Developing a statistical genomics toolbox to decipher the genetic architecture of heterosis using whole-genome sequencing data

• Applicant: Dr. Yong Jiang
• Subject Area: Plant Breeding and Plant Pathology ; Bioinformatics and Theoretical Biology ; Plant Genetics and Genomics
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 540803247

Heterosis is the superiority of a hybrid over its parents. Although the phenomenon has been systematically exploited in crop breeding, its genetic basis is still not clearly explained. Three mutually not exclusive hypotheses have been proposed to explain heterosis, namely dominance, overdominance and epistasis. Since there is strong evidence that heterosis is a polygenic phenomenon, its genetic basis has been investigated mainly using quantitative genetics approaches. Most previous studies were based on biparental populations, for which a systems-oriented approach jointly considering dominance and epistasis was established by Melchinger in 2007. Nevertheless, results obtained in bi-parental populations may have limited generality as they may depend on the two founder lines. Inspired by Melchinger’s work, we established a quantitative genetic framework for studying heterosis with diverse hybrid populations in 2017. Since the linkage disequilibrium in diverse populations decays fast, it is necessary to use high-density marker data, such as whole-genome sequencing (WGS), to increase the mapping resolution. Moreover, the use of WGS data could avoid misinterpreting the results due to apparent epistasis, a phenomenon that the epistatic effect between two different loci can be explained by the additive or dominance effect of an unobserved locus. Several studies on the heterosis in rice and maize have exploited WGS data. However, epistatic effects were either excluded from consideration or restricted to the interaction between markers showing significant dominance effects in these studies. Therefore, studies based on a proper approach jointly considering dominance and epistasis, by using a large diverse hybrid population with WGS data, are still lacking. The principle of our established quantitative genetic framework is suitable for this purpose, but the current mathematical model is computationally not feasible to analyze data sets with thousands of individuals and millions of markers. In view of the above, the main goal of this project is to design an efficient statistical genomics toolbox that improves the existing quantitative genetic framework for studying the genetic basis of mid-parent heterosis. The core of the toolbox is a new mathematical model that can directly test the heterotic effect of each marker, instead of first performing a two-dimensional scan of all digenic epistatic effects and then integrating them into heterotic effects. For a data set with 1,000 individuals genotyped by 1,000,000 markers, the new model is expected to be 100 times faster than the current one. Furthermore, the new model has the potential to be generalized as a novel GWAS algorithm that can handle multiple levels of genetic relatedness. As an application, we plan to utilize the new toolbox to dissect and compare the genetic architecture of heterosis in a wheat and a maize data set, each with more than 4,000 hybrids whose parental lines were genotyped by WGS.

• DFG Programme: Research Grants
• Co-Investigator: Professor Dr. Jochen Reif