2.2.2 Metabolic networks

Cellular metabolism is the set of chemical reactions that take place in the cell and ensure its life.

The concept of metabolism is essential to understand life itself.

It converts available substances to the cell into energy and molecular blocks to build cellular components.

It obeys the laws of chemistry and thermodynamics (like any set of chemical reactions) and it has the ability to shift and modulate its structure to adapt to a wide range of external environments and ensure the cell’s survival.

Metabolic networks (Silva et al.

2008) are composed by metabolites and the reactions that relate them.

Metabolites are normally small molecules such as monosaccharides, amino acids or inorganic ions.

The biochemical reactions are often catalysed by enzymes, a type of protein that works lowering the activation energy of

Figure 2.7. Metabolic network representing the glycolysis pathway in Escherichia coli K12 MG1655 from the EcoCyc. those reactions making them occur at much faster rates.

Metabolites and reactions have been traditionally organized in metabolic pathways.

Those are defined as a group or successive metabolic reactions that carry out a specific cellular function as a group.

The genome-scale metabolic network of an organism includes all the different pathways that represent sections or branches of the network.

How to divide a network in functional modules is one of the core problems this thesis tackles.

Biochemical pathways is the traditional metabolic network division.

It progressively originated when biochemists started to determine the functions of particular enzymes.

If the product of one reaction was the substrate of another then those reactions were linked.

This manually curated method eventually formed a network of metabolites and reactions.

Pathways are simply the groups of reactions that made most biochemical sense to group together.

Boehringer Mannheim chart (http://web.expasy.org/pathways/ ) is the quintessential division of cellular metabolism.

However, recently some efforts have been made to analyse the partitioning of the metabolic networks from a more network-level approach (Papin et al. 2003).

Visualizing and displaying metabolic networks is much harder than it seems.

In the most popular metabolic databases, such as EcoCyc (Keseler et al. 2013) and KEGG (Kanehisa and Goto 2000), metabolites are presented by vertices and reactions by edges joining them.

Reactions usually carry the names of the enzymes that catalyse them.

Sometimes the EC number (Enzyme Commission number) can be present too.

EC numbers contain four digits in the form x.x.x.x that represent the type of reaction being catalysed by one particular enzyme.

Reaction name, enzyme and EC number are used as synonims in these representations.

Figures 2.7 and 2.8 show those features clearly.

Figure 2.8. Metabolic network representing the glycolysis pathway in Escherichia coli K12 MG1655 from the KEGG database.

This approach is very intuitive and useful for understanding the basic layout of the system in a quick look.

However, it lacks a rigorous mathematical description that can be used to study the network much further.

This issue has significant implications for metabolic network analysis and will be addressed in detail in Chapter 5.

Some metabolites, usually called pool metabolites (Ma and Zeng 2003b) such as H2O, CO2 or adenosin triphosphate (ATP), appear in these representations many times distorting the real connectivity of the network.

These pool metabolites are usually overlooked when studying metabolic networks; sometimes are downweighted to be less important or even removed completely from the network.

There are valid reasons to support these simplifications but each one distance us from a rigorous mathematical description of the network.

Other example is the clumping together of the edges representing a concrete reaction.

That means that usually a reaction with two substrates and two products is represented by only one link instead of four.

Figure 2.9 shows the traditional and the direct unipartite graph representation.

Once again, this makes the representation simpler but complicates the network analysis.

Figure 2.9. Metabolic network representing the general glycolysis pathway (from Ma and Zeng 2003b).

(a) Traditional representation of glycolysis. (b) Unipartite graph representation of glycolysis.