I would like to do blastx analysis of a de novo assembled transcriptome against KEGG database. I read the related material on the KEGG web site, but it is not still clear for me as a beginner. How I can download the database? Please share if you have any experience with it?
Yes, but I don't know why some uncommon pathway like those involved in the human disease represented in the results while I'm working on plants. That's why I would like to do local blastx, which sound that not free.
ADD REPLY
• link
updated 20 months ago by
Ram
43k
•
written 8.6 years ago by
seta
★
1.9k
0
Entering edit mode
KOs are often associated with multiple pathways and as such you should be looking at complete pathways (all KOs present) instead of single pathway annotations.
So having unusual pathways in the results is normal? sorry, how I can look at the complete pathway instead of single pathway, can you please more explanation what's your mean?
For example K00016 is associated with 5 different pathways. The same way, some of your KO annotations that you interpret as "human disease" can be associated with many different pathways. Just because you have some KO in your data does not mean that you have all the pathways that are associated with said KO. If you want to be relatively sure about some pathway being present, you have to check if all the KOs that are associated with said pathway are present in your data (and even that doesn't necessarily mean anything).
Pay $5,000 fee for individual non-commercial license or $10,000 for institutional license and you'll gain access to their ftp. It's possible to also download every single protein in the KEGG db utilizing the rest api, but it will take a long, long time and include millions of calls to their server. I don't think that kind of usage of their API is encouraged and could maybe even lead to ban. If you can't get a license, then your only KO-annotation option is what they offer for free. Or maybe the last free KEGG is out there somewhere, I don't know. Even if it was, it wouldn't be up to date though..
did you try KEGG KAAS? It takes sequences (both nucleotide and protein) - annotates and assigns KO id to the sequences.
Yes, but I don't know why some uncommon pathway like those involved in the human disease represented in the results while I'm working on plants. That's why I would like to do local blastx, which sound that not free.
KOs are often associated with multiple pathways and as such you should be looking at complete pathways (all KOs present) instead of single pathway annotations.
So having unusual pathways in the results is normal? sorry, how I can look at the complete pathway instead of single pathway, can you please more explanation what's your mean?
For example K00016 is associated with 5 different pathways. The same way, some of your KO annotations that you interpret as "human disease" can be associated with many different pathways. Just because you have some KO in your data does not mean that you have all the pathways that are associated with said KO. If you want to be relatively sure about some pathway being present, you have to check if all the KOs that are associated with said pathway are present in your data (and even that doesn't necessarily mean anything).
yes. I also observed unusual pathway when I used it. It might be what 5heikki is saying. But, I myself do not know exact reason.