayesian heritability estimation using genomic relationship matrix.

References

Zhao JH, Luan JA, Congdon P (2018). “Bayesian Linear Mixed Models with Polygenic Effects.” Journal of Statistical Software, 85(6), 1 - 27. doi:10.18637/jss.v085.i06 .

Author

Jing Hua Zhao keywords htest

Examples

if (FALSE) { # \dontrun{
require(gap.datasets)
set.seed(1234567)
meyer <- within(meyer,{
    y[is.na(y)] <- rnorm(length(y[is.na(y)]),mean(y,na.rm=TRUE),sd(y,na.rm=TRUE))
    g1 <- ifelse(generation==1,1,0)
    g2 <- ifelse(generation==2,1,0)
    id <- animal
    animal <- ifelse(!is.na(animal),animal,0)
    dam <- ifelse(!is.na(dam),dam,0)
    sire <- ifelse(!is.na(sire),sire,0)
})
G <- kin.morgan(meyer)$kin.matrix*2
library(regress)
r <- regress(y~-1+g1+g2,~G,data=meyer)
r
with(r,h2G(sigma,sigma.cov))
eps <- 0.001
y <- with(meyer,y)
x <- with(meyer,cbind(g1,g2))
ex <- h2.jags(y,x,G,sigma.p=0.03,sigma.r=0.014)
print(ex)
require(coda)
ex.mcmc <- as.mcmc(ex)
traceplot(ex.mcmc)
densplot(ex.mcmc)
} # }