abf.Wakefield: Calculate approximate Bayes factor (ABF) using method of...
In tobyjohnson/gtx: Genetics ToolboX

Description Usage Arguments Details Value Author(s) Examples

Calculates an approximation to the Bayes factor for an alternative model where the parameter beta is a priori normal, against a smaller model where beta is zero, by approximating the likelihood function with a normal distribution.

1	abf.Wakefield(beta, se, priorsd, log = FALSE)

`beta`	Vector of effect size estimates.
`se`	Vector of associated standard errors.
`priorsd`	Scalar specifying the standard deviation of the prior on true effect sizes.
`log`	Whether to return results on a natural log scale.

See “Bayes factors for genome-wide association studies: comparison with P-values” by John Wakefield, 2009, Genetic Epidemiology 33(1):79-86 at http://dx.doi.org/10.1002/gepi.20359.

The original definition was for the Bayes factor for the model where beta is zero (no association), relative to the model with a normal prior for beta (association). The definition used for this function inverts the original definition, so that higher values of the Bayes factor (or log Bayes factor) indicate stronger support for the model with association.

For strong associations, calculating Bayes factors may be numerically unstable, and it is recommended to work on a log scale and rescale appropriately before attempting to calculate Bayes factors or posterior probabilities.

A vector of approximate Bayes factors, on a log scale if log=TRUE. Higher values indicate stronger support for association (which is inverted relative to the original definition).

Toby Johnson Toby.x.Johnson@gsk.com

data(agtstats)
agtstats$pval <- with(agtstats, pchisq((beta/se.GC)^2, df = 1, lower.tail = FALSE))
max1 <- function(bf) return(bf/max(bf, na.rm = TRUE))
agtstats$BF.normal <- with(agtstats, max1(abf.Wakefield(beta, se.GC, 0.05)))
agtstats$BF.t <- with(agtstats, max1(abf.t(beta, se.GC, 0.0208)))
with(agtstats, plot(-log10(pval), log(BF.normal)))
with(agtstats, plot(-log10(pval), log(BF.t)))