Heterotic groups
Background
Hybrid breeding programs are generally based on combining inbred lines from different heterotic groups. Melchinger and Gumber (1998) defined a heterotic group “as a group of related or unrelated genotypes from the same or different populations, which display similar combining ability and heterotic response when crossed with genotypes from other genetically distinct germplasm groups.” Furthermore, a heterotic pattern refers to a specific pair of two heterotic groups, which express high heterosis and consequently high hybrid performance in their cross. Heterotic groups are very important for hybrid breeding due to several reasons. First, they lead to a lower ratio of specific compared to general combining ability. This increases the efficiency of hybrid prediction based on GCA effects and makes early testing as well as genomic selection more effective. Second, handling of sterility systems like cytoplasmic male sterility (CMS) would be largely facilitated. For instance, CMS-cytoplasm needs to be introgressed only in the female pool while restorers are introgressed solely into the male pool. The same holds true for major genes, where functional markers have been developed largely influencing the trait of interest. In wheat, many of those genes are dominant for the positive allele. Thus, it would be sufficient to introgress these genes only in one gene pool. Finally, the handling of flowering biology problems for hybrid wheat is largely facilitated by the use of two different pools. The major problem of cost-efficient hybrid wheat seed production is the limited amount and spread of pollen. To maximize pollination, seed production is based on tall males, which flower later than females, properties which can be easily fixed by using a two-pool concept. Furthermore, a bottleneck due to selection for increased pollination capability is then only a concern in the male pool.
Approaches for the establishment of heterotic groups
If heterotic groups don't exists naturally, they need to be established by the breeder. Unfortunately, their establishment is challenging, takes many years and, thus, needs a good long-term strategy. For instance, heterotic groups have been successfully established in maize by long-term reciprocal recurrent selection. In wheat, however, hybrid breeding is even more challenging compared to allogamous crops like maize because flowering and floral traits add an additional layer of complexity (see item seed production). Before grouping of germplasm into heterotic pools, knowledge about combining ability, hybrid performance, and also genetic diversity is required and therefore test hybrids have to be produced. The number of possible single-cross hybrids increases quadratically as the number of parental inbred lines increases. Hence, not all possible hybrids can be produced and tested under field conditions but recent advances in genomics facilitate improved hybrid predictions.
Recently, two complementary approaches for the establishment of heterotic groups in wheat have been proposed. The first approach proposed by Zhao et al. (2015) is based on a large dataset generated within the HYWHEAT project. Here, phenotypic and genomic data of 1,604 hybrids and their 15 paternal and 120 maternal lines was available. The hybrids were produced in an incomplete factorial mating design. The three-step approach to identify a promising heterotic pattern proposed by Zhao et al. starts by predicting the non-tested hybrids using genomic data. The prediction scenario corresponds to the T2 scenario and yielded a high prediction accuracy of 0.89 based on G-BLUP models. By the way, metabolite profiling failed to increase the genome-based prediction accuracies while tremendously increaising the work load.
When all data on hybrid performance is available, robust and efficient algorithms are required for potential groupings and the identification of promising heterotic patterns. Therefore, we developed a simulated annealing algorithm and employed it to search for a high yielding heterotic pattern within the potential groupings in our large data set. This procedure enabled the identification of a high-yielding heterotic pattern which greatly outperformed the average performance of all possible hybrids. In a last step, the long-term success of the identified heterotic groups was evaluated. This aims to maximize the short-term selection gains without promptly diminishing genetic variance. This was based on the usefulness criterion and additionally taking the theoretical selection limit into account. About 16 individuals per heterotic group seem to be a good compromise between the long- and short-term response to selection. Although this group size appears low it is comparable to the number of ancestors for the Iowa Stiff Stalk Synthetic and Iodent heterotic groups used in the US maize breeding. The proposed three-step strategy can be seen as a starting point for the development of heterotic groups in wheat and other autogamous crops.
The second approach “HyBFrame” proposes a unified framework illustrating how global genetic diversity can be used to complement reciprocal recurrent selection and accelerate the development of genetically distinct heterotic groups. This approach suggests a long-term strategy for hybrid breeding in wheat and shows how initial pools can be further developed with special consideration of genetic distance. Genetic distance itself is not sufficient but required for high hybrid performance and should be taken into account for a successful establishment of heterotic groups. In order to establish genetic distance between heterotic groups, knowledge about genetic diversity is required. We assessed genetic diversity and phenotypic performance in a global collection of 1,110 winter wheat genotypes released during the past decades in 35 countries. Our analyses revealed no major population structure within the Western European material. This phenomenon can be explained by the constant germplasm exchange between the different breeding programs and underscores that no ad hoc heterotic grouping is possible. However, we identified genetically distinct subgroups with potential for hybrid wheat breeding. For instance, Eastern European varieties are highly related with US lines and both groups are separated from the Western European material. These subgroups might be interesting for hybrid wheat breeding in Germany.
Briefly, the HyBFrame approach suggests the following strategy: Adapted germplasm serves as starting material due to the high correlations between line per se performance and hybrid performance. Initial pools are formed based on combining ability and taking floral characteristics into account. Then, the combining ability of the lines in both pools is improved by reciprocal recurrent selection, which over time will result in a divergence of the two groups. At some point, new diversity needs to be introgressed in both pools. Here, it is crucial that this process does not disrupt the established heterotic pattern and the genetic distance between the both groups. Therefore, the lines should be molecularly characterized to assess their genetic distance to the two groups and to allow the genomics-based prediction of their GCA or the hybrid performance. Some candidates may lack in adaptation and must be first crossed with adapted elite lines before entering one of the pools. The whole process can be supported by using the already identified genetically distinct subgroups.
HyBFrame approach (modified)