High-throughput phenotyping for everyone: A low-cost, all-in-one plant growth phenotyping system

Abstract

Scientists rely heavily on quantitative measurements to develop and test scientific hypotheses. However, certain processes of biology can be truly complex, making it challenging to quantify. Plant phenotyping is the science of characterizing plants' physiological, anatomical, or biochemical properties (Walter et al. 2015). While some phenotypes arise purely from the genetic makeup of the plant, others may develop as a result of an organism's interactions with its environment, which is determined by various genetic and environmental factors (Pieruschka and Schurr 2019). Therefore, certain plant phenotypes, like plant growth, are dynamic by nature and constantly build on feedback mechanisms between the environment and the organisms' genetic makeup. The “living” nature of these plant phenotypes, dependent on time and space, makes it particularly difficult for quantification. Despite the challenges, plant phenotyping has been crucial to scientists since the beginning of plant science. What started as a simple visual observation by farmers, often referred to as the breeder's eye, has now evolved into high-throughput phenotyping, where plant physiological status and growth are analyzed in a multidimensional manner (Xiao et al. 2022). High-throughput phenotyping techniques focusing on plant growth mostly use continuous, usually short-interval, measurements on plants to monitor dynamic changes in plants over time. As a result, high-throughput phenotyping thrives on expensive equipment such as cameras, drones, and infrared spectroscopy techniques (Fahlgren et al. 2015). Furthermore, data collected from high-throughput phenotyping are usually analyzed with complex computer algorithms and require advanced knowledge programming, standing as another challenge in front of accessibility of these techniques to a broader group of plant scientists. In this issue of Plant Physiology, Yu et al. (2024) report the development of a new high-throughput phenotyping system that is particularly optimized for drought sensitivity screening. This new phenotyping system's low-cost design and open-source nature are compelling to users looking for alternatives to expensive, high-throughput phenotyping platforms. This system consists of 3 low-cost hardware setups, namely, PhenoRig, PhenoCage, and AWWESmo, a computational pipeline for image analysis. PhenoRig system consists of a wooden frame to hold pots and 2 Raspberry Pi cameras and computers that collect plant images every 30 min (Fig. 1). PhenoCage is another low-cost framing system with a rotating platform within which the plants are placed. While the PhenoRig system was used to collect top-view images of plants, PhenoCage is built and used for side-view images, ensuring adequate information about the 3D architecture of the plant. The third component of this system, AWWESmo, was developed to monitor plant evapotranspiration in a more automatized way. Using an Arduino-Watering and Weighing unit, pots are automatically weighted and watered to their target weight. The authors also developed computational tools, RasPiPheno Pipe, and shiny app, RaspiPheno App, for performing downstream data analysis and statistics. Data collected from PhenoRig and PhenCage are then analyzed to generate a digital plant biomass using RasPiPheno Pipe, which benefits from a previous platform, PlantCV (Gehan et al. 2017). Differences between different genotypes were analyzed using the RaspiPheno App throughout time and treatments. This app uses R studio to analyze the digital biomass data statistically, as well as leaf area using t test, Wilcox, ANOVA, or 2-way ANOVA, and can generate graphs using ggplot2 and ggpubr packages. A summary for the new high-throughput, low-cost phenotyping system developed by Yu et al. (2024). The phenotyping system relies on 2 house-built, low-cost pieces of equipment: PhenoCage and PhenoRig. The PhenoRig system is used to collect top-view images of plants, and the PhenoCage is used to collect side-view images. Images are collected every 30 min and then analyzed using the RaspiPheno pipeline to extract digital biomass information. The RaspiPheno App can take the output of the RaspiPheno pipeline and analyze the data statistically using and create images R software (adapted from Yu et al. 2024). To show the potential of their new system, the authors performed high-throughput phenotyping using 3 different plant systems, model plant Arabidopsis thaliana, cowpea, and tepary beans, under drought stress. The authors also conducted a genome-wide association study on a natural diversity panel of cowpea consisting of 368 genotypes. Using this new phenotyping platform, they identified multiple new drought-responsive loci in cowpeas. These loci were screened for annotated genes within the linkage disequilibrium (30 kbp) of the identified SNP, and their homologs were further studied in model plant A. thaliana. To test the functional role of the newly identified drought-responsive genes, the authors took advantage of the Arabidopsis T-DNA insertional mutagenesis collection. They grew 13 Arabidopsis mutant lines under control and drought conditions. They discovered that mutations in 1,8-cineole synthase (AtTPS27, EVT2-2), CAAX amino terminal protease (EVT8), Alpha carbonic anhydrase 7 (AtACA7, EVT3-1, EVT3-2) showed significantly higher rosette size under drought conditions, suggesting that they play a role in plant performance under drought conditions. Increasing the accessibility of high-throughput plant phenotyping is important, considering the challenges agriculture faces due to changing climate. Yu et al. (2024) offer an exciting alternative to the costly, high-throughput phenotyping techniques. Equipment needed for building the system is low cost, and the authors provided detailed instructions (in manuscript and as videos) to build the system in-house. Data analysis followed by phenotyping is made simple with open source data analysis pipeline, RasPiPheno Pipe, and the shiny app, RasPiPheno App. Phenotyping software of this kind will provide an opportunity for plant scientists hoping to perform high-throughput phenotyping but who do not have access to expensive resources. B.A. wishes to thank Plant Physiology for the opportunity to act as an Assistant Features Editor. No new data were generated or analyzed in support of this research.

Plant phenotyping relevance

低コストの画像取得・蒸散計測装置と解析パイプライン／アプリから成る植物表現型解析プラットフォームの開発を中心に扱っているため。

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.