SeptoSympto: a precise image analysis of Septoria tritici blotch disease symptoms using deep learning methods on scanned images.

BACKGROUND: Investigations on plant-pathogen interactions require quantitative, accurate, and rapid phenotyping of crop diseases. However, visual assessment of disease symptoms is preferred over available numerical tools due to transferability challenges. These assessments are laborious, time-consuming, require expertise, and are rater dependent. More recently, deep learning has produced interesting results for evaluating plant diseases. Nevertheless, it has yet to be used to quantify the severity of Septoria tritici blotch (STB) caused by Zymoseptoria tritici-a frequently occurring and damaging disease on wheat crops. RESULTS: We developed an image analysis script in Python, called SeptoSympto. This script uses deep learning models based on the U-Net and YOLO architectures to quantify necrosis and pycnidia on detached, flattened and scanned leaves of wheat seedlings. Datasets of different sizes (containing 50, 100, 200, and 300 leaves) were annotated to train Convolutional Neural Networks models. Five different datasets were tested to develop a robust tool for the accurate analysis of STB symptoms and facilitate its transferability. The results show that (i) the amount of annotated data does not influence the performances of models, (ii) the outputs of SeptoSympto are highly correlated with those of the experts, with a similar magnitude to the correlations between experts, and (iii) the accuracy of SeptoSympto allows precise and rapid quantification of necrosis and pycnidia on both durum and bread wheat leaves inoculated with different strains of the pathogen, scanned with different scanners and grown under different conditions. CONCLUSIONS: SeptoSympto takes the same amount of time as a visual assessment to evaluate STB symptoms. However, unlike visual assessments, it allows for data to be stored and evaluated by experts and non-experts in a more accurate and unbiased manner. The methods used in SeptoSympto make it a transferable, highly accurate, computationally inexpensive, easy-to-use, and adaptable tool. This study demonstrates the potential of using deep learning to assess complex plant disease symptoms such as STB.

Quantitative and qualitative plant-pathogen interactions call upon similar pathogenicity genes with a spectrum of effects.

Septoria leaf blotch is a foliar wheat disease controlled by a combination of plant genetic resistances and fungicides use. R-gene-based qualitative resistance durability is limited due to gene-for-gene interactions with fungal avirulence (Avr) genes. Quantitative resistance is considered more durable but the mechanisms involved are not well documented. We hypothesize that genes involved in quantitative and qualitative plant-pathogen interactions are similar. A bi-parental population of Zymoseptoria tritici was inoculated on wheat cultivar 'Renan' and a linkage analysis performed to map QTL. Three pathogenicity QTL, Qzt-I05-1, Qzt-I05-6 and Qzt-I07-13, were mapped on chromosomes 1, 6 and 13 in Z. tritici, and a candidate pathogenicity gene on chromosome 6 was selected based on its effector-like characteristics. The candidate gene was cloned by Agrobacterium tumefaciens-mediated transformation, and a pathology test assessed the effect of the mutant strains on 'Renan'. This gene was demonstrated to be involved in quantitative pathogenicity. By cloning a newly annotated quantitative-effect gene in Z. tritici that is effector-like, we demonstrated that genes underlying pathogenicity QTL can be similar to Avr genes. This opens up the previously probed possibility that 'gene-for-gene' underlies not only qualitative but also quantitative plant-pathogen interactions in this pathosystem.

Resistance of the Wheat Cultivar 'Renan' to Septoria Leaf Blotch Explained by a Combination of Strain Specific and Strain Non-Specific QTL Mapped on an Ultra-Dense Genetic Map.

Quantitative resistance is considered more durable than qualitative resistance as it does not involve major resistance genes that can be easily overcome by pathogen populations, but rather a combination of genes with a lower individual effect. This durability means that quantitative resistance could be an interesting tool for breeding crops that would not systematically require phytosanitary products. Quantitative resistance has yet to reveal all of its intricacies. Here, we delve into the case of the wheat/Septoria tritici blotch (STB) pathosystem. Using a population resulting from a cross between French cultivar Renan, generally resistant to STB, and Chinese Spring, a cultivar susceptible to the disease, we built an ultra-dense genetic map that carries 148,820 single nucleotide polymorphism (SNP) markers. Phenotyping the interaction was done with two different Zymoseptoria tritici strains with contrasted pathogenicities on Renan. A linkage analysis led to the detection of three quantitative trait loci (QTL) related to resistance in Renan. These QTL, on chromosomes 7B, 1D, and 5D, present with an interesting diversity as that on 7B was detected with both fungal strains, while those on 1D and 5D were strain-specific. The resistance on 7B was located in the region of Stb8 and the resistance on 1D colocalized with Stb19. However, the resistance on 5D was new, so further designated Stb20q. Several wall-associated kinases (WAK), nucleotide-binding and leucine-rich repeats (NB-LRR) type, and kinase domain carrying genes were present in the QTL regions, and some of them were expressed during the infection. These results advocate for a role of Stb genes in quantitative resistance and for resistance in the wheat/STB pathosystem being as a whole quantitative and polygenic.