High-Throughput Phenotyping and Random Regression Models Reveal Temporal Genetic Control of Soybean Biomass...

Area of Science:

• Plant genetics
• Crop science
• Quantitative genetics

Background:

• Understanding temporal accumulation of soybean above-ground biomass (AGB) is key for yield enhancement and developing stress-resilient cultivars.
• High-throughput phenotyping and genomic prediction are essential for dissecting complex traits like AGB.

Purpose of the Study:

• To develop a high-throughput phenotyping method for predicting soybean AGB over time.
• To investigate the temporal quantitative genomic properties of soybean AGB.
• To utilize random regression models (RRM) for analyzing dynamic genetic architecture.

Main Methods:

• Utilized a SoyNAM population (n = 383) in multi-environment trials.
• Collected destructive AGB measurements and multispectral/RGB imaging from 27 to 83 days after planting (DAP).
• Applied machine learning for phenotypic prediction, genomic prediction, and GWAS using RRM.

Main Results:

• Achieved high AGB phenotypic predictions (R² = 0.92-0.94).
• Estimated narrow-sense heritabilities for AGB ranged from 0.02 to 0.28 across different DAP.
• Observed high prediction accuracies and low biases for genomic breeding values, with time-varying significant genomic regions.

Conclusions:

• Random regression models are powerful for modeling the temporal genetic architecture of soybean AGB.
• This study provides a foundation for integrating phenotyping and genomic analyses for complex longitudinal plant traits.
• The findings offer valuable insights for soybean crop improvement strategies targeting biomass accumulation.