Can OMA calculate orthologs for single query sequence?

Question

[Andrzej Zielezinski]

Is it possible to search for orthologs/inparalogs only for specific sequences in the query genome?

I'm aware that searching for orthologs requires complete proteomes. I would like to avoid a step where OMA calculates all-versus-all sequence comparisons of all proteins from query and subject genomes. In fact, I don't have to BLAST all proteins in a proteome when I'm interested only in one protein. I was wondering whether OMA can be used this way? For example, Reciprocal Smallest Distance (RSD) approach lets users to find orthologs for specific sequences in the query genome.

Answer 1

[Christophe Dessimoz]

If your specific sequence is part of a genome in the OMA database, it is of course possible to retrieve the inferred orthologs for this sequence only. For this, use the search function with an identifier or the sequence itself.

If the sequence is not in OMA, you can export a few closely related genomes and run the OMA standalone on them and your sequence. Assuming that the genomes are somewhat complete, the orthologs inferred for your sequence should be fairly reliable. Thanks to the reciprocal criterion (roughly the RSD criterion you mention) and the use of "witnesses of non-orthology", things only start to fall apart when most genomes are very incomplete.

Answer 2

[cyril-cros]

You have a very large many number of databases with precomputed groups of orthologs. Try http://orthodb.org/ which is quite complete.

Another technique is to look at genes in the vicinity of your protein, they may be part of a synteny group.