Improved genome annotation of Rhynchosporium commune isolate UK7 using Illumina short reads of in vitro and in plantae conditions

Improved genome annotation of the Rhynchosporium commune isolate UK7 using Illumina short reads of in vitro and in plantae conditions. The short reads used for the annotation are available at https://doi.org/10.5281/zenodo.5729968 and https://doi.org/10.5281/zenodo.5729863.To create the gene models, we used tophat v. 2.0.14 to align short reads to the UK7 reference genome (Trapnell et al., 2009). The Intron splice site hints were generated using bam2hints, included in the AUGUSTUS v. 3.2.1 software (Stanke et al., 2006). Due to the very high RNA-sequencing depth available, intron splice hints were filtered for a minimum coverage of 20 reads to avoid an impact of spurious splice signals on gene prediction. To produce ab initio gene models, the BRAKER v. 1.0 pipeline (Hoff et al., 2016) combining GeneMark-ET ab initio gene model predictions and AUGUSTUS v. 3.2.1. GeneMark-ET was trained using the RNA-seq-based splice information as hints. AUGUSTUS was automatically trained using ab initio gene models that were fully supported by splice information. Finally, AUGUSTUS was used to predict gene models using both RNA-seq splice information and coding sequence hints based on exonerate protein alignments as extrinsic evidence.