Link to Pubmed [PMID] – 25917918
G3 (Bethesda) 2015 Apr;5(7):1323-33
The fungal pathogen Zymoseptoria tritici (synonym Mycosphaerella graminicola) is a prominent pathogen of wheat. The reference genome of the isolate IPO323 is one of the best-assembled eukaryotic genomes and encodes more than 10,000 predicted genes. However, a large proportion of the previously annotated gene models are incomplete, with either no start or no stop codons. The availability of RNA-seq data allows better predictions of gene structure. We here used two different RNA-seq datasets, de novo transcriptome assemblies, homology-based comparisons, and trained ab initio gene callers to generate a new gene annotation of Z. tritici IPO323. The annotation pipeline was also applied to re-sequenced genomes of three closely related species of Z. tritici: Z. pseudotritici, Z. ardabiliae, and Z. brevis. Comparative analyses of the predicted gene models using the four Zymoseptoria species revealed sets of species-specific orphan genes enriched with putative pathogenicity-related genes encoding small secreted proteins that may play essential roles in virulence and host specificity. De novo repeat identification allowed us to show that few families of transposable elements are shared between Zymoseptoria species while we observe many species-specific invasions and expansions. The annotation data presented here provide a high-quality resource for future studies of Z. tritici and its sister species and provide detailed insight into gene and genome evolution of fungal plant pathogens.