Closed vkkodali closed 10 months ago
Hello @vkkodali,
Thank you very much for reporting this issue.
I updated the README to make it clear how the OMAmerDB available through omark
's (Zenodo)[https://zenodo.org/record/7359861#.Y_p493_MJQo] relate to the ones available through the OMA Browser, described in the omamer
README. To sum it up, the OMAmerDB we made available through Zenodo is a copy of the LUCA.h5 database from the December2021 release of the OMA database (The current OMA release is from November 2022).
I apologize for the confusion and I hope this updates clear it up, we will consider additional steps to make these differences more clear fin the future.
We recommend using omark
with an OMAmer database constructed from all the species in the OMA database ( so-called LUCA.h5 database), which allows the species and contamination detection module to function as expected. Other databases (Metazoa.h5, Viridiplantea.h5, Primates.h5) only cover species from specific taxa and will only allow omark
to detect species from these taxa.
I believe using either the above mentioned (OMAmerDB)[https://zenodo.org/record/7359861/files/OMAmerDB.gz] or (LUCA.h5)[https://omabrowser.org/All/LUCA.h5] from the OMA Browser should allow you to run the omark
pipeline as expected with no more errors.
However, OMArk should in principle be also compatible with more specific databases. I will work on fixing this issue you reported as soon as possible and will keep you updated.
Thank you for your patience and let us know if you need any additional assistance.
Thank you @YanNevers, using LUCA.h5 for both omamer search
and omark
commands worked! Going forward, I will only use the latest version of LUCA.h5 for my work.
Do you know if there's any versioning of the LUCA.h5 database? How frequently is it updated? Is there any way to extract the metadata, especially the version information, from the database file itself? It will help when reporting results to specify which version of LUCA.h5 was used instead of saying something like "downloaded from www.omabrowser.org on 2023-02-24".
I am glad to read omark
is now working for you!
To answer your questions, we aim to have release new release of the OMA database every 6 to 12 months, although specific time may vary depending on the amount of computation needed.They are generally announced through the OMA Browser main page, our newsletter, RSS feed and social media. You can access those in the bottom right corner of the Browser home page. Corresponding OMAmer databases will be released at the same time.
As for now, there is no explicit versioning of the LUCA.h5 database, but there is still a way to confirm which version of the dataset you are using. In the OMAmer output file, the first letter of the HOG identifiers is release specific since the Apr2021 release of OMA. Thus given the first letter you can infer:
However, we agree we need a more straightforward way to access metadata from the database itself. We plan to soon issue an update to OMAmer where release information and the parameters used to construct the database will be directly available through the command line. Additionally, we will add a feature where the databases issued from a main release of OMA will issue a warning if a more recent database is available online.
I just issued a patched omark
release (0.2.5) where the initial issue (omark
not working with Metazoa.h5 database) is fixed. This should have no impact over the use of LUCA.h5 database but will make sure omark
is able to run with any omamer
database going forward, as would be expected. Thank you again for reporting this!
As of previous comment, detailed OMAmer metadata is now available with the command omamer info --db DB
. Hopefully, it will hlep with reporting results.
Pardon the uninformative title for the issue; I wasn't sure how exactly I should describe the error. I am trying to run OMArk (version 0.2.4) on a set of about 38000 proteins from the species Mercenaria mercenaria. Here's how I went about it:
Can you please help me with the error? Am I supposed to use a different database? It is unclear to me what the difference between
OMAmerDB
mentioned in the OMArk README and the other databases (such as LUCA.h5, Metazoa.h5, etc) mentioned in theomamer
README page is. In this instance, I have used the latest version ofMetazoa.h5
.