Root architecture simulation improves the inference from seedling root phenotyping towards mature root systems.

Area of Science:

• Plant Science
• Agricultural Science
• Computational Biology

Background:

• High-throughput phenotyping of plant root systems is crucial for breeding but often limited to seedlings.
• Current methods struggle to predict mature root system architecture and traits from early-stage data.
• Bridging this gap is essential for advancing breeding programs that rely on root characteristics.

Purpose of the Study:

• To evaluate the efficacy of root system architecture (RSA) models in predicting mature root traits from seedling data.
• To identify key seedling root traits that are most predictive of mature root system performance.
• To overcome the inference problem in high-throughput seedling root phenotyping for plant breeding.

Main Methods:

• Phenotyped 16 pea genotypes using seedling (Petri dish) and mature (sand-filled column) root platforms.
• Parameterized the RSA model RootBox with seedling traits to simulate mature root systems.
• Compared measured and modeled root traits (length, lateral number, distribution) to assess prediction accuracy.

Main Results:

• No direct correlation was found between seedling and mature root traits across phenotyping systems.
• RootBox modeling significantly improved trait inference compared to direct phenotyping platforms alone.
• Seedling elongation rates and interbranch distance were sufficient to predict genotype ranking for total root length (RSpearman = 0.83).
• Incorporating uneven lateral spacing via a scaling function enhanced prediction of diverse root architectures.

Conclusions:

• Root system architecture (RSA) models effectively solve the inference problem for seedling root phenotyping.
• RSA models provide reliable predictions of mature root systems, valuable for breeding research.
• Integrating RSA models into phenotyping pipelines is recommended for robust breeding data.