Restricted maximum likelihood procedures for the estimation of additive and nonadditive genetic variances and covariances in multibreed populations.

摘要: Restricted maximum-likelihood procedures were developed to estimate additive and nonadditive genetic environmental covariances for multiple traits in multibreed populations. The computational procedure follows the expectation-maximization (EM) algorithm, where set of equations maximization step is solved by successive approximations. This does not guarantee convergence a symmetric positive-definite covariance matrix. Thus, computer programs will need incorporate restrictions ensure positive definiteness each Additive modeled subclass form (zeros ones design matrices). Nonadditive regression (any value between including zero one Computational requirements be larger than intrabreed analyses. Appropriate simplifying assumptions numerical techniques (e.g., sparse iterative techniques) required implementation these estimation procedures. Number iterations (5 12) computing times (57 113 min) achieve when estimating 21 five small simulated data sets (two breeds, 25,200 50,400 calves, 120 135 unrelated bulls) suggest that are computationally feasible.