Estimation of changes in genetic parameters in selected lines of mice using REML with an animal model. 1. Lean mass

Abstract

Analysis was undertaken using Restricted Maximum Likelihood (REML) with an animal model of the results of selection for 20 generations for predicted lean mass in 10-week-old male mice. There were three replicates, each comprising high, low and unselected control lines. The overall estimates of heritability (h2) and common environmental correlations (c2) from results of the first seven generations were 0.51 0.03 and 0.21 0.01, respectively. Analyses of data from different lines and different numbers of generations were undertaken but with all pedigrees and data included, which enabled inferences to be drawn on changes in variance that were not due simply to inbreeding or short-term effects of selection. Estimates of h2 were lower in selected lines than the control, increasingly so in later generations, indicating departure from the infinitesimal model assumption of unlinked additive genes each of very small effect. In addition, values of c2became higher in high than in control or low selected lines.

Estimation of changes in genetic parameters in selected lines of mice using REML with an animal model. 2. Body weight, body composition and litter size

“…Random effects were additive genetic (variance VA), litter or common environmental (Va) and residual error (VE), with h2 = VAJVJ,, c2 V/V and V being the phenotypic variance. Different variances were fitted in individual lines or over differing periods of the experiment, using a heterogeneity of variance facility with REML (Visscher & Thompson, 1990;Beniwal et a!., 1992). Univariate analyses were undertaken of body weight data up to generation 38 and bivariate analyses of body weight and gonadal fat pad weight data up to generation 20.…”

“…There is a change in the genetic variance and heritability of body weight as a result of selection, just as for lean mass with which it is highly correlated (Beniwal et a!., 1992), even though inferences are made to the base population. In the model used to analyse P6 data after generation 20, in which records of both males and females are included, it is assumed that there is no sex X genotype interaction.…”

“…1; it can be seen that substantial responses occurred, but only until about generation 33 in the high line. Responses for lean mass (to generation 20) are detailed by Beniwal et al (1992). A summary of means at generation 20 is given in Table 1, and other details, including chemical body analysis, by Hastings & Hill (1989).…”

“…Establishment of the P-lines and results of the first 20 generations of selection for predicted lean mass (LM) where LM=BW-8 )< FW, and BW is body weight (g) of males at 10 weeks and F W the weight (g) of their gonadal fat pads, have been described previously Hastings & Hill, 1989;Beniwal et a!., 1992). There were three replicates each with high, low and unselected control lines with selection within families on LM with 16 full sib families per line for seven generations, eight families subsequently.…”

“…In a previous paper (Beniwal et a!., 1992) a univariate analysis was carried out on the lean mass index for the 20 generations of selection conducted for that trait; here the aim was to investigate relationships among the components of the index.…”