Machine learning-based prediction of dynamic heterosis for plant height with pathway biomarkers in rice

The metabolomic landscape of dynamic heterosis for plant height was displayed in rice. Heterosis-associated pathways across most developmental stages were developed into robust pathway biomarkers. The development of robust biomarkers enables accurate prediction of complex phenotypes, which contribute to the advancement of precision breeding. However, the dynamic nature of biomarkers is often underestimated, as their quantitative changes during development are directly connected to phenotypic transformations, influencing crop agronomic traits. Here, we performed comparative metabolomic analyses to investigate dynamic heterosis for plant height in rice, which is an agronomic trait varying throughout development and is a key determinant of yield heterosis. We found that the levels of pyruvaldehyde were predictive of height heterosis specifically at the seedling stage, while 4-hydroxycinnamic acid positively correlated with height heterosis across four developmental stages. We identified metabolic pathways associated with height heterosis and found that metabolomic changes during the elongation stage had a greater impact than those in other stages. Finally, 11 heterosis-associated pathways were developed into pathway biomarkers using random forest analysis, enabling the prediction of height heterosis in an independent population under different growth conditions. We elucidate the metabolomic landscape of dynamic height heterosis in rice through the identification of heterosis-associated analytes and pathways across stages. Our findings provide a strategy to develop robust biomarkers for heterosis of important agronomic traits by integrating metabolic pathways involved in heterosis across most developmental stages, contributing to the establishment of precision pairing in hybrid crop breeding.