Phenomics: conceptualization and importance for plant physiology

CucumberAI: Cucumber Fruit Morphology Identification System Based on Artificial Intelligence

Cucumber is an important vegetable crop that has high nutritional and economic value and is thus favored by consumers worldwide. Exploring an accurate and fast technique for measuring the morphological traits of cucumber fruit could be helpful for improving its breeding efficiency and further refining the development models for pepo fruits. At present, several sets of measurement schemes and standards have been proposed and applied for the characterization of cucumber fruits; however, these manual methods are time-consuming and inefficient. Therefore, in this paper, we propose a cucumber fruit morphological trait identification framework and software called CucumberAI, which combines image processing techniques with deep learning models to efficiently identify up to 51 cucumber features, including 32 newly defined parameters. The proposed tool introduces an algorithm for performing cucumber contour extraction and fruit segmentation based on image processing techniques. The identification framework comprises 6 deep learning models that combine fruit feature recognition rules with MobileNetV2 to construct a decision tree for fruit shape recognition. Additionally, the framework employs U-Net segmentation models for fruit stripe and endocarp segmentation, a MobileNetV2 model for carpel classification, a ResNet50 model for stripe classification and a YOLOv5 model for tumor identification. The relationships between the image-based manual and algorithmic traits are highly correlated, and validation tests were conducted to perform correlation analyses of fruit surface smoothness and roughness, and a fruit appearance cluster analysis was also performed. In brief, CucumberAI offers an efficient approach for extracting and analyzing cucumber phenotypes and provides valuable information for future cucumber genetic improvements.

Plant microphenotype: from innovative imaging to computational analysis

The microphenotype plays a key role in bridging the gap between the genotype and the complex macro phenotype. In this article, we review the advances in data acquisition and the intelligent analysis of plant microphenotyping and present applications of microphenotyping in plant science over the past two decades. We then point out several challenges in this field and suggest that cross-scale image acquisition strategies, powerful artificial intelligence algorithms, advanced genetic analysis, and computational phenotyping need to be established and performed to better understand interactions among genotype, environment, and management. Microphenotyping has entered the era of Microphenotyping 3.0 and will largely advance functional genomics and plant science.

Differentiated responses of the phyllosphere bacterial community of the yellowhorn tree to precipitation and temperature regimes across Northern China

Introduction

As an ephemeral and oligotrophic environment, the phyllosphere harbors many highly diverse microorganisms. Importantly, it is known that their colonization of plant leaf surfaces is considerably influenced by a few abiotic factors related to climatic conditions. Yet how the dynamics of phyllosphere bacterial community assembly are shaped by detailed climatological elements, such as various bioclimatic variables, remains poorly understood.

Methods

Using high-throughput 16S rRNA gene amplicon sequencing technology, we analyzed the bacterial communities inhabiting the leaf surfaces of an oilseed tree, yellowhorn (Xanthoceras sorbifolium), grown at four sites (Yinchuan, Otogqianqi, Tongliao, and Zhangwu) whose climatic status differs in northern China.

Results and Discussion

We found that the yellowhorn phyllosphere's bacterial community was generally dominated by four phyla: Proteobacteria, Firmicutes, Actinobacteria, and Bacteroidetes. Nevertheless, bacterial community composition differed significantly among the four sampled site regions, indicating the possible impact of climatological factors upon the phyllosphere microbiome. Interestingly, we also noted that the α-diversities of phyllosphere microbiota showed strong positive or negative correlation with 13 bioclimatic factors (including 7 precipitation factors and 6 temperature factors). Furthermore, the relative abundances of 55 amplicon sequence variants (ASVs), including three ASVs representing two keystone taxa (the genera Curtobacterium and Streptomyces), exhibited significant yet contrary responses to the precipitation and temperature climatic variables. That pattern was consistent with all ASVs' trends of possessing opposite correlations to those two parameter classes. In addition, the total number of links and nodes, which conveys community network complexity, increased with rising values of most temperature variables. Besides that, remarkably positive relevance was found between average clustering coefficient and most precipitation variables. Altogether, these results suggest the yellowhorn phyllosphere bacterial community is capable of responding to variation in rainfall and temperature regimes in distinctive ways.

Towards the synthesis of spectral imaging and machine learning-based approaches for non-invasive phenotyping of plants

High-throughput phenotyping is now central to the progress of plant sciences, accelerated breeding, and precision farming. The power of phenotyping comes from the automated, rapid, non-invasive collection of large datasets describing plant objects. In this context, the goal of extracting relevant information from different kinds of images is of paramount importance. We review both the spectral and machine learning-based approaches to imaging of plants for the purpose of their phenotyping. The advantages and drawbacks of both approaches will be discussed with a focus on the monitoring of plants. We argue that an emerging approach combining the strengths of the spectral and the machine learning-based approaches will remain a promising direction in plant phenotyping in the nearest future.

Pitfalls and potential of high-throughput plant phenotyping platforms

Automated high-throughput plant phenotyping (HTPP) enables non-invasive, fast and standardized evaluations of a large number of plants for size, development, and certain physiological variables. Many research groups recognize the potential of HTPP and have made significant investments in HTPP infrastructure, or are considering doing so. To make optimal use of limited resources, it is important to plan and use these facilities prudently and to interpret the results carefully. Here we present a number of points that users should consider before purchasing, building or utilizing such equipment. They relate to (1) the financial and time investment for acquisition, operation, and maintenance, (2) the constraints associated with such machines in terms of flexibility and growth conditions, (3) the pros and cons of frequent non-destructive measurements, (4) the level of information provided by proxy traits, and (5) the utilization of calibration curves. Using data from an Arabidopsis experiment, we demonstrate how diurnal changes in leaf angle can impact plant size estimates from top-view cameras, causing deviations of more than 20% over the day. Growth analysis data from another rosette species showed that there was a curvilinear relationship between total and projected leaf area. Neglecting this curvilinearity resulted in linear calibration curves that, although having a high r2 (> 0.92), also exhibited large relative errors. Another important consideration we discussed is the frequency at which calibration curves need to be generated and whether different treatments, seasons, or genotypes require distinct calibration curves. In conclusion, HTPP systems have become a valuable addition to the toolbox of plant biologists, provided that these systems are tailored to the research questions of interest, and users are aware of both the possible pitfalls and potential involved.