Genome-wide conservation and locus-specific coding variation in Teratoramularia rumicicola, a candidate microbial herbicide for Rumex weeds

The Rumex pathogen Teratoramularia rumicicola is a candidate microbial herbicide, but its functional genomic repertoire and variation among isolates remain poorly characterised. We generated a PacBio HiFi reference and evidence-supported gene annotation for TR4 and a short-read assembly for GR1, then compared the isolates using assembly- and read-based analyses. The TR4 and GR1 nuclear assemblies comprised 22.34 Mb (13 contigs; N50, 2.90 Mb) and 21.23 Mb (486 contigs; N50, 0.350 Mb), respectively. The TR4 annotation predicted 8,898 protein-coding genes, and the representative protein set was 96.7% complete in the Benchmarking Universal Single-Copy Orthologues assessment. Functional annotation identified 593 putatively secreted proteins, 258 candidate effectors, 355 carbohydrate-active enzyme (CAZyme) candidates and 20 predicted secondary-metabolite biosynthetic regions. Seven secreted CAZymes were prioritised as potential host-barrier-associated factors based on predicted activities and homology to virulence-related entries in the Pathogen-Host Interactions database. Despite 99.81% average nucleotide identity across aligned regions, read-based comparison identified 28,040 high-confidence isolate-fixed differences. Of 158 reconstructed candidate GR1 alleles, 147 showed strong assembly support. Notable differences included predicted truncations of two Zn(II)2Cys6 transcription factors, one adjacent to an intact nonribosomal peptide synthetase-like core gene, and a separate gene showing a localised repeat-induced point mutation (RIP)-like substitution pattern and predicted premature termination. Together, these resources and candidate loci provide a foundation for investigating pathogenicity and strain variation relevant to microbial herbicide development.