MULTIPLE-TRAIT GIBBS SAMPLER FOR ANIMAL-MODELS - FLEXIBLE PROGRAMS FOR BAYESIAN AND LIKELIHOOD-BASED (CO)VARIANCE COMPONENT INFERENCE

A set of FORTRAN programs to implement a multiple-trait Gibbs sampling algorithm for (co)variance component inference in animal models (MTGS AM) was developed. The MTGSAM programs are available to the public. Th e programs support, models with correlated genetic effects and arbitra ry numbers of covariates, fixed effects, and independent random effect s for each trait. Any combination of missing traits is allowed. The pr ograms were used to estimate variance components for 50 replicates of simulated data. Each replicate consisted of 50 animals of each sex in each of four generations, for 400 animals in each replicate for two tr aits. For MTGSAM, informative prior distributions for variance compone nts were inverted Wishart random variables with 10 df and means equal to the simulation parameters. A total of 15,000 Gibbs sampling rounds were completed for each replicate, with 2,000 rounds discarded for bur n-in. For multiple-trait derivative free restricted maximum likelihood (MTDFREML), starting values for the variance components were the simu lation parameters. Averages of posterior mean of variance components e stimated using MTGSAM with informative and flat prior distributions fo r variance components and REML estimates obtained using MTDFREML indic ated that all three methods were empirically unbiased. Correlations be tween estimates from MTGSAM using flat priors and MTDFREML all exceede d .99.