Zero-inflated gaussian model in MCMCglmm

zig.R

dat <- data.frame(obs=c(rep(0, times=100), rnorm(100, 10, 1)))

# observed data looks something like
hist(dat$obs)

# break into the zero-part and the normal part of the distribution
# observed zeros get zero for the first part, otherwise 1
# observed zeros get NA for the second part, otherwise observed value
dat <- transform(dat, zero_part=ifelse(obs == 0, yes=0, no=1),
                      norm_part=ifelse(obs != 0, yes=obs, no=NA))

zero_part <- rep(c(0, 1), each=100)
norm_part <- rnorm(100, 10, 1)

library(MCMCglmm)

# 2 latent traits for gaussian + categorical
prior=list(R=list(V=diag(2), nu=1, fix=2))

m <- MCMCglmm(cbind(norm_part, zero_part) ~ trait - 1,
              rcov=~idh(trait):units,
              data=dat,
              family=c('gaussian', 'categorical'),
              prior=prior)