hwe.ibf {HWEintrinsic}

Testing Hardy-Weinberg Equilibrium Using an Intrinsic Prior Approach

Description

This function implements the exact calculation of the Bayes factor based on intrinsic priors for the Hardy-Weinberg testing problem as described in Consonni et al. (2011).

Usage

hwe.ibf(y, t)

Arguments

y

an object of class "HWEdata".

t

training sample size.

Details

This function implements the exact formula for the Bayes factor based on intrinsic priors.

Value

hwe.ibf returns the value of the Bayes factor based on intrinsic priors.

Note

The Bayes factor computed here is for the unrestricted model (M_1) against the Hardy-Weinberg case (M_0). This function provides the output only for the two alleles case.

Author(s)

Sergio Venturini [e-mail]

References

Consonni, G., Moreno, E., and Venturini, S. (2011). "Testing Hardy-Weinberg equilibrium: an objective Bayesian analysis". Statistics in Medicine, 30, 62–74. https://onlinelibrary.wiley.com/doi/10.1002/sim.4084/abstract

See Also

hwe.ibf.plot, hwe.ibf.mc.

Examples

## Not run: 
# ATTENTION: the following code may take a long time to run! #

data(Lindley)
hwe.ibf.exact <- Vectorize(hwe.ibf, "t")
f <- seq(.05, 1, .05)
n <- sum([e-mail], na.rm = TRUE)

# Dataset 1 #
plot(dataL1)
npp.exact <- 1/(1 + hwe.ibf.exact(round(f*n), y = dataL1))
npp.std <- 1/(1 + hwe.bf(dataL1))
plot(f, npp.exact, type="l", lwd = 2, xlab = "f = t/n",
	ylab = "Null posterior probability")
abline(h = npp.std, col = gray(.5), lty = "longdash")

# Dataset 2 #
plot(dataL2)
npp.exact <- 1/(1 + hwe.ibf.exact(round(f*n), y = dataL2))
npp.std <- 1/(1 + hwe.bf(dataL2))
plot(f, npp.exact, type="l", lwd = 2, xlab = "f = t/n",
	ylab = "Null posterior probability")
abline(h = npp.std, col = gray(.5), lty = "longdash")

# Dataset 3 #
plot(dataL3)
npp.exact <- 1/(1 + hwe.ibf.exact(round(f*n), y = dataL3))
npp.std <- 1/(1 + hwe.bf(dataL3))
plot(f, npp.exact, type="l", lwd = 2, xlab = "f = t/n",
	ylab = "Null posterior probability")
abline(h = npp.std, col = gray(.5), lty = "longdash")

# Dataset 4 #
plot(dataL4)
npp.exact <- 1/(1 + hwe.ibf.exact(round(f*n), y = dataL4))
npp.std <- 1/(1 + hwe.bf(dataL4))
plot(f, npp.exact, type="l", lwd = 2, xlab = "f = t/n",
	ylab = "Null posterior probability")
abline(h = npp.std, col = gray(.5), lty = "longdash")

## End(Not run)