Fig. 1

Workflow diagram for the OrthoFiller algorithm. a Proteomes are subdivided into orthogroups using OrthoFinder. b Protein sequences in each orthogroup are subject to multiple sequence alignment, back-translated to DNA and used to create hidden Markov models (HMMs). These HMMs are used to search each genome in the set. c The set of hits are evaluated and filtered to remove low quality hits. d Gene models are constructed around each retained hit using Augustus. e The new gene models are compared to the hints that were used to generate them, and filtered to remove those which bear insufficient similarity to the hints. f The filtered genes are clustered into orthogroups and genes that are successfully placed into the orthogroup that was used to identify them are retained. g The process may be run once, or iteratively until no further genes are found