bayesIndirect: bayesIndirect

In SharonLutz/bayesIndirect: Runs the bayesIndirect Function

Description

bayesIndirect examines the indirect effect of a SNP on the outcome through the mediator in a Bayesian framework with a spike and slab prior.

Usage

bayesIndirect(x, k, y, z = NULL, nCov = 0, plot0 = FALSE, nIts = 50000, nBurn = 10000, propGamma1Pi = 0.7, cx = 10, pix = 0.5, SEED = 1)

Arguments

x	is the exposure, such as the SNP.
k	is the mediator.
y	is the normally distributed outcome.
z	is the matrix of covariates.
nCov	is the number of covariates. nCov = 0, 1 or 2.
plot0	true if density plot should be outputted or false otherwise.
nIts	is the number of iterations of the MCMC.
nBurn	is the Burn-in period.
propGamma1Pi	is the probability of gamma from the Bernoulli distribution.
cx	is the variance of beta X from the normal distribution.
pix	is the probability that beta x is sampled from the normal distribution.
SEED	is the seed for the SKAT function, default = 1.

Value

The posterior mean, standard deviation and quantiles for each variable.

Warning

library(coda) is needed to run this function.

Author(s)

Sharon Lutz, Annie Thwing

References

Lutz SM, Hokanson JE, Sharma S, Weiss S, Raby B, Lange C. (2013) On the Integration of Expression Profiles in Genetic Association Studies: A Bayesian Approach to Determine the Path from Gene to Disease. Open Journal of Genetics. 3(3).

SharonLutz/bayesIndirect documentation built on May 20, 2018, 11:12 a.m.