Hi,
Given the following input data:
EC numbers for the enzymes involved + compound ids of reactants involved in each reaction (we actually have to do conversion steps from gene names and arbitrary compound ids in a SBML file before in order to generate these, but I am positive that this is the easiest step).
Does anybody know or has anybody already implemented a programmatic way of determining whether or not a reaction in KEGG or maybe ecocyc metacyc is reversible or not, and would like to share the principle or code, in whatever language?
The background: We are trying to evaluate and quality check a genome scale metabolic model (in SBML) and we suspect that it erroneously has be added a single (hopefully not more) reversible reaction where the reaction in reality is irreversible. So it would be possible but tedious process to check manually with the kegg maps if the reversibility in our model conforms with the reaction in KEGG.
I hope this has become clear, if not please comment for more information.
•
written
by
Michael Dondrup
A while ago I have asked a similar question to KEGG's maintainers, and they told me in metabolism pathways, they define only binary relations and do not give the information of directions to relations.
If you look at the KGML's specifications you can see that inside the Reaction element there is an attribute called 'directionality'. However, as the same documentation explains, that should not be interpreted as the directionality of a reaction, but is rather an inference made from the literature.
MetaCyc does have information about reaction direction. You can get it via the web service, eg:
% curl http://websvc.biocyc.org/getxml?META:1.8.4.8-RXN
<ptools-xml ptools-version="15.0" xml:base="<a href=" <a="" href="http://BioCyc.org/getxml?META:1.8.4.8-RXN" rel="nofollow">http://BioCyc.org/getxml?META:1.8.4.8-RXN" "="" rel="nofollow">http://BioCyc.org/getxml?META:1.8.4.8-RXN'>
...
<Reaction ID="META:1.8.4.8-RXN" orgid="META" frameid="1.8.4.8-RXN" detail="full">
....
<reaction-direction>REVERSIBLE</reaction-direction>
...
I don't think KEGG or *cyc have this info, but BRENDA has a table "Substrate/Product" with a column "Reversibility":
r=reversible
ir=irreversible
?=not specified
Here's an example entry: direct link to substrate table, complete frame
You can download BRENDA, but I haven't tried it now to check how easy it is to access the reversibility information.
•
written
by
Michael Kuhn
thank you Michael we will look at the example, if possible we would like to stick with kegg though because the model was originally derived from kegg.
Log in to answer this question.
