Converting to homologs

Question

[Lkal]

I was wondering what you guys do when converting gene IDs to another species homologous genes. Specifically, when one gene maps to multiple genes in the target species. If I apply the “value”, say RNA-seq CPMs for example, to all of the newly mapped homologs it over represents those genes when they multi-map.

Picking one at random is what I typically do, as the homologs are likely of similar function anyways, so its probably representative of the multi-map groups.

There is a similar issue when going from gene IDs -> proteins, to map protein protein interactions in cytoscape for example.

Maybe there is a simple and better solution I am not seeing. Any suggestions?

Edit: I generally need to convert gene IDs to another species as GO terms are not available for every species yet. So for this use the exact gene matters less and function more so.

Answer 1

[Geparada]

I would suspect that other people have already done efforts in the gene homology direction already, as it is something that many people wants to do when you are working with more than one model organism or when you want to homologate everything to human so you can get the best usage of the available databases to solve your problem.

Thus, I would recommend looking for homology databases first, like this one, but I am genuinely interested to know how do people solve the multi-mapping issue that you are stating.