Predicting lean growth while accounting for correlated traits.

Abstract

Lean tissue growth rate is usually estimated from indirect measurements including growth rate. A procedure to determine prediction equations for lean tissue growth rate is proposed. The procedure restricts the regression of fat growth rate on predicted lean growth rate to be equal to the regression of fat growth rate on actual lean growth rate. The restriction can be phenotypic or genetic if suitable parameter estimates are available. When applied phenotypically, selection on predicted lean tissue growth rate will result in selection differentials for both fat and lean tissue growth rates that are proportional to those obtained by direct selection for lean tissue growth rate. This restriction is desirable because expected correlated changes in fat are used to justify selection for lean tissue growth. Conventional prediction procedures have ignored correlated changes and obscured the original intent of using lean tissue growth rate as a biological selection criterion. When using conventional procedures to predict a biological selection criterion from indirect measurements, changes in important correlated traits may depend more on the choice of indirect measurements than on the choice of selection criterion.