[Paper Review] Propagation of fluctuations in interaction networks governed by the law of mass action

Sergei Maslov, Kim Sneppen+1 more|arXiv (Cornell University)|Nov 7, 2006

Gene Regulatory Network Analysis18 citations

TL;DR

This paper develops a matrix formalism to model how small perturbations in protein concentrations propagate through interaction networks governed by the law of mass action. It demonstrates that such fluctuations decay exponentially with network distance, with the decay rate determined by network topology, binding strength, and concentration correlations—providing a theoretical basis for understanding signal propagation and cross-talk in biological networks.

ABSTRACT

Using an example of physical interactions between proteins, we study how perturbations propagate in interconnected networks whose equilibrium state is governed by the law of mass action. We introduce a comprehensive matrix formalism which predicts the response of this equilibrium to small changes in total concentrations of individual molecules, and explain it using a heuristic analogy to a current flow in a network of resistors. Our main conclusion is that on average changes in free concentrations exponentially decay with the distance from the source of perturbation. We then study how this decay is influenced by such factors as the topology of a network, binding strength, and correlations between concentrations of neighboring nodes. An exact analytic expression for the decay constant is obtained for the case of uniform interactions on the Bethe lattice. Our general findings are illustrated using a real biological network of protein-protein interactions in baker's yeast with experimentally determined protein concentrations.

Motivation & Objective

• To understand how localized perturbations in protein concentrations propagate through biological interaction networks governed by the law of mass action.
• To quantify the attenuation of these fluctuations with distance from the perturbation source, especially in the context of minimizing cross-talk between biological pathways.
• To investigate how network topology, binding affinities, and correlations between interacting protein concentrations influence the propagation and decay of fluctuations.
• To develop a resistor network analogy for predicting perturbation responses in non-linear, mass-action-driven systems.
• To validate theoretical predictions using a real-world yeast protein-protein interaction network with experimentally measured concentrations.

Proposed method

• Formulates a linearized matrix approach to relate small changes in total protein concentrations (δC_i) to changes in free concentrations (δF_i), using the Jacobian of the mass action equilibrium equations.
• Derives a system of linear equations (Eq. 4) that models the relative perturbation response across the network, expressed as δF_i / F_i in terms of a connectivity-weighted matrix Λ.
• Applies a resistor network analogy, where perturbations propagate like current through a network of resistors and shunt conductances, with effective conductances derived from binding constants and free concentrations.
• Uses the Bethe lattice as an analytically tractable model to derive an exact expression for the exponential decay constant λ (Eq. 12) in regular, homogeneous networks.
• Extends the model to include non-bipartite structures (e.g., homodimers and odd-length loops), treating them as small perturbations that enhance dissipation and reduce signal propagation.
• Validates results using a real yeast PPI network with measured concentrations, comparing real and reshuffled concentration distributions to assess the impact of concentration correlations.

Experimental results

Research questions

• RQ1 How do small changes in total protein concentrations propagate through a network of mass-action-governed interactions?
• RQ2 What determines the rate of exponential decay of perturbations with increasing network distance from the source?
• RQ3 How does network topology—especially degree distribution and presence of odd-length loops—affect signal propagation and attenuation?
• RQ4 To what extent do correlations between concentrations of interacting proteins influence the stability and propagation of fluctuations?
• RQ5 Can the resistor network analogy accurately predict the response of non-linear mass-action systems to perturbations?

Key findings

• Perturbations in free protein concentrations decay exponentially with network distance, with the decay constant λ determined by binding strength and network structure.
• On the Bethe lattice with uniform parameters, the exact decay constant is given by λ = -1/(d−1) × [(d + k/F)/2 − √((d + k/F)/2)² − (d−1)], which is always between -1 and 0, indicating oscillatory decay.
• In the strong binding limit (k/F ≪ 1), the decay constant approaches λ ≈ -1/(d−1) × (1 − 1/(d−2) × √(dk/C)) for d > 2, showing that higher degrees and stronger binding reduce decay rate.
• For a linear chain (d=2), the decay constant simplifies to λ_L = [ (1+8C/k)^{1/4} − 1 ] / [ (1+8C/k)^{1/4} + 1 ], approaching -1 in the strong binding limit, indicating near-undamped propagation.
• Any variation in total concentration (e.g., oscillating C_i) increases the average free concentration and thus accelerates perturbation decay, with |λ_±| ≤ |λ_L|, confirming that inhomogeneity suppresses propagation.
• The observed positive correlation (ρ = 0.27, p < 10^−54) between interacting protein concentrations in the yeast network enhances network susceptibility to perturbations compared to randomized networks, as confirmed by simulation.