Ecological Archives A015-009-A2

Toshinori Okuyama and Benjamin M. Bolker. 2005. Combining genetic and ecological data to estimate sea turtle origins. Ecological Applications 15:315–325.

Appendix B. BUGS (Bayesian inference Using Gibbs Sampling) code (WinBUGS 1.4 version).

# BUGS code for mixed stock analysis with a single gyre


   for(i in 1:N){
    y[i,1:H] ~ dmulti(q[Z[i],],1);
    Z[i] ~ dcat(theta[1:R]);

  for(j in 1:R){x[j,1:H] ~ dmulti(q[j,1:H],T[j])}
  for(j in 1:R){T[j] <- sum(y[j,])}
  for(j in 1:R){q[j,1:H] ~ ddirch(beta[j,])
  for(j in 1:R){
    for(h in 1:H){beta[j,h] <- pb[h]}
   for(j in 1:(R-1)){theta[j] <- xi[j]/(sum(xi[])+1)}
   theta[R] <- 1-sum(theta[1:(R-1)])
   for(j in 1:(R-1)){ xi[j] <- exp(xi0[j])}
   for(j in 1:(R-1)){xi0[j] ~  dnorm(mu[j],sigma2)}
   for(j in 1:(R-1)){ mu[j] <- a0+a1*alrn[j]}
   a0 ~ dnorm(0,1)
   a1 ~ dnorm(0,1)I(0,)
   sigma2 ~ dgamma(1,1)


# Most notation follows Appendix A:
#      N: number of individuals sampled from the mixed stock
#      R: number of rookeries/source populations
#      H: number of haplotypes
#      y: mixed population haplotype data (N by H array)
#      Z: imputed origin
#      q: rookery haplotype frequencies (R by H array)
#  theta: rookery contributions
#      x: rookery haplotype data (R by H array)
#   beta: parameters for rookery haplotype frequencies 
#     mu: expected mean contribution (alr)
#     a0: intercept of alr-transformed size and contribution relationship
#     a1: slope of alr-transformed size and contribution relationship
# sigma2: variance of alr-transformed size and contribution relationship
#   alrn: alr-transformed rookery sizes
#     pb: prior parameters for haplotype frequencies

[Back to A015-009]