Meta-analysis for estimates of genetic correlation between birth and weaning weights of cattle

Abstract

The problem of combining information from independent studies permeates almost all fields of science. Because the number of scientific articles being published has increasingly grown in past years, combining information has became a necessity. A meta-analysis was conducted to summarize published estimates of genetic correlation between birth and weaning weights of zebu beef cattle, through the collection of estimates of 39 reports, from 1968 to 2000. Data were analyzed by fixed and random effect models. Components of variance were obtained by the restricted maximum likelihood technique. Strong lack of homogeneity among the studies was observed. As a consequence, fixed model estimates of combined results were inappropriate. Results suggest that random effect models produce better estimates. Subgroup comparisons, for decade publication, showed heterogeneity. Meta-analysis techniques were recommended for quantitative reviews of genetic parameters.