Genomic Insights Into Heterosis: Dominance or Additive × Additive Interaction?

Abstract

Heterosis was documented in the 18th century, but its biological basis has been debated since. The theoretical framework proposed by Hill, and adapted by Lynch, is based on two central parameters: admixed composition (S), and the heterozygosity (H). Using genomic information, it is now possible to estimate independently the individual realized Si and Hi. In this research, a methodology for estimating the contribution of dominance and additive &#xd7; additive effects to heterosis is proposed. This approach would be especially relevant in cases where there is insufficient phenotypic information available, or an adequate genetic group experimental design, common in humans, wild species and other admixed populations. We also provide theoretical arguments highlighting the enhanced precision of the estimations of heterosis parameters through this method. Furthermore, we exemplify this procedure by analysing data from an experimental F2 pig population, which was initially designed for QTL mapping. Notably, all animals in this population were genotyped (including F1 and parental breeds), but phenotypic information was only available for F2 individuals and included 13 traits related to growth, fat deposition, carcass characteristics and meat quality. Significant additive effects (p&#x2009;<&#x2009;0.05) were detected for longissimus muscle area and carcass temperature, suggesting complementary additive effects for these traits. Significant dominance and additive &#xd7; additive effects were also detected for birth weight and carcass length, respectively (p&#x2009;<&#x2009;0.05), indicating that heterosis for these traits is primarily attributable to dominance and additive &#xd7; additive interactions. These results demonstrate that the proposed methodology can successfully estimate the genetic components underlying heterosis and underscores the utility of this approach in&#xa0;situations where we possess genomic data but limited phenotypic data.