Evaluating different genomic coancestry matrices for managing genetic variability in turbot

Abstract

In population management, the most efficient method to control the increase of inbreeding and the associated loss of genetic variability is the Optimal Contributions method. This method optimizes the contributions of breeding candidates by minimizing the weighted global coancestry. Traditionally, coancestry coefficients have been estimated from pedigree data but the current availability of genome-wide information allows us to estimate them with higher precision. In recent years, developments of genomic tools in aquaculture species have been very significant. For turbot, a species with an increasing aquaculture value, the whole genome has been recently assembled and genetic and physical maps have been refined. Although several measures of genomic coancestry have been proposed, their relative efficiency for maintaining genetic variability is unknown. The objectives of this study were to compare different measures of genomic coancestry for turbot, and to evaluate their efficiency for retaining genetic variability when using the Optimal Contributions method. We used genomic data obtained through 2b-RAD technology for a domesticated population to achieve the objectives. The different genome-wide coancestry matrices compared were based on: i) the proportion of shared alleles; ii) deviations of the observed number of alleles shared by two individuals from the expected number; iii) the realized relationship matrix obtained by VanRaden's method 1; iv) the realized relationship matrix obtained by VanRaden's method 2; v) the realized relationship matrix obtained by Yang's method; and vi) identical by descent segments. The amount of genetic variability retained when using each coancestry matrix was measured as the expected heterozygosity in the next generation. Results revealed that coancestry coefficients showing high correlations between them gave similar results from the optimization. The genetic variability retained was about 5% higher when using the matrices based on the proportion of shared alleles, deviations of the observed number of alleles shared or segments than when using the three genomic relationship matrices. Matrices retaining more variability showed a higher ability to discriminate relationships between individuals. The higher the diversity achieved the lower was the number of fish selected to contribute to the next generation.