Three faces of metabolites:
pathways, localizations and
network positions
Jing Zhao(赵静）1, Petter Holme2,3
1 Department of Mathematics, Logistical Engineering University, China
2 Department of Physics, Umeå University, Sweden
3 Department of Energy Science, Sungkyunkwan University, Korea
The Fourth International Conference on Computational Systems Biology
(ISB 2010) , September 9-11, 2010 Suzhou, China
Outline





Background
Relation between the categories of
metabolites
Quantitative difference between categories
Network structures of the categories
Conclusion
Metabolism
metabolic network
Subcellular compartments
Cytoplasm, [c]
Extracellular Space, [e]
Mitochondrion, [m]
Golgi Apparatus, [g]
Endoplasmic Reticulum, [r]
Lysosome, [l]
Peroxisome, [x]
Nucleus, [n]
Pathways
Carbohydrate Metabolism, C
Energy Metabolism, E
Lipid Metabolism, L
Nucleotide Metabolism, N
Amino Acid Metabolism, A
Glycan Biosynthesis and Metabolism, G
Metabolism of Cofactors and Vitamins, V
Xenobiotics Biodegradation and Metabolism, X
Biosynthesis of Secondary Metabolites, S
Transport, T
Modular topology of metabolic network
Simulated annealing method
r
max M  max [eii  ( eij ) 2 ]
i 1
j
M: the modularity metric for a given
decomposition of the network
eij: the fraction of edges that leads
between vertices of cluster i and j
the sum is over the partition into
clusters
This study…
Data used
BiGG database
Duarte, N. C.; Becker, S. A.; Jamshidi, N.; Thiele, I.; Mo, M. L.;
Vo, T. D.; Srivas, R.; Palsson, B. O., Global reconstruction of the
human metabolic network based on genomic and bibliomic data.
PNAS 2007, 104, (6), 1777-1782.
Relation between the categorizations
M=0.676
Quantitative difference between categories:
overlap score
Prototypical overlap score

 
XY
(x, y)  X (x) Y (y)
x X y Y
Normalized overlap score
 XY 
 XY
max( XX , YY )

X ,Y : two categorizations
X(x) ,Y(y) : the fraction of metabolites in category x  X, y Y, respectively
XY(x,y): the joint frequency of x and y, i.e. the fraction of vertices that are
categorized both as x  X and y  Y.
Quantative difference between a feature of the real metabolic
network and its randomized counterparts: Z-score
f  fr
Z 
f r
f : the metric of the feature in the real network
fr
: the mean of the corresponding metric in the randomized ensemble
f r : the standard deviation of the corresponding metric in the randomized
ensemble
Overlap scores between categories
Localization vs. network clusters :  = 0.72, z =108.46;
Network clusters vs. pathways:  = 0.37, z = 45.94;
Pathways vs. localization : = 0.42, z = 54.69
Network structure of the categories of the three categorizations
Conclusions
• We have investigated three ways of grouping metabolites
according to pathways, localization, and network clusters
respectively.
• The compartmentalization is clearly organized into a core of
extracellular media and cytoplasm, and a periphery of
organelles; while the core and periphery modules of the
pathway interactions correspond to housekeeping and
advanced specific functions respectively.
• Qualitative and quantitative analysis shows that the three
categories are significantly correlated with each other,
suggesting the interrelationship between biochemical specific
functions, spatial isolations in cells, and topological compact
regions in metabolic networks.
Acknowledges
National Natural Science Foundation of China (10971227)
Swedish Foundation for Strategic Research
Swedish Research Council
the WCU program through NRF Korea funded by MEST
R31–2008–000–10029–0
Thanks!