ADVANCED SENSING SYSTEMS FOR APHANOMYCES ROOT ROT DISEASE MONITORING IN PEA

Milton Orlando Valencia Ortiz

doi:10.7273/000007492

Plant biomarkers and spectral reflectance responses resulting from plant-pathogen interactions can be informative about significant changes during the early disease infection process. This dissertation utilized field asymmetric ion mobility spectrometry (FAIMS) and hyperspectral imaging (HSI) systems to monitor Aphanomyces root rot (ARR) disease in pea (Pisum sativum L.). ARR is caused by Aphanomyces euteiches Drechs and is one of the most devastating diseases that affects peas worldwide. To ensure consistent data acquisition with the FAIMS system, a semiautomated volatile organic compounds (VOC) sampling system was developed. In addition, inoculation time and inoculum concentration were optimized using two plant ages (5 and 7 days after emergence) and three inoculum concentrations (1 × 105, 1 × 106, and 2.79 × 106 zoospores ml−1) with a near-isogenic line (NIL). Four NILs were then utilized to monitor the variations in VOC at the early inoculation phase. Two NILs contained the major quantitative trait locus (QTL) Ae-Ps7.6, associated with partial resistance to ARR, and two NILs without QTL. The same NILs were also utilized to detect changes in plant spectral reflectance at the early inoculation phase using the HSI system. A concentration of 1 × 105 zoospores ml−1 was used to inoculate the plants seven days after emergence. Plants were grown in controlled, hydroponic conditions. A split-plot design with two treatments (non-inoculated and inoculated) and six replications was used to assess the disease's effect on VOC and hyperspectral reflectance traits. In addition, unsupervised (PCA, k-means clustering, and agglomerative hierarchical clustering) and supervised (random forest and gradient boosting classifiers) machine-learning methods were employed to uncover patterns associated with the evaluated conditions. The results demonstrated early ARR detection with a consistent VOC biomarker in the FAIMS spectra at 2 and 4 days after inoculation (DAI). This biomarker also showed significant differences between lines with QTL Ae-Ps7.6 and lines without QTL, mostly in non-inoculated treatment. Using HSI analysis, the red-edge wavelength (745 nm) was found as a key feature in detecting ARR at 3 DAI. This dissertation demonstrated the application of FAIMS and HSI technologies for ARR monitoring before visual symptoms appear. These technologies enable breeders to monitor ARR disease at early stages, facilitating the development of partially resistant varieties.

ADVANCED SENSING SYSTEMS FOR APHANOMYCES ROOT ROT DISEASE MONITORING IN PEA

Files and links (1)

Abstract

Metrics

Details