Figure 2:

Method overview to identify split protein-coding regions. Genes are depicted as boxes, protein-coding regions are indicated with gray areas. White areas indicate introns or untranslated regions (UTRs). If two CDSs annotated as part of different genes in the partial genome consistently map to non-overlapping parts of a common gene in several reference genomes, this suggests that the two CDSs are part of a split protein-coding region and should be merged (refer to Methods section for details).