We have completed annotation of our rosid de novo genome assemblies with MAKER. This is a fairly complete package of tools that does repeat masking, aligns transcript/protein evidence, produces ab initio predictions, and synthesizes results into annotations with evidence-based quality values. Beyond evidence from alignment with transcript/protein evidence, confidence in annotations can be boosted with matches to a Pfam domain or the Uni/SwissProt database.

Here is a look at our Malus ioensis annotations. There are ~62k annotations, and over 60% of these have align with input transcript/protein evidence, contain a Pfam domain, or align with something in the Uni/SwissProt database.