Search

Septoria tritici blotch caused by Zymoseptoria tritici is a serious threat to wheat in Tunisia, causing yield losses of up to 50%. Epidemics occur regularly, mainly on durum wheat (DW) but also to a lesser extent on bread wheat (BW), that are mainly cropped in the north‐western and north‐eastern regions, respectively. Isolates of Z. tritici exhibit both cultivar specificity and host species specificity. In this study, we investigated the extent of host specialization of Z. tritici in the Tunisian context. Cross‐inoculations of different strains derived from either BW or DW were performed at the seedling stage on a panel of the most cultivated BW and DW varieties in Tunisia using the percentage of pycnidia coverage and necrotic leaf area as disease parameters. Results showed that the effect of strains, the strain‐derived species, the wheat species and of their interactions were highly significant based on pycnidia coverage. This indicates the presence of significant virulence variability among Z. tritici strains on the tested wheat genotypes, with some strains being more virulent than others. It also suggests cultivar and host species specificity, as not only genotypes of the strains interacted differently with the genotypes of the host, but also the origin of the strain, whether BW or DW, influenced the disease outcome. Hence, future research based on comparative genomics coupled with transcriptomics of these strains and their progenies could determine genes whose expression control this host specificity or is influenced by host or strain genotype, or alternatively by life cycle stage. [ABSTRACT FROM AUTHOR]

Crop diseases can cause major yield losses, so the ability to detect and identify them in their early stages is important for disease control. Deep learning methods have shown promise in classifying multiple diseases; however, many studies do not use datasets that represent real field conditions, necessitating either further image processing or reducing their applicability. In this paper, we present a dataset of wheat images taken in real growth situations, including both field and glasshouse conditions, with five categories: healthy plants and four foliar diseases, yellow rust, brown rust, powdery mildew and Septoria leaf blotch. This dataset was used to train a deep learning model. The resulting model, named CerealConv, reached a 97.05% classification accuracy. When tested against trained pathologists on a subset of images from the larger dataset, the model delivered an accuracy score 2% higher than the best‐performing pathologist. Image masks were used to show that the model was using the correct information to drive its classifications. These results show that deep learning networks are a viable tool for disease detection and classification in the field, and disease quantification is a logical next step. [ABSTRACT FROM AUTHOR]