ESTIMATING UNCERTAINTY IN MULTIVARIATE RESPONSES TO SELECTION

摘要: Predicting the responses to natural selection is one of key goals evolutionary biology. Two challenges in fulfilling this goal have been realization that many estimates might be highly biased by environmentally induced covariances between traits and fitness, estimated do not incorporate or report uncertainty estimates. Here we describe application a framework blends merits Robertson-Price Identity approach multivariate breeder's equation address these challenges. The allows genetic covariance matrices, differentials, gradients, without bias, direct indirect distinguished, if implemented Bayesian-MCMC framework, statistically robust on all parameters made. We illustrate our with worked example previously published data. More generally, suggest applying both will facilitate hypothesis testing about selection, constraints, responses.