Clément Plessis

Accurate and reproducible assessment of foliar disease severity is essential for evaluating the performance of heterogeneous plant communities and understanding host-pathogen interactions. However, traditional visual scoring methods remain subjective, with limited precision, and difficult to scale in large phenotyping experiments. Here, we present a semi-automated image analysis workflow designed to quantify multiple foliar disease symptoms simultaneously on wheat flag leaves sampled from varietal mixtures. The workflow combines three methodological components: (i) a standardized protocol for leaf sampling and imaging, (ii) supervised machine learning segmentation using Random Forest implemented in Ilastik to classify multiple symptoms (powdery mildew and yellow rust), and (iii) a graphical user interface facilitating pipeline deployment by non-specialist operators. To evaluate the influence of image representation on classification performance, four color spaces (RGB, HSV, HLS, LAB) were systematically compared. The approach was validated using images of durum wheat flag leaves collected from a field experiment assessing eight-way varietal mixtures under natural fungal pressure. Cross-validation against manually annotated images demonstrated high segmentation accuracy across all symptom. Comparison among color spaces revealed only minor differences in performance. Overall, this workflow offers a cost-effective, annotation-efficient and reproducible alternative to deep learning approaches, leveraging open-source and actively maintained tools while requiring limited training data and enabling objective, reproducible and scalable disease phenotyping.

Biotic interactions between plants and insects can drive key evolutionary processes. In Mediterranean agroecosystems, the harvester ants Messor barbarus (Hymenoptera: Formicidae) frequently collect seeds, including those of cultivated cereals. Yet their potential role in shaping crop traits remains poorly understood. This study investigates whether harvester ant seed predation is driven by genetic and phenotypic variation in durum wheat ( Triticum turgidum ssp. durum ), a major Mediterranean crop derived from wild emmer ( T. turgidum ssp. dicoccoides ). Using a panel of 180 genetically diverse durum wheat inbred lines grown in a field experiment, we visually recorded spike predation and performed a genome‐wide association study (GWAS) using SNP markers to assess the genetic architecture of susceptibility to seed predation by M. barbarus . We identified a significant quantitative trait loci (QTL) on chromosome 2A explaining 21% of the variation in predation rate. This region contains a 3.6 Mb chromosomal inversion and 46 candidate genes, including a MYB transcription factor potentially involved in regulating cuticle and chemical traits. To validate these genetic findings, we conducted a cafeteria experiment with 208 spikes from 26 genotypes, placed at the entrances of eight ant nests. Ants preferentially removed spikes from genotypes carrying the allele identified in the GWAS. Additionally, shorter spikes were more likely to be harvested. However, unlike previous studies on wild plants, seed morphology and protein content did not significantly affect ant preference. Synthesis . Our results demonstrate that M. barbarus exhibits genotype‐specific preferences in durum wheat, associated with a major QTL, and is influenced by spike traits. This study provides the first evidence of ant‐mediated selective pressure in a cereal crop and opens new perspectives on plant–insect dynamics in agroecosystems and the role of plant–insect interactions in the evolutionary history of crop species.

The taxonomic assignment of sequences obtained by high throughput amplicon sequencing poses a limitation for various applications in the biomedical, environmental, and agricultural fields. Identifications are constrained by the length of the obtained sequences and the computational processes employed to efficiently assign taxonomy. Arriving at a consensus is often preferable to uncertain identification for ecological purposes. To address this issue, a new tool called “ASVmaker” has been developed to facilitate the creation of custom databases, thereby enhancing the precision of specific identifications. ASVmaker is specifically designed to generate reference databases for allocating amplicon sequencing data. It uses publicly available reference data and generates specific sequences derived from the primers used to create amplicon sequencing libraries. This versatile tool can complete taxonomic assignments performed with pre-trained classifiers from the SILVA and UNITE databases. Moreover, it enables the generation of comprehensive reference databases for specific genes in cases where no directly applicable database exists for taxonomic classification tools.