Database and Tool updates

[sanyalab]

Hello,

I wanted to know how frequently will the tool and the OMAmer database be updated?
Further, are there any biases in the database that might lead to over-estimation of scores? For example will the OMArk tool perform well for an assessment of Oomycota proteome?
Lastly, the data for OMArk is curated? or is it pulled from UniprotKB

Thanks
Abhijit

[YanNevers]

Hello,

OMArk itself is actively maintained, and we try to patch upcoming issue as soon as possible, when they arise. For now, we do not have new features on our radar, but I will consider any feature suggestions from users.

The OMAmer database, however, is updated as often as the OMA Database. And OMA is updated at least once a year, with update focused mostly in increasing the diversity of species. You can consult the species currently covered here: https://omabrowser.org/oma/release/ and make suggestion for new species here (if you wish to): https://omabrowser.org/oma/suggestion/genome/.

Like other reference-based software, we believe there is indeed some bias in OMArk estimation, mainly due to sample bias and scarcity of reference genomes for some clade (including Oomycota). This may result in a bit of overestimation of completeness and overestimation of erroneous gene models. (You can read the benchmarking section of our paper for more information)
For this reason, we suggest users to always compare their OMArk results to the "best" results of close species. This should provide a good estimation of how well their annotation compare to best standards.

We believe this effect will be less and less impactful as we had more proteomes to our database.

Finally, regarding the source and quality of the data included in OMA. The proteomes come from a diversity of sources, included some from UniProt. However, we do not automatically pull new data from databases, and we curate the list of genomes we decide to include at each release.
Whenever we make a new addition, we consider both the taxonomic importance of a species, as well as quality indicators (OMArk and BUSCO results, general quality of gene models, status of the assembly). Nowadays, we add mainly annotation from trusted sources, like Ensembl and RefSeq.