Comparison of growth curves of mice selected and unselected for postweaning gain

Mice were sampled from a line selected for increased postweaning weight gain from three to six weeks and from a randombred control line originating from the same base population. Body weights were recorded at each of 14 ages from day 5 to day 98. The Richards and logistic growth functions were fitted to the growth trajectories of each individual mouse by a generalized non-linear least squares procedure. Estimated growth parameters (asymptotic weight, rate, shape of curve, age and weight at inflection, mean absolute growth rate and mean relative growth rate) were computed for each individual. The effects of line, litter within line, sex and line × sex interactions on these estimated parameters were then studied.Both the Richards and logistic functions fitted the data equally well and the plotted trajectories coincided over most of the growth curve. There was excellent agreement between the estimates of asymptotic weight and both age and weight at inflection based on the different functions. However, both functions apparently underestimated the asymptotic weight.Analyses of the line differences showed that selection for postweaning gain increased the mean absolute growth rate over the entire curve but had no effect on the relative growth rate or the shape of the growth curve. Full-sib analyses suggested the presence of considerable genetic variation and some high genetic correlations among the estimated growth parameters.