| The world is currently facing the daunting task of addressing food and energy shortages, which are results of a growing global population, decreasing available resources, and severe climate changes. Fortunately, exploitation of heterosis in crops has contributed greatly to improvement in gobal food and energy production over the past several decades. However, available elite hybrid crops seem insufficient to meet the food and energy needs in the future. Screening and breeding new excellent hybrid combintions is urgently required to guarantee food security. Quantitative hybridization trials and hybrid performance evaluation works, which are time-consuming and laborious, form the foundation of present hybrid breeding programs. To increase breeding efficiency and reach breeding program goals, breeders tried phenotypes of parental lines and genetic distance based on various types of molecular markers to predict hybrid phenotypes or heterosis. Whereas they found weakness like low predictive ability or limited application ranges of the prediction approaches. To find suitable prediction strategy, rice, which is the sample food for more than half people on earth, was chosen in this study. After constructing a complete diallel cross population with indica and japonica rice inbred lines, genetic divergence index based on InDel markers and parental metabolite profiling were used to predict performance of F1 hybrids. Finally, result of prediction to gain yield per plant with InDel markers demonstrated that, when the genetic divergence index is 0.37, F1 hybrids can avoid reproductive isolation of intersubspecies and utilize intersubspecific heterosis simultaneously. Meanwhile, high predictive abilities (r>0.8) were obtained in prediction of grain yield per plant, plant height and heading date with the combination of parental metabolite profiling and partial least square to latent factor. Notably, both the above two methods can provide reliable predictions for F1 hybrids either in indica-indica,japonica-japonica, indica-japonica population, or different cytoplasm reciprocal hybrids. Our findings offered workable strategies for screening promising superior hybrid rice, which can contribute to rice production improvement and accelerate breeding programs.Despite the pervasive utilization of heterosis and paramount importance, the mechanism of heterosis remains elusive. Classical theories including dominance, overdominance, epistasis etc. can elucidate the mechanism of heterosis for some traits under specific conditions, while they can not comprehensively explain different types of the phenomenon of heterosis. Here, the mechanisms of heterosis for both vegetative and reproductive traits were analyzed at phenotypic and metabolic two levels. The results of modes of inheritance of all investigated traits in the testcross population demonstrated that additive effects were the foundation of heterosis for complex traits in a hierarchical structure, and multiplicative interactions among the component traits were the framework of heterosis in complex traits. Furthermore, new balances between unit traits and related component traits provided hybrids with the opportunity to achieve an optimal degree of heterosis for complex traits. At the metabolic level, different groups of metabolites were identified to be correlated with specific traits. Metabolomic additive effects along with parent-of-origin effects were proved to contribute to better-parent heterosis. Furthermore, we found that component trait-correlated metabolites additively accumulated to achieve better-parent heterosis for component traits, while better-parent heterosis for complex traits were multiplicatively assembled based on component trait-correlated metabolites. Research focus should be paid to component traits, rather than complex traits. And regarding component trait-correlated metabolites as subunits of multi-molecular complexes in dynamic systems to reveal the mechanism of heterosis should be fruitful and have universal significance. |
PDF Full Text Request