Questions: Stoichiometric Modeling

Question 1

What does the null space of the stoichiometric matrix represent biologically?

• A. The set of metabolites that are not consumed by any reaction
• B. The set of all flux distributions that satisfy steady-state mass balance for every internal metabolite
• C. The set of reactions that are thermodynamically irreversible
• D. The set of metabolites that accumulate over time

Answer: (B)

The equation S * v = 0 defines steady state: no internal metabolite accumulates or depletes. Any flux vector v in the null space of S satisfies this condition. The null space is typically a multi-dimensional subspace, meaning many different flux distributions are stoichiometrically feasible — the cell operates at one point in this space, selected by enzyme expression levels, thermodynamics, and regulation. Additional constraints (flux bounds, objective functions) narrow the feasible space, which is exactly what flux balance analysis does.

Question 2

If a metabolic network has 100 metabolites and 150 reactions, the stoichiometric matrix S is 100 x 150. The system S * v = 0 is underdetermined, meaning it has infinitely many solutions.

• T. True
• F. False

Answer: True

With more reactions (columns) than metabolites (rows), the null space of S has dimension at least 150 - 100 = 50 (assuming full row rank). This means there are at least 50 degrees of freedom in the flux solution — the system is underdetermined. This is typical for genome-scale metabolic models and is the fundamental reason why constraint-based methods add additional constraints (reaction bounds, objective functions, experimental flux measurements) to narrow the solution space to biologically meaningful predictions.

Question 3

Explain why the stoichiometric matrix framework can model genome-scale metabolic networks without requiring any kinetic parameters.

Model answer: The stoichiometric matrix encodes only the topology and stoichiometry of the network — which reactions exist, which metabolites they consume and produce, and in what ratios. The steady-state constraint S * v = 0 depends only on these stoichiometric coefficients, not on enzyme kinetics, binding affinities, or rate constants. This is its key advantage: stoichiometric data is available for entire genomes (from genome annotation and biochemical databases), while kinetic parameters are known for only a tiny fraction of enzymes. By constraining fluxes with stoichiometry, thermodynamics, and capacity bounds rather than kinetics, the framework scales to thousands of reactions.

This parameter-free quality is what enabled the construction of genome-scale metabolic models (GEMs) for hundreds of organisms. The E. coli model iML1515 contains 2,712 reactions and 1,877 metabolites — parameterizing this with Michaelis-Menten kinetics would require tens of thousands of rate constants, most of which are unknown.