Weakly reversible deficiency zero realizations of reaction networks

Abhishek Deshpande; Casian Pantea; Gheorghe Craciun; Neal Buxton

REVIEW 1 major objections 5 minor 52 references

Weakly reversible deficiency zero realizations of reaction networks

T0 review · 1 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read If a reaction network can be realized by a weakly reversible deficiency zero network for every rate vector, that realizing network is unique and rate-independent.

desk verdict A genuine extension of the WR0 realization program with a load-bearing but repairable gap in Lemma 4.4; worth sending to review. read the letter →

arxiv 2502.17461 v1 pith:3OMYLOXP submitted 2025-02-10 q-bio.MN math.DS

Neal Buxton , Gheorghe Craciun , Abhishek Deshpande , Casian Pantea This is my paper · ORCID

classification q-bio.MNmath.DS MSC 92C4237N2552A20

keywords weaklyreversiblenetworksdeficiencyzeroreactionnetworkrealizationmass-actionkineticsrateconstantsconeinclusiondynamicalequivalenceWR0-realizable

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks when a reaction network that is not itself weakly reversible and deficiency zero (WR0) can nevertheless produce, for every choice of rate constants $\kappa$, the same mass-action dynamics as some WR0 network. The central result is Theorem 4.3: if such a WR0 realization exists for all $\kappa$, then the realizing WR0 network is unique and does not change with $\kappa$. The authors also give Algorithm 2, which certifies whether a given network is WR0-realizable and returns the unique realizing network by running a known fixed-rate algorithm at $\kappa = 1$ and checking cone inclusions. A reader should care because WR0 networks have a unique, asymptotically stable equilibrium in every stoichiometric compatibility class, so a network that is merely WR0-realizable inherits these stability guarantees no matter what its actual rate constants are.

What carries the argument

The argument is carried by the cone geometry of a mass-action system. For a source complex $y$, $\operatorname{Cone}_{\mathcal{N}}(y)$ is the cone generated by the reaction vectors $y' - y$ for reactions $y \to y'$ leaving $y$, and $L_{\mathcal{N}}(y)$ is the larger cone generated by all complexes in $y$'s linkage class inside a WR0 realization. The net reaction vector $w_y$ is the coefficient column of the monomial $x^y$ after grouping terms. The proof uses Algorithm 1, which finds the unique WR0 realization of a fixed mass-action system from the extreme rays of $\ker W \cap \mathbb{R}_{\ge 0}^m$; Lemma 4.4 is meant to ensure no source monomial can vanish; Lemma 4.6 shows the linkage class partition is rate-independent; and Theorem 3.6 supplies uniqueness of the WR0 realization for a fixed ODE system.

What would settle it

Take any network that passes Algorithm 2 and, for each source complex $y$, solve $\sum_{y \to y'} \kappa_{y \to y'}(y' - y) = 0$ with all $\kappa_{y \to y'} > 0$; if any solution exists, Lemma 4.4 is false and the uniqueness theorem loses its support.

Watch

Extended reading notes

Core claim

The paper's central claim is that the relation '$\mathcal{N}$ is realizable by $\mathcal{N}'$ with $\mathcal{N}'$ weakly reversible and deficiency zero for every $\kappa$' is rigid: the network $\mathcal{N}'$ is independent of $\kappa$. More precisely, if $\mathcal{N}$ is WR0-realizable, then there exists a unique WR0 network $\mathcal{N}'$ such that for any rate constants $\kappa$ there are rate constants $k_\kappa$ for $\mathcal{N}'$ with $f_{\mathcal{N},\kappa} = f_{\mathcal{N}',k_\kappa}$. The proof proceeds by showing that the set of source complexes is rate-independent, that the linkage class partition of the realizing network is rate-independent, and that the face of the linkage-class cone generated by each net reaction vector is rate-independent; these three facts force the reaction set of $\mathcal{N}'$ to be fixed.

Load-bearing premise

The proof in Section 4 depends on Lemma 4.4 — no reactant complex can ever have its outgoing reaction vectors sum to exactly zero — and as written that lemma is not established, because the zero-sum complex it constructs is not a complex of the network used to reach the contradiction.

Editorial extensions

If this is right

Because the realizing WR0 network is unique, checking WR0-realizability needs only one run of the fixed-rate algorithm (at all rate constants equal to 1) followed by cone-inclusion checks.
Every mass-action system generated by a WR0-realizable network inherits the WR0 stability guarantees — a unique equilibrium in each compatibility class, asymptotically stable — for any rate constants.
The set of source complexes of the original network is rate-independent, so no source monomial can disappear from the dynamics as parameters vary.
The unique realizing network $\mathcal{N}'$ provides a fixed certificate: once found, it can be used to decide dynamical realizability by WR0 networks for the whole network, not just for one rate vector.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If Lemma 4.4 fails, the uniqueness theorem could fail too: rate constants might change which source complexes appear, and the 'unique' realization could depend on the parameter vector.
The set of rate constants for which a given network has a WR0 realization (Question 3 of the paper) is plausibly described by polynomial inequalities defining cones; this is a concrete next problem the paper leaves open.
The numerical tolerances used in the implementation suggest that an exact symbolic computation of extreme rays would make the certificate reliable near cone boundaries, where small numerical errors could flip the output.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Here is the short version: the main claim (Theorem 4.3) is a genuine extension of the fixed-parameter WR0 realization work of Craciun, Jin, and Yu, and Algorithm 2 is a natural, useful tool with a working CoNtRol implementation. But the proof as written has a gap in Lemma 4.4, which is load-bearing. The lemma is likely true and can be repaired, but the written argument does not do the job.

The paper does several things well. It states a clean question: if a reaction network is realizable by a weakly reversible deficiency zero network for every choice of rate constants, does the same WR0 network work for all of them? Theorem 4.3 says yes, and the uniqueness is not present in the earlier work. Algorithm 2 is correctly structured: run the existing fixed-parameter algorithm at κ=1, then check the cone inclusion condition from Theorem 2.8. The examples clarify where the fixed-parameter and all-parameters questions differ, and the software is a concrete deliverable.

Now the soft spot. Lemma 4.4 claims that for a WR0-realizable network, no source complex has 0 in the interior of its cone, so the source monomial set is constant across rate constants. The proof has two errors. First, it projects extreme rays of ker Wκ1 onto the first m−1 coordinates and claims these are extreme rays of ker Wκ0. But for a vector c=(d,c_m) in ker Wκ1, we only get Wκ0 d = −c_m w_ym, not Wκ0 d=0. So the projection need not lie in ker Wκ0. Second, the contradiction with weak reversibility of Nκ0 does not follow, because y0 has zero net vector in (N,κ0) and is therefore not a source complex, hence not a vertex, of Nκ0. Absence of reactions into y0 is not a weak reversibility violation.

Both issues are repairable. As the stress-test note observes, choose a generic κ with no cancellations; then y is a source complex of the WR0 realization Nκ, and Theorem 2.8 gives Cone_N(y) ⊆ Cone_{Nκ}(y). Since Nκ is WR0 deficiency zero, its linkage class cone is pointed, so 0 cannot lie in the interior of Cone_N(y). That is a short alternative proof of Lemma 4.4. So Theorem 4.3 is probably correct, but the paper needs a corrected proof before it is fully trustworthy.

Who should read it: people working on reaction network realizability, disguised toric systems, and computational structural analysis of mass-action networks. It is a solid subfield contribution, not a paradigm shift. I would recommend sending it to peer review, with a referee explicitly asked to check Lemma 4.4 and the projection argument.

Referee Report

1 major / 5 minor

Summary. The paper studies reaction networks N whose mass-action ODE systems admit a weakly reversible deficiency zero (WR0) realization for every choice of rate constants. The main result, Theorem 4.3, asserts that if such realizations exist for all κ, then the underlying WR0 network is independent of κ and unique. The authors propose Algorithm 2, which runs the known Algorithm 1 at κ=1 and then checks a cone-inclusion condition, to decide whether a given network is WR0-realizable and to return the unique WR0 network. Section 4 proves two lemmas (4.4 and 4.6) used to establish Theorem 4.3, and Section 5 reports an implementation in CoNtRol.

Significance. The uniqueness statement is natural and valuable: it turns the property 'has a WR0 realization for all rate constants' into a finite, checkable condition and strengthens the robustness conclusions available from the Deficiency Zero Theorem. The algorithmic component is a useful extension of Algorithm 1, and the paper is generally clearly written. The use of earlier published results by overlapping authors (Algorithm 1, Proposition 3.4, Theorem 3.6, Theorem 2.8) appears legitimate; these are used as tools, not as assumptions of the target conclusion. However, the proof of the key Lemma 4.4 is incomplete, so the main theorem is not yet established as written.

major comments (1)

[Section 4, Lemma 4.4] The proof of Lemma 4.4 is not valid as written. The step claiming that the first m−1 coordinates of the extreme rays of ker Wκ1 form a complete set of extreme rays of ker Wκ0 is false: for c=(d,c_m) in ker Wκ1, the equation Wκ1 c=0 gives Wκ0 d = −c_m w_{y_m}. If c_m>0, then d is not in ker Wκ0 at all, so the projected vectors are not extreme rays of ker Wκ0. Consequently the support-partition argument based on Proposition 3.4 does not apply. The final contradiction is also not a contradiction: because y0 has zero net vector in (N,κ0), y0 is not a source complex and not a vertex of Nκ0, so the absence of reactions yi→y0 does not violate weak reversibility of Nκ0. Since Lemma 4.4 is load-bearing for Theorem 4.3, this gap must be repaired; a correct proof will need to use the global 'for all κ' nature of the hypothesis rather than only the existence of a pointwise realization.

minor comments (5)

[Section 4, Lemma 4.4] In the sentence 'This implies that only the vertex y0 has zero net reaction vector in (N,κ)', the parameter should be κ0, not κ; likewise, the network denoted N0 near the end of the lemma should be Nκ0.
[Section 4, Lemma 4.6] The claim 'there is a linkage class of Nκ1 with the same complexes as L and so c_l0 is an extreme ray of Nκ1 as well' is too terse; after equality in (8), counting forces exactly one extreme ray of Nκ1 whose support is L, and that ray must be a positive multiple of c_l0. Please spell out this counting argument.
[Proof of Algorithm 2] The sentence beginning 'If, on the other hand, the algorithm goes through line 8' should be rephrased: failure of Cone_N(y)⊆Cone_N1(y) implies by Theorem 2.8 that N is not realizable by N1, and if N were WR0-realizable, Theorem 4.3 would force the realizing network to be N1, a contradiction. The current wording suggests that a particular κ0 with no N1-realization has been found, which is stronger than what the cone-inclusion check directly gives.
[Section 2, Definition 2.1] Please state explicitly whether reactions with y=y′ are allowed. If self-loops are permitted, Lemma 4.4 needs an additional hypothesis, since a source complex with only a self-loop has Cone_N(y)={0} and hence 0 lies in its relative interior.
[Section 4, Theorem 4.3 proof, Eq. (9)] Equation (9) would benefit from a one-line justification: since wy lies in the relative interior of Cone_N(y), the face of L(y) generated by wy is the face generated by the whole cone Cone_N(y). This is true, but it is not immediate from the notation used.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the all-kappa uniqueness theorem is derived from fixed-system lemmas and geometric cone facts, not assumed in its own proof; the self-citations are prior published tools, and the known gap in Lemma 4.4 is a correctness issue, not a circular step.

full rationale

Theorem 4.3 asks whether the underlying WR0 network can vary with the rate constants kappa when N is WR0-realizable. The proof does not assume that conclusion. Lemma 4.4 is a separate claim that no source-complex monomial cancels (0 not in int Cone_N(y)); Lemma 4.6 is a separate claim that the linkage-class partition of the WR0 realization is constant; and the final face-comparison step uses the fixed cone L(y) together with the fact that w_y lies in the relative interior of Cone_N(y). None of these steps is defined in terms of the theorem being proved. The cited results -- Theorem 2.8 from [50,19], Proposition 3.4 from [33], and Theorem 3.6 from [40] -- are prior published theorems about fixed mass-action systems, not restatements of the all-kappa uniqueness conclusion. They are used as lemmas, and they do not embed the target result as an assumption. Algorithm 2's correctness proof invokes Theorem 4.3, but Theorem 4.3 is proved before the algorithm's correctness is established and does not depend on the algorithm. The manuscript does contain a likely defect in the written proof of Lemma 4.4: the projection of extreme rays of ker W_kappa1 is not generally a complete set of extreme rays of ker W_kappa0, and y0 has zero net vector so the stated weak-reversibility contradiction is not immediate. This is a mathematical correctness gap, not circular reasoning: the lemma is a premise of Theorem 4.3, and the theorem is not assumed in the lemma's proof. No fitted parameter is renamed as a prediction, and no conclusion is equivalent to an input by construction. Therefore no circular steps are identified.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

No new physical entities or fitted parameters are introduced. The paper is a pure mathematical structural result; all assumptions are standard reaction network theory facts or cited prior theorems.

assumptions (6)

standard math Polyhedral cone facts: faces, extreme rays, simplicial cones.
Used throughout Section 2.1 and in the proofs.
domain assumption Theorem 2.8 characterization of dynamical realizability via cone inclusion.
Cited from Craciun and Pantea 2008 and Anderson et al. 2020; used in Algorithm 2's correctness proof.
domain assumption Proposition 3.4: for a WR0 mass-action system, ker W has a basis of nonnegative vectors with supports partitioning the monomial indices.
Cited from Craciun, Jin, Yu [33, Lemma 3.1]; central to the linkage class analysis.
domain assumption Proposition 3.5: complexes of any linkage class of a deficiency zero network are affinely independent.
Cited from Feinberg and Horn/Jackson; used to argue about faces of L(y).
domain assumption Theorem 3.6: a mass-action system has at most one WR0 realization.
Cited from Craciun, Jin, Yu [40, Theorem 3.11]; used for uniqueness.
domain assumption Algorithm 1 correctly decides whether a fixed mass-action system has a WR0 realization.
Taken from [33,51]; used as a subroutine in Algorithm 2.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Weakly reversible deficiency zero realizations of reaction networks." pith.science (2026). https://pith.science/paper/3OMYLOXP

@misc{pith2026250217461,
  author       = {Pith},
  title        = {Pith review of: Weakly reversible deficiency zero realizations of reaction networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3OMYLOXP}},
  note         = {Machine review of arXiv:2502.17461}
}

abstract

We prove that if a given reaction network $\mathcal{N}$ has a weakly reversible deficiency zero realization for all choice of rate constants, then there exists a $\textit{unique}$ weakly reversible deficiency zero network $\mathcal{N}'$ such that $\mathcal{N}$ is realizable by $\mathcal{N}'$. Additionally, we propose an algorithm to find this weakly reversible deficiency zero network $\mathcal{N}'$ when it exists.

Figures

Figures reproduced from arXiv: 2502.17461 by the authors.

**Figure 1.** Network N in (a) and N0 in (b) give rise to the same ODE system if we choose all rate constants equal to 3 for N and all rates equal to 1 for N0. Mass-action systems (N , 3) and (N0, 1) are called realizations of massaction system (5). Definition 2.6 (Realizations of mass-action systems). The mass-action system (N ′ , κ′ ) is called a realization of the mass-action ODE system dx dt = W xY if 6 [PITH_FULL_IMAGE:fig… view at source ↗

**Figure 2.** Mass-action system N in (a) has WR0 realization (b) when κ2 = κ3 We see that if (and in fact only if ) κ2 = κ3, then such a realization exists: the network N0 in [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

**Figure 3.** Mass-action system N in (a) has WR0 realization (b) when κ1 > κ4 8 [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 50 canonical work pages

[1]

Craciun and M

G. Craciun and M. Feinberg. Multiple equilibria in complex chemical rea c- tion networks: I. the injectivity property. SIAM J. Appl. Math. , 65(5):1526– 1546, 2005

work page 2005
[2]

Craciun and M

G. Craciun and M. Feinberg. Multiple equilibria in complex chemical re- action networks: II. the species-reaction graph. SIAM J. Appl. Math. , 66(4):1321–1338, 2006

work page 2006
[3]

Wilhelm and R

T. Wilhelm and R. Heinrich. Smallest chemical reaction system with ho pf bifurcation. J. Math. Chem. , 17(1):1–14, 1995

work page 1995
[4]

Wilhelm and R

T. Wilhelm and R. Heinrich. Mathematical analysis of the smallest che mical reaction system with hopf bifurcation. J. Math. Chem. , 19(2):111–130, 1996

work page 1996
[5]

´Erdi and J

P. ´Erdi and J. T´ oth.Mathematical models of chemical reactions: theory and applications of deterministic and stochastic models . Manchester University Press, 1989

work page 1989
[6]

Deshpande and M

A. Deshpande and M. Gopalkrishnan. Autocatalysis in reaction ne tworks. Bull. Math. Biol. , 76(10):2570–2595, 2014

work page 2014
[7]

Craciun and A

G. Craciun and A. Deshpande. Homeostasis and injectivity: a rea ction network perspective. J. Math. Biol. , 85(6):67, 2022

work page 2022
[8]

Craciun, A

G. Craciun, A. Deshpande, B. Joshi, and P. Yu. Autocatalytic re combina- tion systems: A reaction network perspective. Math. Biosci. , 345:108784, 2022. 17

work page 2022

Show all 52 references

[9]

Horn and R

F. Horn and R. Jackson. General mass action kinetics. Arch. Ration. Mech. Anal., 47(2):81–116, 1972

1972
[10]

F. Horn. Necessary and suﬃcient conditions for complex balanc ing in chem- ical kinetics. Arch. Ration. Mech. Anal. , 49(3):172–186, 1972

1972
[11]

Feinberg

M. Feinberg. Chemical reaction network structure and the st ability of com- plex isothermal reactors - I. the deﬁciency zero and deﬁciency on e theorems. Chem. Eng. Sci. , 42(10):2229–2268, 1987

1987
[12]

Toric dynamical systems

Gheorghe Craciun, Alicia Dickenstein, Anne Shiu, and Bernd Stur mfels. Toric dynamical systems. Journal of Symbolic Computation , 44(11):1551– 1565, 2009

2009
[13]

Feinberg

M. Feinberg. Foundations of chemical reaction network theory . Springer, 2019

2019
[14]

A reaction network appr oach to the convergence to equilibrium of quantum boltzmann equations for bose gases

Gheorghe Craciun and Minh-Binh Tran. A reaction network appr oach to the convergence to equilibrium of quantum boltzmann equations for bose gases. 2021

2021
[15]

C. Pantea. On the persistence and global stability of mass-act ion systems. SIAM J. Math. Anal. , 44(3):1636–1673, 2012

2012
[16]

Craciun, F

G. Craciun, F. Nazarov, and C. Pantea. Persistence and perm anence of mass-action and power-law dynamical systems. SIAM J. Appl. Math. , 73(1):305–329, 2013

2013
[17]

Craciun, F

G. Craciun, F. Nazarov, and C. Pantea. Persistence and perm anence of mass-action and power-law dynamical systems, 2013

2013
[18]

Anderson

D. Anderson. A proof of the global attractor conjecture in t he single linkage class case. SIAM J. Appl. Math. , 71(4):1487–1508, 2011

2011
[19]

Anderson, J

D. Anderson, J. Brunner, G. Craciun, and M. Johnston. On cla sses of reaction networks and their associated polynomial dynamical syst ems. J. Math. Chem. , 58:1895–1925, 2020

1925
[20]

Gopalkrishnan, E

M. Gopalkrishnan, E. Miller, and A. Shiu. A geometric approach to the global attractor conjecture. SIAM J. Appl. Dyn. Syst. , 13(2):758–797, 2014

2014
[21]

G. Craciun. Polynomial dynamical systems, reaction networks , and toric diﬀerential inclusions. SIAM J. Appl. Algebra Geom. , 3(1):87–106, 2019. 18

2019
[22]

Y. Ding, A. Deshpande, and G. Craciun. Minimal invariant regions and minimal globally attracting regions for variable-k reaction systems. Discrete Continuous Dyn. Syst. Ser. B. , 28(3):696–1718, 2023

2023
[23]

Minimal in vari- ant regions and minimal globally attracting regions for toric diﬀerent ial inclusions

Yida Ding, Abhishek Deshpande, and Gheorghe Craciun. Minimal in vari- ant regions and minimal globally attracting regions for toric diﬀerent ial inclusions. Adv. Appl. Math. , 136:102307, 2022

2022
[24]

Craciun and A

G. Craciun and A. Deshpande. Endotactic networks and toric d iﬀerential inclusions. SIAM J. Appl. Dyn. Syst. , 19(3):1798–1822, 2020

2020
[25]

Craciun, A

G. Craciun, A. Deshpande, and J. Jin. Weakly reversible deﬁcien cy one realizations of polynomial dynamical systems. Discrete Continuous Dyn. Syst. Ser. B. , 29(6):2786–2816, 2024

2024
[26]

Craciun, A

G. Craciun, A. Deshpande, and J. Jin. Weakly reversible single link age class realizations of polynomial dynamical systems: an algorithmic perspe ctive. J. Math. Chem. , 62(2):476–501, 2023

2023
[27]

Craciun, J

G. Craciun, J. Jin, and P. Yu. An eﬃcient characterization of co mplex- balanced, detailed-balanced, and weakly reversible systems. SIAM J. Appl. Math., 80(1):183–205, 2020

2020
[28]

Craciun, A

G. Craciun, A. Deshpande, and J. Jin. On the connectivity of th e disguised toric locus of a reaction network. J. Math. Chem. , 62(2):386–405, 2024

2024
[29]

Craciun, A

G. Craciun, A. Deshpande, and J. Jin. A lower bound on the dimen sion of the-disguised toric locus of a reaction network. SIAM Journal on Applied Algebra and Geometry , 8(4):877–912, 2024

2024
[30]

Craciun, A

G. Craciun, A. Deshpande, and J. Jin. The dimension of the disgu ised toric locus of a reaction network. arXiv preprint arXiv:2412.02620 , 2024

2024 arXiv
[31]

Kothari and A

S. Kothari and A. Deshpande. Endotactic and strongly endot actic networks with inﬁnitely many positive steady states. J. Math. Chem. , pages 1–25, 2024

2024
[32]

Kothari, J

S. Kothari, J. Jin, and A. Deshpande. Realizations through wea kly reversible networks and the globally attracting locus. arXiv preprint arXiv:2409.04802, 2024

2024 arXiv
[33]

Craciun, J

G. Craciun, J. Jin, and P. Yu. An algorithm for weakly reversible d eﬁ- ciency zero realizations of polynomial dynamical systems. arXiv preprint arXiv:2205.14267, 2022. 19

2022 arXiv
[34]

G. Ziegler. Lectures on polytopes, volume 152. Springer Science & Business Media, 2012

2012
[35]

Rockafellar

R. Rockafellar. Convex analysis , volume 18. Princeton university press, 1970

1970
[36]

Introduction to Toric Varieties

W Fulton. Introduction to Toric Varieties . Princeton University Press, Princeton, New Jersey, 1993

1993
[37]

Boyd and L

S. Boyd and L. Vandenberghe. Convex optimization . Cambridge university press, 2004

2004
[38]

Craciun, A

C. Craciun, A. Deshpande, and J. Yeon. Quasi-toric diﬀerentia l inclusions. Discrete and Continuous Dynamical Systems - B , 26(5):2343–2359, 2021

2021
[39]

G. Craciun. Toric diﬀerential inclusions and a proof of the global attractor conjecture. arXiv preprint arXiv:1501.02860 , 2015

2015 arXiv
[40]

Craciun, J

G. Craciun, J. Jin, and P. Yu. Uniqueness of weakly reversible an d deﬁ- ciency zero realizations of dynamical systems. Math. Biosci. , 342:108720, 2021

2021
[41]

Deshpande

A. Deshpande. Source-only realizations, weakly reversible deﬁ ciency one networks, and dynamical equivalence. SIAM J. Appl. Dyn. Sys. , 22(2):1502–1521, 2023

2023
[42]

Feinberg

M. Feinberg. Lectures on chemical reaction networks. Notes of lec- tures given at the Mathematics Research Center, University of Wisconsin , page 49, 1979

1979
[43]

Gunawardena

J. Gunawardena. Chemical reaction network theory for in-silic o biologists. Notes available for download at http://vcp. med. harvard. e du/papers/crnt. pdf, 2003

2003
[44]

Guldberg and P

C. Guldberg and P. Waage. Studies Concerning Aﬃnity. CM Forhan- dlinger: Videnskabs-Selskabet I Christiana , 35(1864):1864, 1864
[45]

E. Voit, H. Martens, and S. Omholt. 150 years of the mass actio n law. PLOS Comput. Biol. , 11(1):e1004012, 2015

2015
[46]

Yu and G

P. Yu and G. Craciun. Mathematical Analysis of Chemical Reactio n Sys- tems. Isr. J. Chem. , 58(6-7):733–741, 2018. 20

2018
[47]

Adleman, M

L. Adleman, M. Gopalkrishnan, M. Huang, P. Moisset, and D. Reis hus. On the mathematics of the law of mass action. In A Systems Theoretic Approach to Systems and Synthetic Biology I: Models and Syst em Charac- terizations, pages 3–46. Springer, 2014

2014
[48]

Hars and J

V. Hars and J. T´ oth. On the inverse problem of reaction kinetics. volume 30, 01 1979

1979
[49]

Conradi, D

C. Conradi, D. Flockerzi, J. Raisch, and J. Stelling. Subnetwork analysis reveals dynamic features of complex (bio) chemical networks. Proceedings of the National Academy of Sciences , 104(49):19175–19180, 2007

2007
[50]

Craciun and C

G. Craciun and C. Pantea. Identiﬁability of chemical reaction ne tworks. J. Math. Chem. , 44(1):244–259, 2008

2008
[51]

Craciun, J

G. Craciun, J. Jin, and P. Yu. An algorithm for ﬁnding weakly reve rsible deﬁciency zero realizations of polynomial dynamical systems. SIAM J. Appl. Math. , 83(4):1717–1737, 2023

2023
[52]

Eaton, David Bateman, Søren Hauberg, and Rik Wehbrin g

John W. Eaton, David Bateman, Søren Hauberg, and Rik Wehbrin g. GNU Octave version 6.4.0 manual: a high-level interactive lang uage for numer- ical computations, 2021. 21

2021

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.

[1] [1]

Craciun and M

G. Craciun and M. Feinberg. Multiple equilibria in complex chemical rea c- tion networks: I. the injectivity property. SIAM J. Appl. Math. , 65(5):1526– 1546, 2005

work page 2005

[2] [2]

Craciun and M

G. Craciun and M. Feinberg. Multiple equilibria in complex chemical re- action networks: II. the species-reaction graph. SIAM J. Appl. Math. , 66(4):1321–1338, 2006

work page 2006

[3] [3]

Wilhelm and R

T. Wilhelm and R. Heinrich. Smallest chemical reaction system with ho pf bifurcation. J. Math. Chem. , 17(1):1–14, 1995

work page 1995

[4] [4]

Wilhelm and R

T. Wilhelm and R. Heinrich. Mathematical analysis of the smallest che mical reaction system with hopf bifurcation. J. Math. Chem. , 19(2):111–130, 1996

work page 1996

[5] [5]

´Erdi and J

P. ´Erdi and J. T´ oth.Mathematical models of chemical reactions: theory and applications of deterministic and stochastic models . Manchester University Press, 1989

work page 1989

[6] [6]

Deshpande and M

A. Deshpande and M. Gopalkrishnan. Autocatalysis in reaction ne tworks. Bull. Math. Biol. , 76(10):2570–2595, 2014

work page 2014

[7] [7]

Craciun and A

G. Craciun and A. Deshpande. Homeostasis and injectivity: a rea ction network perspective. J. Math. Biol. , 85(6):67, 2022

work page 2022

[8] [8]

Craciun, A

G. Craciun, A. Deshpande, B. Joshi, and P. Yu. Autocatalytic re combina- tion systems: A reaction network perspective. Math. Biosci. , 345:108784, 2022. 17

work page 2022