A Detailed Curation of the Zymoseptoria tritici IPO323 Gene Annotation

Functional and comparative genomics based projects are dependent upon the availability of a genome sequence and accompanying genome annotation. The quality of the annotation can have numerous downstream effects on the research. If a gene model is inaccurate or not present then SNP consequences may be misinterpreted or overlooked respectively. Zymoseptoria tritici is an important foliar pathogen of wheat. A complete genomic reference sequence is available for Z. tritici isolate IPO323, however the currently available three genome annotations from JGI (10,933 genes), Rothamsted Research (10,688 genes) and Max Plank Institute (11,839 genes) contain a number of inaccurately predicted gene models. For example, some gene models are lacking either start or stop codon, and some have two physically closely linked genes miss-annotated as one. Moreover, several gene models can only be found in one of the three genome annotations but not in the other two. Here we present results of a comprehensive re-annotation of IPO323 followed by its manual curation. To achieve this, the three earlier genome annotations were merged together with a fourth, new annotation that we produced using Maker2 with new transcript and protein supporting evidences. We identified a large number of gene models that were missed or miss-annotated in previous genome annotations. Our resulting annotation consists of 13,862 gene models with 899 of those representing genes encoding secreted proteins, which is a 55% increase (317 new genes) on the JGI annotation set of 582 and 7.5% increase (125 new genes) on the Max Plank Institute set of 774 genes.