betaFeedback

WorksMathematics

Bulk stability of bimolecular stochastic chemical reaction networks

Abstract

We establish criteria for stability in the bulk of planar stochastic chemical reaction networks with mass-action kinetics. Using Poincaré compactification, we characterize the asymptotic drift and show how stochastic fluctuations can prevent escape along transversally repulsive branches of the limiting deterministic flow. We construct a Lyapunov function by combining local estimates near these branches with dissipative estimates in the neighboring regions. The criteria cover a wide class of networks and in particular apply beyond weakly reversible and strongly endotactic networks. Together with compatible boundary estimates, these criteria imply positive recurrence.

Introduction

Stability is a fundamental requirement for the study of the long-time behavior of chemical reaction networks. For stochastic models, stability is often understood as (positive) recurrence, giving an invariant probability measure on each closed irreducible class and allowing one to study the system at stationarity. This is especially important in the multiscale setting, where the effective dynamics of slow variables are often obtained by averaging over the stationary distribution of a fast subsystem [10, 25]. Establishing such a reduction requires control of the fast dynamics, including their stability and the relevant moment bounds.

Chemical reaction networks admit two classical descriptions. The deterministic model follows concentrations through a system of mass-action ordinary differential equations. The stochastic model tracks molecular copy numbers through a continuous-time Markov chain, retaining the fluctuations that are particularly important at low copy numbers [21]. Under the natural volume scaling, the deterministic equations arise as a law of large numbers limit of the stochastic dynamics on finite time intervals [28, 27]. However, this connection does not, by itself, identify their long-time stability properties.

Stability in this context has been extensively studied in both deterministic and stochastic settings. Starting with the theory of complex balance [23, 18], the deterministic literature has developed structural conditions relating the reaction graph to persistence and permanence, notably weak reversibility and endotacticity [14, 13, 22]. Related conditions also play a central role in the stochastic theory. Complex balance yields product-form stationary distributions [8, 7], while the geometry of strongly endotactic networks underlies stochastic Lyapunov estimates and large-deviation results [1, 2, 6].

The two descriptions can nevertheless behave very differently near the boundary of the state space, where some species have low copy numbers. Extinction, trapping near the boundary and changes of recurrence class can occur in the stochastic model without a corresponding deterministic behavior [9, 5, 4, 6]. In the bulk, where all species have large copy numbers, the stochastic drift is instead well approximated by the deterministic vector field. Much of the stability analysis in this region has consequently treated the stochastic dynamics as a perturbation of the deterministic one, using deterministic inward drift to construct stochastic Lyapunov functions [2, 6]. This suggests a close correspondence between the two models in the bulk, but does not establish an equivalence of their stability conditions.

In our previous article [3], we showed that this correspondence can fail even without boundary effects. We constructed a network whose deterministic system contains a trajectory that diverges in finite time in the interior of the state space, while the stochastic model is positive recurrent. The divergent trajectory is repulsive in the transverse direction, and stochastic fluctuations push the process away towards regions where the dynamics dissipate. While an earlier example based on a special choice of rate constants appears in [7 Example 4], the system from [3] displays the stabilization by noise for generic choices of such constants. These examples show that deterministic bulk stability is not necessary for stochastic stability.

This paper continues our program by giving general criteria for bulk stability of two-species networks. The goal is to give conditions on the deterministic vector field that guarantee sufficient stability (encoded through Lyapunov-type dissipation inequalities) in the bulk of the stochastic system. Building on the ideas of [3], we focus on its behavior at infinity: which curves lead to infinity, and whether they attract or repel neighboring trajectories. Our aim is to introduce new ideas and a new approach to studying stochastic dynamics in the bulk, without seeking a complete classification of bulk stability.

To obtain this information, we carry out a compactification of the deterministic vector field, using Poincaré and Poincaré-Lyapunov compactifications. These compactifications are particularly well suited to polynomial planar vector fields, see [15], and allow us to “spread” infinity along a circle, so that different asymptotic directions can be studied separately. Different scaling regimes must be analysed, determined by the graph of the CRN studied.

The singular points at infinity of the compactified vector field describe the asymptotic behavior of invariant curves. We distinguish convergent and divergent curves, along which the flow goes towards smaller and larger states, respectively. Any such (convergent or divergent) curve can be further classified as attractive or repulsive according to whether neighboring trajectories move towards it or away from it. The example in [3] shows that a repulsive divergent curve in the deterministic system can still allow stability of the stochastic system. Under the non-degeneracy and drift assumptions stated below, we show stability in the bulk when all attractive curves are convergent.

To prove this result we construct, for each such system, an energy function dissipated in the bulk, which can be combined with compatible boundary estimates to obtain positive recurrence. The proof of the dissipation inequality relies on a careful approximation of the infinitesimal generator of the process in each scaling regime. These approximations, given in Propositions 5.3 and 6.3, are used throughout the paper, both to show pointwise dissipation of the energy function and to derive scaling limits for the process near repulsive curves. We then patch the local energy functions together to obtain dissipation throughout the bulk.

Our result significantly extends previous work on the stability of chemical reaction systems with stochastic mass-action kinetics by exploiting the purely stochastic phenomenon of stabilization by noise. In particular, when combined with complementary dissipation estimates on the boundary, it allows us to establish positive recurrence for networks outside the endotactic family, thereby extending the class of stochastically stable networks well beyond the deterministically stable ones. This further clarifies the relationship between the asymptotic behavior of deterministic and stochastic models of chemical reaction networks.

At the conceptual level, the main novelty of our approach lies in using the Poincaré compactification of the deterministic vector field to derive criteria for stability in the bulk of stochastic CRNs. The main technical challenge stems from the generality of our results: the broad family of networks considered gives rise to a number of qualitatively different cases requiring separate, ad hoc treatment. To address this challenge, we develop a principled approach based on a careful decomposition of the generator in various scaling regimes, allowing us to obtain the desired dissipation estimates for all the different asymptotic behaviors of the model. To combine these estimates, we also develop a conceptually straightforward but rather general patching lemma that may be of independent interest. We believe our assumptions to be close to optimal, apart from some asymptotically degenerate cases that we explicitly discuss and leave for future work.

The paper is organized as follows. Section 2 presents the general setting of CRNs. Section 3 introduces Poincaré compactification and illustrates it with two examples. The main results and their scope of application are presented in Section 4. The rest of the paper is devoted to the proof of the main theorem. In Sections 5 and 6, we construct energy functions near repulsive curves in the diagonal and sub-diagonal regions, respectively, and prove their dissipation. Section 7 deals with attractive curves, completing the analysis within each compactification. Section 8 treats the regions between the compactifications, called drift regions. Finally, Section 9 patches the local energy functions together to complete the proof.

Model and notations

General notations

2.1. General notations. We write Z\mathbb{Z} for the integers and N=Z≥0\mathbb{N} = \mathbb{Z}_{\ge0} for the non-negative integers. As r→∞r \to\infty, O(r)O(r) denotes a quantity whose ratio to rr is bounded, and o(r)o(r) one whose ratio to rr tends to zero. The notation x∼rx \sim r means that xx is “of order r”,i.e.,r”, i.e.,|x/r|$ is upper and lower bounded (away from 0) when r→+∞r \to+\infty. We use the Euclidean norm ∥⋅∥\lVert\cdot\rVert and scalar product ⟨⋅,⋅⟩\langle\cdot,\cdot\rangle.

Chemical reaction networks

2.2. Chemical reaction networks. We introduce here the standard notations used in the theory of chemical reaction networks (CRN).

A CRN is a triplet of finite sets (S,C,R)(\mathcal{S},\mathcal{C},\mathcal{R}) with

  • S\mathcal{S} the set of species. We restrict the study to planar CRNs, with only two species. Therefore, S={S1,S2}\mathcal{S} = \{S_1, S_2\}.

  • C\mathcal{C} is the set of complexes. It is a subset of N∣S∣\mathbb{N}^{|\mathcal{S}|}, each coordinate giving the number of molecules of the corresponding species in the complex.

  • R\mathcal{R} is the set of reactions. It is a subset of {(y1,y2)∈C2:y1≠y2}\{(y_1,y_2) \in\mathbb{C}^2 : y_1 \ne y_2\}.

A given CRN is usually represented as a directed graph, with vertices CC and edges R\mathcal{R}. An element y=(y1,y2)∈Cy=(y_1,y_2)\in C is often written as y1S1+y2S2y_1S_1+y_2S_2, and (0,0)(0,0) as ∅\varnothing. An element (y−,y+)∈R(y^-,y^+)\in\mathcal{R} is written y−→y+y^-\to y^+.

We introduce the set of input vertices, or source complexes,

C−:={y∈C:∃y→y+∈R}.C^-:=\{y\in C:\exists y\to y^+\in\mathcal{R}\}.

We assume that each of the species S1S_1 and S2S_2 is present in at least one of the source complexes. For a reaction ν:yν−→yν+\nu:y^-_\nu\to y^+_\nu, we write ξν=yν+−yν−\xi^\nu=y^+_\nu-y^-_\nu for the reaction vector, giving the net change in molecule numbers when the reaction occurs. We set K0:=max⁡ν∈R∥ξν∥K_0:=\max_{\nu\in\mathcal{R}}\|\xi^\nu\|. For S⊂Z≥02S\subset\mathbb{Z}_{\geq0}^2, its two-sided K0K_0-boundary is

∂K0(S):={x∈Z≥02:dist⁡(x,S)≤K0, dist⁡(x,Sc)≤K0},\partial^{K_0}(S):=\{x\in\mathbb{Z}_{\geq0}^2:\operatorname{dist}(x,S)\leq K_0,\ \operatorname{dist}(x,S^c)\leq K_0\},

where Sc=Z≥02∖SS^c=\mathbb{Z}_{\geq0}^2\setminus S. This contains both endpoints of any reaction jump crossing the boundary of SS.

Deterministic system. The state of the deterministic system is given by the concentrations x(t)∈R≥02x(t)\in\mathbb{R}_{\geq0}^2. The kinetics of the reactions follow the law of mass action. Given a set of positive reaction rate constants κ=(κν)ν∈R\kappa=(\kappa_\nu)_{\nu\in\mathcal{R}}, the ODE associated to the system (S,C,R,κ)(S,C,\mathcal{R},\kappa) is

x˙=Δ(x):=∑ν∈Rκνxyν−xyν+ξν.(1)\dot{x}=\Delta(x):=\sum_{\nu\in\mathcal{R}}\kappa_\nu x^{y^-_\nu}x^{y^+_\nu}\xi^\nu. \tag*{(1)}

For x∈R≥02x\in\mathbb{R}_{\geq0}^2 and y∈Cy\in C, we use the notation xy:=x1y1x2y2x^y:=x_1^{y_1}x_2^{y_2}.

Stochastic system. The stochastic system tracks the number of molecules of each species. At time tt, its state is given by X(t)∈Z≥02X(t)\in\mathbb{Z}_{\geq0}^2. The reaction kinetics again follow the law of mass action. The rate of a reaction ν\nu depends on the state of the system, and is zero when there are not enough reactant molecules. Explicitly,

λν(x):={0if x1<yν,1− or x2<yν,2−,κνx1!(x1−yν,1−)!x2!(x2−yν,2−)!otherwise(2)\lambda_\nu(x):= \begin{cases} 0 & \text{if }x_1<y^-_{\nu,1}\text{ or }x_2<y^-_{\nu,2},\\ \kappa_\nu\frac{x_1!}{(x_1-y^-_{\nu,1})!}\frac{x_2!}{(x_2-y^-_{\nu,2})!} & \text{otherwise} \end{cases} \tag*{(2)}

Equivalently,

λν(x)=κνx1(x1−1)⋯(x1−yν,1−+1)x2(x2−1)⋯(x2−yν,2−+1)(3)\lambda_\nu(x)=\kappa_\nu x_1(x_1-1)\cdots(x_1-y^-_{\nu,1}+1)x_2(x_2-1)\cdots(x_2-y^-_{\nu,2}+1) \tag*{(3)}

for x∈Z≥02x\in\mathbb{Z}_{\geq0}^2, with empty products equal to one.

Given a set of reaction rate constants κ=(κν)ν∈R\kappa=(\kappa_\nu)_{\nu\in\mathcal{R}}, the stochastic process associated to the system (S,C,R,κ)(S,C,\mathcal{R},\kappa) is a continuous time Markov chain whose infinitesimal generator acts on a test function gg as

A(g)(x)=∑ν∈Rλν(x)(g(x+ξν)−g(x)).(4)A(g)(x)=\sum_{\nu\in\mathcal{R}}\lambda_\nu(x)\bigl(g(x+\xi^\nu)-g(x)\bigr). \tag*{(4)}

Polytope

The source polytope is the convex hull of C−C^-. To determine which reactions dominate at large states, we consider scalings xi∼Nωix_i\sim N^{\omega_i}, with N→∞N\to\infty, and normalize the directions by max⁡{ω1,ω2}=1\max\{\omega_1,\omega_2\}=1. Thus, we introduce

Ω={(1,ω2),ω2∈[0,1]}∪{(ω1,1),ω1∈[0,1]}.\Omega=\{(1,\omega_2),\omega_2\in[0,1]\}\cup\{(\omega_1,1),\omega_1\in[0,1]\}.

We order Ω\Omega in the trigonometric sense, i.e. for ω,ω′∈Ω\omega,\omega'\in\Omega,

ω≤ω′ if and only if ω2≤ω2′ and ω1≥ω1′.\omega\leq\omega'\ \text{if and only if }\omega_2\leq\omega'_2\text{ and }\omega_1\geq\omega'_1.

Also, ω<ω′\omega< \omega' if ω≤ω′\omega\leq\omega' and ω≠ω′\omega\ne\omega'. For ω−≤ω+\omega^- \leq\omega^+,

[ω−,ω+]:={ω∈Ω:ω−≤ω≤ω+},[\omega^-,\omega^+] := \{\omega\in\Omega: \omega^- \leq\omega\leq\omega^+\},
(ω−,ω+):={ω∈Ω:ω−<ω<ω+}.(\omega^-,\omega^+) := \{\omega\in\Omega: \omega^- < \omega< \omega^+\}.

Finally, we introduce Ω≤={ω∈Ω:ω≤(1,1)}\Omega_{\leq} = \{\omega\in\Omega: \omega\leq(1,1)\}.

Along the scaling xi∼Nωix_i \sim N^{\omega_i}, the source monomial xy−x^{y^-} has order N⟨y−,ω⟩N^{\langle y^-,\omega\rangle}. The largest exponent therefore identifies the dominant source complexes and their reactions.

References

  1. [1]Andrea Agazzi, Amir Dembo, and Jean-Pierre Eckmann. Large deviations theory for markov jump models of chemical reaction networks. The Annals of Applied Probability, 28(3):1821–1855, 2018.arxiv.org/abs/1701.02126
  2. [2]Andrea Agazzi, Amir Dembo, and Jean-Pierre Eckmann. On the geometry of chemical reaction networks: Lyapunov function and large deviations. Journal of Statistical Physics, 172:321–352, 2018.DOI
  3. [3]Andrea Agazzi and Lucie Laurence. Noise-induced stabilization in a chemical reaction network without boundary effects. 2025.arxiv.org/abs/2506.12163
  4. [4]Andrea Agazzi and Jonathan C. Mattingly. Seemingly stable chemical kinetics can be stable, marginally stable or unstable. Communications in Mathematical Sciences, 18(6):1605–1642, 2020.arxiv.org/abs/1810.06547
  5. [5]David F Anderson and Daniele Cappelletti. Discrepancies between extinction events and boundary equilibria in reaction networks. Journal of mathematical biology, 79(4):1253–1277, 2019.DOI
  6. [6]David F Anderson, Daniele Cappelletti, Jinsu Kim, and Tung D Nguyen. Tier structure of strongly endotactic reaction networks. Stochastic Processes and their Applications, 130(12):7218–7259, 2020.arxiv.org/abs/1808.05328
  7. [7]David F Anderson, Daniele Cappelletti, Masanori Koyama, and Thomas G Kurtz. Non-explosivity of stochastically modeled reaction networks that are complex balanced. Bulletin of mathematical biology, 80:2561–2579, 2018.DOI
  8. [8]David F Anderson, Gheorghe Craciun, and Thomas G Kurtz. Product-form stationary distributions for deficiency zero chemical reaction networks. Bulletin of mathematical biology, 72:1947–1970, 2010.DOI
  9. [9]David F. Anderson, Germán A. Enciso, and Matthew D. Johnston. Stochastic analysis of biochemical reaction networks with absolute concentration robustness. Journal of the Royal Society Interface, 11(93):20130943, 2014.DOI
  10. [10]Karen Ball, Thomas G. Kurtz, Lea Popovic, and Greg Rempala. Asymptotic analysis of multiscale approximations to reaction networks. The Annals of Applied Probability, 16(4):1925 – 1961, 2006.DOI
  11. [11]Patrick Billingsley. Convergence of probability measures. Wiley Series in Probability and Statistics: Probability and Statistics. John Wiley & Sons Inc., New York, second edition edition, 1999. A Wiley-Interscience Publication.
  12. [12]Daniele Cappelletti, Aidan Howells, and Chuang Xu. Stability of randomly switching stochastic reaction networks with asymptotically linear transition rates, 2025.arxiv.org/abs/2507.09421
  13. [13]Gheorghe Craciun. Polynomial dynamical systems, reaction networks, and toric differential inclusions. SIAM Journal on Applied Algebra and Geometry, 3(1):87–106, 2019.arxiv.org/abs/1901.02544
  14. [14]Gheorghe Craciun, Fedor Nazarov, and Casian Pantea. Persistence and permanence of mass-action and power-law dynamical systems. SIAM Journal on Applied Mathematics, 73(1):305–329, 2013.arxiv.org/abs/1010.3050
  15. [15]Freddy Dumortier, Jaume Llibre, and Joan C Artés. Qualitative theory of planar differential systems. Springer, 2006.DOI
  16. [16]Kacha Dzhaparidze and JH Van Zanten. On bernstein-type inequalities for martingales. Stochastic processes and their applications, 93(1):109–117, 2001.DOI
  17. [17]S.N. Ethier and T.G. Kurtz. Markov Processes: Characterization and Convergence. Wiley Series in Probability and Statistics. Wiley, 2009.
  18. [18]Martin Feinberg. Complex balancing in general kinetic systems. Archive for rational mechanics and analysis, 49(3):187–194, 1972.DOI
  19. [19]Y. Filonov. A criterion for the ergodicity of discrete homogeneous Markov chains. Akademiya Nauk Ukrainskoi SSR. Institut Matematiki. Ukrainskii Matematicheskii Zhurnal, 41(10):1421–1422, 1989.DOI
  20. [20]David A Freedman. On tail probabilities for martingales. the Annals of Probability, pages 100–118, 1975.DOI
  21. [21]Daniel T. Gillespie. Exact stochastic simulation of coupled chemical reactions. The Journal of Physical Chemistry, 81(25):2340–2361, 1977.DOI
  22. [22]Manoj Gopalkrishnan, Ezra Miller, and Anne Shiu. A geometric approach to the Global Attractor Conjecture. SIAM Journal on Applied Dynamical Systems, 13(2):758–797, 2014.DOI
  23. [23]Fritz Horn and Roy Jackson. General mass action kinetics. Archive for rational mechanics and analysis, 47:81–116, 1972.DOI
  24. [24]Jean Jacod and Albert N. Shiryaev. Limit theorems for stochastic processes, volume 288 of Grundlehren der mathematischen Wissenschaften. Springer-Verlag, Berlin, second edition, 2003.
  25. [25]Hye-Won Kang and Thomas G. Kurtz. Separation of time-scales and model reduction for stochastic reaction networks. The Annals of Applied Probability, 23(2):529 – 583, 2013.arxiv.org/abs/1011.1672
  26. [26]G. S. Katzenberger. Solutions of a stochastic differential equation forced onto a manifold by a large drift. Ann. Probab., 19(4):1587–1628, 1991.DOI
  27. [27]Thomas Kurtz. Limit theorems for sequences of jump markov processes approximating ordinary differential processes. Journal of Applied Probability, 8(2):344–356, 1971.
  28. [28]Thomas G Kurtz. Solutions of ordinary differential equations as limits of pure jump markov processes. Journal of applied Probability, 7(1):49–58, 1970.
  29. [29]Lucie Laurence and Philippe Robert. Stochastic chemical reaction networks with discontinuous limits and AIMD processes. Stochastic Processes and their Applications, 186:Paper No. 104643, 32, 2025. Arxiv (PDF).arxiv.org/abs/2406.12604
  30. [30]Lucie Laurence and Philippe Robert. Scaling methods for stochastic chemical reaction networks. Stochastic Processes and their Applications, 194:104855, April 2026. Arxiv PDF.arxiv.org/abs/2310.01949
  31. [31]Jaume Llibre and Clàudia Valls. Global qualitative dynamics of the brusselator system. Mathematics and Computers in Simulation, 170:107–114, 2020.DOI
  32. [32]Lawrence Perko. Differential equations and dynamical systems, volume 7. Springer Science and Business Media, 2013.
  33. [33]Henri Poincaré. Mémoire sur les courbes définies par une équation différentielle. Journal de mathématiques pures et appliquées, 7:375–422, 1881.
  34. [34]L. C. G. Rogers and David Williams. Diffusions, Markov processes, and martingales. Vol. 1. Cambridge Mathematical Library. Cambridge University Press, Cambridge, 2000. Foundations, Reprint of the second (1994) edition.
  35. [35]Galen R Shorack and Jon A Wellner. Empirical processes with applications to statistics. SIAM, 2009.

Paper details

vlnc:W_ffynfjtr
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 1

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Artificial intelligence tools

View statement
Looking for related papers…
Contents