Comparing multiple related genomes can help to improve their structural annotation. The accuracy and consistency of the predicted exon–intron structures of the protein coding genes can be higher when considering all genomes at once rather than annotating one genome at a time. The comparative gene prediction algorithm of AUGUSTUS performs such a multi-genome annotation. A multiple alignment of genomes is used to exploit evolutionary clues to conservation and negative selection. Further, AUGUSTUS exploits the fact that orthologous genes typically have congruent exon–intron structures. Comparative AUGUSTUS simultaneously predicts the genes in all input genomes. In this chapter we walk the reader through a small example from eight vertebrate species, including the construction of an alignment of the input genomes and how to integrate RNA-Seq evidence from multiple species for gene finding.
Nachtweide, S., & Stanke, M. (2019). Multi-genome annotation with AUGUSTUS. In Methods in Molecular Biology (Vol. 1962, pp. 139–160). Humana Press Inc. https://doi.org/10.1007/978-1-4939-9173-0_8