Learn R Programming

BMRV (version 1.32)

blvcm_bin: Bayesian latent variable collapsing model (BLVCM) for binary data with probit link

Description

The function implements BLVCM for binary traits using a Gibbs sampler with probit link function.

Usage

blvcm_bin(pheno, geno, model = 3, iter = 30000, burnin = 500, var = -1, lambda = 0.2, cov = 0, init = c(0, 0))

Arguments

pheno
An $N$ x $3$ phenotypic data matrix (trait, family number, zyg=1 for MZ, 2 for DZ), where $N$ is the number of subjects. The trait must be 0 or 1.
geno
An $N$ x $K$ genotypic data matrix, where $N$ is the number of subjects and $K$ is the number of rare variants. The value can be 0 or 1. A missing genotype is represented by -9, which will be imputated by BLVCM based on HWE.
model
Twin model: 3 for ACE model, 2 for AE model, 1 for independent subjects
iter
The number of MCMC iterations (must be positive). The default value is 30000.
burnin
The number of burn-ins (must be positive). The default value is 500.
var
The variance hyperparameters (must be positive) in the priors for $\beta$ and $\gamma$. The default value is 1.
lambda
The threshold $\lambda$ (must be positive) for hypothesis test. The default value is 0.2.
cov
A matrix of other covariates to be adjusted.
init
Initial values for $\beta$ and $\gamma$. The default values are 0. The initial value for $\beta$ must be non-negative.

Value

Details

The Gibbs sampler uses the variable augmentation method for probit link described in Albert, J. H., & Chib, S. (1993). Since the variance of a binary variable is determined by its mean compared to quantitative traits, $\theta$(s) are eliminated to avoid overfitting.

References

He, L., Sillanp, M. J., Ripatti, S., & Pitkniemi, J. (2014). Bayesian Latent Variable Collapsing Model for Detecting Rare Variant Interaction Effect in Twin Study. Genetic epidemiology, 38(4), 310-324.

Albert, J. H., & Chib, S. (1993). Bayesian analysis of binary and polychotomous response data. Journal of the American statistical Association, 88(422), 669-679.

Examples

Run this code
data(blvcm_bin_data)
blvcm_bin(blvcm_bin_data$pheno, blvcm_bin_data$geno[,1:3], iter=5000, burnin=500, model=2)

Run the code above in your browser using DataLab