Résumé
An understanding of the evolutionary dynamics that allow plant pathogens to break down host resistance and develop fungicide resistance, is needed to develop durable and efficient control methods. To address the plasticity of fungal genomes, long-read genome sequences and adequate bioinformatic support is required to generate high-quality chromosome-level assemblies. In this study, the variability in the structural genomes of two important banana pathogens, Pseudocercospora fijiensis (causing the black Streak) and Fusarium oxysporum f.sp. cubense TR4 (causing the Fusarium wilt), were investigated. We first developed the workflow " Podium ASM " to assess the quality of long-read genome assemblies, based on contig numbers, genome completeness, and the presence of telomeric sequences. A second workflow, named " EffiCAZ ", was then used to improve functional annotation of pathogenicity-related effectors and CAZymes. Chromosome-level genome assemblies were obtained from four Nanopore sequences of each species, and compartmentalised into a core and an accessory part, with contrasting gene densities and effector distribution. A much higher proportion of transposable elements was observed in P. fijiensis than in F. oxysporum f.sp. cubense TR4. These new pipelines pave the way for the comprehensive characterization of the pangenomes of the two banana pathogens.