Assessing the compatibility of genotypes in the rice—Magnaporthe oryzae interaction by hyperspectral imaging

Abstract

The optical properties of plants are valuable indicators of their health and can be used to detect, monitor, and characterize both biotic and abiotic stresses. Spectral techniques like hyperspectral imaging (HSI), have gained attention in plant phenotyping due to their ability to capture distinct spectral signatures associated with disease symptoms and nutrient stress. Understanding the complex interactions between host plants, pathogen and plant nutritional status is pivotal for effective disease assessment and management. Here the feasibility of HSI in the visible/near-infrared range in assessing the compatibility of genotypes in the rice (<em>Oryza sativa</em>)—<em>Magnaporthe oryzae</em> pathosystem has been investigated. The potential of HSI as a tool for assessing the sporulation of <em>M. oryzae</em> isolates on resistant and susceptible rice genotypes was examined as well as the effect of varying rates of mineral nitrogen (N) supply on the complex host-pathogen interactions. Hyperspectral data were analyzed by using the spectral angle mapper (SAM) algorithm for supervised classification. Spectral signatures differed between healthy and diseased tissue of rice genotypes varying in susceptibility to <em>M. oryzae</em>. Gene-for-gene-specific interactions between rice and <em>M. oryzae</em> resulted in diverse blast symptom types, enabling the grading of host-pathogen interactions. Time-series imaging of symptoms revealed significant differences in the manifestation and progression of blast symptom subareas. Distinct spectral signatures of these symptom subareas enabled their differentiation and classification by SAM algorithm with higher accuracy than visual assessments. The influence of mineral N supply on chlorophyll content, blast severity, lesion size, and number of lesions depended on the genotypes of rice. Analysis of reflectance spectra of healthy leaves and disease symptoms spectra revealed significant effects of mineral N supply, particularly in the visible and red-edge range of spectra, and interactions with rice genotypes and blast symptom subareas. The red edge inflection point was linked to the rate of mineral N supply. Among rice genotypes infected with <em>M. oryzae</em>, blast symptoms significantly differed in conidia production. Spectral signatures associated with grey tissue – corresponding to the sporulating area of <em>M. oryzae</em> lesions - differed between rice genotypes and a significant, positive correlation was identified between the area under the difference spectrum, representing spectral changes due to sporulation and the number of conidia per lesion and per lesion area. This study demonstrates the efficiency of HSI in assessing the compatibility of rice genotypes to <em>M. oryzae</em> interactions, thereby promoting the phenotyping process, and supporting plant breeding for disease resistance.