Yes, this all is correct, except that re-aligning sequencing data with yet another reference is simply infeasible in current settings. Meanwhile, I guess there is a way. At least I found it for thioredoxin 1: its KOALA page maps it to COG0526, which seems to be 404red at NCBI, but I guess this way can solve my issue.
Hi folks:
From what I have learned from biologists, COG use to be (or even is) widely used for measuring "stress response" genes expression in microbial communities. In the ongoing project I am relying on KEGG for analyses and wondering if there a way to map KOs to COG -- so we can see how our results align with the previous research?
UPDATE: well, turned out that KEGG provides a file ko_cog.list through their FTP which makes mapping super easy.
Thanks!
2 answers
It turned out that KEGG provides a file named ko_cog.list through their FTP which makes mapping super easy. Thanks!
Given that underlying functional hierarchy is different (KOs map to GO, COGs don't), I don't think you can obtain one-to-one mapping for all identifiers (you will get one-to-many quite frequently, plus lots of missing ones). This will make any comparison with the previous research quite difficult.
As for practice, I don't think mapping file exists. Uniprot's id mapping service maps KOs only to eggNOG identifiers (while eggNOG shares the same functional hierarchy as COG, I didn't find eggNOG to COG mapping file). The fastest way is to map your data to selected COGs is to use sequence similarity (assuming you have sequences). Blast, rapsearch or something like that.
Cool, I didn't know about KOALA. BTW, you could add your solution as an answer and later accept it. That way the question will be marked as answered (you don't get points for answering your own question, so it's not seen as wrong behaviour) and site will be kept clean.
Log in to answer this question.