I've been looking through the files available at www.reactome.org along with some others but I only find images/pdfs of metabolic pathway diagrams. Where is the actual data these pathways are drawn from? I imagine there should be some kind of directed graph representation. I'm limited to freely available datasets, so unfortunately I'm not able to check out those such as KEGG.
-
$\begingroup$ KEGG should be open for anyone, if I am not completely wrong. $\endgroup$– Chris ♦Jan 22, 2014 at 19:10
-
$\begingroup$ According to their download page, access to the ftp files for KEGG require a paid academic subscription (through bioinformatics.jp/en/keggftp.html). I suppose the tools on their website are free, but I haven't found the pathway data I'm looking for (that is, I see where I'm only able to browse pathway diagrams). $\endgroup$– AaronJan 22, 2014 at 19:21
-
$\begingroup$ Well, now I see that KEGG has a XML markup of their pathways, but it is only available from querying their API (kegg.jp/kegg/xml). I wonder if there is an easy way to grab all pathway files. $\endgroup$– AaronJan 22, 2014 at 19:34
1 Answer
KEGG is definitely your choice!
It is possible to get the whole list(s) of pathways: http://rest.kegg.jp/list/pathway (reference pathways) or http://rest.kegg.jp/list/pathway/hsa (for human).
Then you can use these IDs to retrieve information about individual pathways: http://rest.kegg.jp/get/hsa00053 (GeneBank-like format), http://rest.kegg.jp/get/hsa00053/image (image), http://rest.kegg.jp/get/hsa00053/kgml (KEGG xml). It will take a while, of course, and I'd recommend not to overload their server and make delays between queries. With the GeneBank-like format batch querying is also possible: http://rest.kegg.jp/get/hsa00053+hsa00061
This should likely be a one-time or at least rare operation, because of the load on their server (and I don't think they make big changes very often). You can retrieve info for new pathways, once they appear in http://rest.kegg.jp/list/pathway/...
-
$\begingroup$ Yes, this is actually precisely what I ended up doing. I wrote a python script which iterated through the human pathway file, sent a get request to their API and just saved each kgml file to hard disk. For the GeneBank-like format, those don't indicate any relationship among gene or compounds involved right? It looks like it is simply a list. $\endgroup$– AaronJan 23, 2014 at 16:15
-
1$\begingroup$ @Aaron: as far as I know, they are just listed. It's a pity, the API has no batching for the xml output, but KEGG is great any case! $\endgroup$ Jan 23, 2014 at 18:57