Impact of pedigree depth and inclusion of historical data in the estimation of variance components and breeding values in Macrobrachium rosenbergii

Data from a Macrobrachium rosenbergii selection program were analyzed to investigate how different pedigree depths and the volume of data can affect the estimation of variance components on harvest body weight and the breeding values of candidate parents. Seven years of data and reliable pedigree information back to the founding parents were available. The harvest body weights of 81,418 prawns from 724 families were recorded. The variance components were estimated using an animal model and the restricted maximum likelihood method. The estimated breeding values of the candidate parents were calculated using best linear unbiased prediction. Additive genetic variances, common environmental variances, their standard errors and the estimated breeding values of candidate parents obtained from different pedigree depths and historical data were compared. The volume of data had a much larger impact than pedigree depth on variance component estimation and estimated breeding values of candidate parents. The additive genetic variances showed significant differences among the data sets and no obvious trend was found as more years of data were included in the analysis. The common environmental variances showed a downward trend as more years of data were included. The standard errors for additive genetic variance and common environmental variance decreased substantially when two years of data were included, and errors decreased only slightly as more data were included. The accuracy of estimated breeding values improved as more pedigree information and historical data were included in the analysis for all candidate parents. The effect of different pedigree depths and historical data were more striking for family selection than within-family selection on harvest body weight. The results are discussed in a practical context of genetic evaluation for aquatic animals.