Back

Hybran: Hybrid Reference Transfer and ab initio Prokaryotic Genome Annotation

Elghraoui, A.; Gunasekaran, D.; Ramirez-Busby, S. M.; Bishop, E.; Valafar, F.

2022-11-10 bioinformatics
10.1101/2022.11.09.515824 bioRxiv
Show abstract

De novo assembly has become commonplace for microbial organisms, increasing the demand for reliable genome annotation. Ab initio annotation is not an ideal approach for closely related strains due to suboptimal matching of the short or hypervariable genomic features that reference-based annotation transfer can overcome through identification of conserved synteny. At the same time, reference-based annotation methods leave gaps in the annotation where structural variations introduce unique sequence. We present Hybran, a hybrid reference-based and ab initio prokaryotic genome annotation pipeline that transfers features from a curated reference annotation and supplements unannotated regions with ab initio predictions. It builds on existing tools to create initial annotations using both approaches, then compares and resolves them to produce the hybrid annotation. With this pipeline, full advantage is taken of the communitys experimental efforts on reference strains to propagate as many known features as possible without sacrificing best-effort ab initio predictions for the remaining unannotated loci. Genome annotation performed in this way can facilitate comparative genomics and the investigation of evolutionary dynamics in microbial populations. Hybran is freely available at https://lpcdrp.gitlab.io/hybran

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.