Figure 3

Identification of incorrect gene models in the C. graminicola database. A. Statistical analysis of incorrect gene models. In total, 906 gene models were improved with respect to incorrect introns or exons, incorrect splice sites, incorrectly split or joined genes and sequencing errors. B-E. Examples of the four categories of incorrectly annotated genes. B. Read coverage of the coding sequence of GLRG_02449 convincingly shows that the first intron is somewhat longer than predicted. C. Exon five of GLRG_06933 harbours an unidentified intron. Using RNA-Seq data, an additional short exon at the immediate 3′-terminus needed to be added to the gene structure. D. Gene GLRG_08413 and GLRG_08414 were incorrectly split due to the occurrence of two introns (blue bars). The correct annotation shows a single ORF that consists of four exons (green bar). E. Insertion of an adenine into the genomic DNA sequence of GLRG_08997 led to frame-shift and premature stop of translation. In the current annotation, the conflict was apparently solved by inserting an intron which could be identified as incorrect. Blue bars indicate incorrect cds predicted by automated gene annotation, green bars highlight the re-annotated gene structure based on RNA-Seq data, grey bars represent individual RNA-Seq reads and red arrowheads indicate the positions of errors in the current annotation of the C. graminicola genome.