Evaluation of Effective Class-Balancing Techniques for CNN-Based Assessment of Aphanomyces Root Rot Resistance in Pea (Pisum sativum L.).

Aphanomyces root rot (ARR) is a devastating disease that affects the production of pea. The plants are prone to infection at any growth stage, and there are no chemical or cultural controls. Thus, the development of resistant pea cultivars is important. Phenomics technologies to support the selection of resistant cultivars through phenotyping can be valuable. One such approach is to couple imaging technologies with deep learning algorithms that are considered efficient for the assessment of disease resistance across a large number of plant genotypes. In this study, the resistance to ARR was evaluated through a CNN-based assessment of pea root images. The proposed model, DeepARRNet, was designed to classify the pea root images into three classes based on ARR severity scores, namely, resistant, intermediate, and susceptible classes. The dataset consisted of 1581 pea root images with a skewed distribution. Hence, three effective data-balancing techniques were identified to solve the prevalent problem of unbalanced datasets. Random oversampling with image transformations, generative adversarial network (GAN)-based image synthesis, and loss function with class-weighted ratio were implemented during the training process. The result indicated that the classification F1-score was 0.92 ± 0.03 when GAN-synthesized images were added, 0.91 ± 0.04 for random resampling, and 0.88 ± 0.05 when class-weighted loss function was implemented, which was higher than when an unbalanced dataset without these techniques were used (0.83 ± 0.03). The systematic approaches evaluated in this study can be applied to other image-based phenotyping datasets, which can aid the development of deep-learning models with improved performance.

Integrating Phenotypic and Gene Expression Linkage Mapping to Dissect Rust Resistance in Chickling Pea.

Rusts are among the most important foliar biotrophic fungal diseases in legumes. Lathyrus cicera crop can be severely damaged by Uromyces pisi, to which partial resistance has been identified. Nevertheless, the underlying genetic basis and molecular mechanisms of this resistance are poorly understood in L. cicera. To prioritise the causative variants controlling partial resistance to rust in L. cicera, a recombinant inbred line (RIL) population, segregating for response to this pathogen, was used to combine the detection of related phenotypic- and expression-quantitative trait loci (pQTLs and eQTLs, respectively). RILs' U. pisi disease severity (DS) was recorded in three independent screenings at seedling (growth chamber) and in one season of exploratory screening at adult plant stage (semi-controlled field conditions). A continuous DS range was observed in both conditions and used for pQTL mapping. Different pQTLs were identified under the growth chamber and semi-controlled field conditions, indicating a distinct genetic basis depending on the plant developmental stage and/or the environment. Additionally, the expression of nine genes related to U. pisi resistance in L. cicera was quantified for each RIL individual and used for eQTL mapping. One cis-eQTL and one trans-eQTL were identified controlling the expression variation of one gene related to rust resistance - a member of glycosyl hydrolase family 17. Integrating phenotyping, gene expression and linkage mapping allowed prioritising four candidate genes relevant for disease-resistance precision breeding involved in adaptation to biotic stress, cellular, and organelle homeostasis, and proteins directly involved in plant defence.