Submitted to: Journal of Animal Science Supplement
Publication Type: Abstract Only
Publication Acceptance Date: June 10, 1996
Publication Date: N/A
The Multiple Trait Gibbs Sampler for Animal Models (MTGSAM) programs were extended to allow analysis of ordered categorical data using a Bayesian threshold model. The algorithm is based on data augmentation, where an observation on the unobserved underlying normally distributed variable is generated in each round of iteration for each categorical observation. The programs allow for a flexible number of continuous and categorical variables. Categorical levels must be ordered but are not required to be sequential or starting at one. The program allows any number of fixed and random effects for categorical or continuous traits. The model can be different for each trait. The programs were used to analyze twinning and ovulation rates from cattle selected for twinning rate at the U.S. Meat Animal Research Center. Data included number of calves born in each parturition for the lifetime of a cow and number of eggs ovulated for several estrous cycles before first breeding as heifers. Observations of triplets born (n=25) or three eggs ovulated (n=29) were included as counts of two. A total of 6411 calvings was recorded with 16.8% multiple births. A total of 19,849 ovulations was recorded with 14.8% multiple ovulations. The model for twinning rate included fixed effects for combinations of year, season, and age and random permanent environmental (PE) effect. The model for ovulation rate included fixed effects of birth year-season, age group, and calendar month and random PE effects. Gibbs sampler was run for 11500 rounds with the first 1500 rounds discarded as burn-in. Mean posterior estimates of heritability and fraction of variance accounted for by PE were .17 and .07 for twinning and .19 and .06 for ovulation. Mean posterior estimate of genetic correlation was .77 and correlation of PE was .43.