Try to initialise a metabolic reaction bipartite graph as a DB and we link out from there.
currently the data is listed as below:

ko -> cpd -> ko

1. every rxn is given a rxn ID as well as cpd
   in 2 formats (xml or raw)

XML

R00001: C00404 + n C00001 <=> (n+1) C02174

filepath: kegg_dump/ligand/reaction/reaction.lst (eqn) or reaction (rxn name & details)
filepath: kegg_dump/ligand/compound/compound (names, id & details)

::More information::

A single entry in reaction (perl script can be reused /// is the entry separator

perl -0777 -ne '@A=split(qq(///),$_); END{print $a[1]}' ligand/reaction/reaction

ENTRY R00002 Reaction
NAME Reduced ferredoxin:dinitrogen oxidoreductase (ATP-hydrolysing)
DEFINITION 16 ATP + 16 H2O + 8 Reduced ferredoxin <=> 8 e- + 16 Orthophosphate + 16 ADP + 8 Oxidized ferredoxin
EQUATION 16 C00002 + 16 C00001 + 8 C00138 <=> 8 C05359 + 16 C00009 + 16 C00008 + 8 C00139
COMMENT a part of multi-step reaction (see R05185, R00002+R00067+R00153+R02802+R04782)
RPAIR RP00003 C00002_C00008 main
RP00010 C00002_C00009 leave
RP05676 C00001_C00009 leave
ENZYME 1.18.6.1

rxns2KOs
every enz/KO comes with a list of associated reactions

filepath: /export2/home/uesu/kegg_dump/genes/ko/ko

perhaps need to output to gephi then to igraph?