RootXplorer: A computer vision-based 3D phenotyping platform for high-throughput quantification and spatio-temporal analysis of root system penetrability.

Abstract

Studying the mechanisms that promote deep rooting in crops is crucial for engineering plant varieties with enhanced drought resilience and increased carbon sequestration capacity. Soil compaction is a major constraint on rooting depth and, to overcome this, root system penetrability needs to be enhanced. However, because of the limitations of current methods, phenotyping root penetrability remains a bottleneck. Here, we developed RootXplorer, a computer vision-based 3D phenotyping platform for high-throughput quantification of root penetration-related traits/phenotypes across dicot and monocot species. RootXplorer integrates a novel Phytagel-based cylinder system, a 3D imaging unit, and an automated software pipeline to extract root penetration-related traits with high precision and at a large scale. We demonstrate that RootXplorer enables large-scale diversity screenings in conditions replicating soil compaction effects in multiple species, revealing species-specific strategies for overcoming mechanical impedance. These findings highlight the utility and promise of RootXplorer for accelerating research on root architectural plasticity under controlled compaction conditions, identifying genotypes with varying tolerance to mechanical impedance, and supporting data-driven breeding decisions for developing soil compaction-resilient crop varieties. This technology has important implications for future plant breeding strategies and supports ongoing climate change mitigation efforts.

Plant phenotyping relevance

Root penetrability関連形質を対象に、3D画像計測と自動ソフトウェアで抽出する高スループット表現型解析プラットフォームを開発しており、方法が研究の中心です。

Code and data availability

The paper's Data Availability statement explicitly deposits the complete analysis pipeline and time-lapse video generation code in two public GitHub repositories under the authors' Salk Harnessing Plants Initiative organization. These directly support the paper's RootXplorer phenotyping analysis (image cropping, U-Net+