Leveraging the Marine Microbial Eukaryote Transcriptome Sequencing Project for Algal Genome Annotation

Algorithms utilizing protein homology for gene model prediction and annotation continue to be widely adopted. However, annotation quality is highly dependent on the input sequences included in your gene finding pipeline. Especially for highly divergent or understudied taxa, identification of appropriate sequences can be a difficult process.

A genome annotation project is underway for the photosynthetic marine microalgal species Phaeocystis globosa at the DOE Joint Genome Institute, where initial work has included homology seeds for GenomeScan and FgenesH algorithms from a set of 15 full transcriptomes and heavily filtered subsets of both Swiss-prot and TrEMBL databases. Phaeocystis is a component of the large-scale Marine Microbial Eukaryote Transcriptome Sequencing Project (MMETSP) that includes a total of 305 diverse species, totaling nearly 10 million protein sequences. Beginning with a set of 811,000 MCL sequence clusters (and 2.3 million singletons), we report a method for exploring the homology of the full MMETSP dataset to a single genome assembly. Phaeocystis annotation refinement to include the resulting filtered dataset of highly homologous MMETSP protein sequences, and this method’s general applicability to annotation of other marine microbial eukaryotes, is also discussed.