arXiv reader

Absolute Concentration Robustness in Rank-One Kinetic Systems

Eduardo R. Mendoza, Dylan Antonio SJ. Talabis, Editha C. Jose and Lauro L. Fontanil

arXiv
2304.03611v1 · PDF · 7 April 2023
Categories
math.DS
Licence
CC BY 4.0 The site shows the text.

Abstract

A kinetic system has an absolute concentration robustness (ACR) for a molecular species if its concentration remains the same in every positive steady state of the system. Just recently, a condition that sufficiently guarantees the existence of an ACR in a rank-one mass-action kinetic system was found. In this paper, it will be shown that this ACR criterion does not extend in general to power-law kinetic systems. Moreover, we also discussed in this paper a necessary condition for ACR in multistationary rank-one kinetic system which can be used in ACR analysis. Finally, a concept of equilibria variation for kinetic systems which are based on the number of the system’s ACR species will be introduced here.

Items 39

Definition

Definition 2.1.1

OYC9

A chemical reaction network is a triple \mathscr N = (\mathscr S, \mathscr C, \mathscr R) of three non-empty finite sets:

  1. a set species \mathscr S;

  2. a set \mathscr C of complexes, which are nonnegative integer linear combinations of the species; and

  3. a set \mathscr R \subseteq \mathscr C \times \mathscr C of reactions such that

    • (y,y)\notin \mathscr R for all y\in \mathscr C, and

    • for each y\in \mathscr C, there exists a y'\in \mathscr C such that (y,y')\in \mathscr R or (y',y)\in \mathscr R.

Used by Definition 3.1.1, Example 3.1.1, Theorem 3.1.1, Example 3.2.1, Proof, Theorem 4.1.1, Proof, Corollary 4.1.1, Corollary 4.1.2, Definition 5.1.1, Proposition 5.1.1, Proof, Example 5.1.1, Definition 5.1.2, Proposition 5.3.1, Proof
Definition

Definition 2.2.1

P24U

A kinetics K: \mathbb R^{\mathscr S}_{>0} \rightarrow \mathbb R^{\mathscr R} is a power law kinetics if

K_i(x)=k_ix^{F_{i,\cdot}} for i=1,\dots, r

where k_i\in \mathbb R_{>0}, F_{i,j} \in \mathbb R, and F_{i,\cdot} is the row of F associated to reaction R_i.

Used by Example 3.2.1
Definition

Definition 2.2.2

NF0X

A PLK system has reactant-determined kinetics (of type PL-RDK) if for any two branching reactions R_i, R_j\in \mathscr R, the corresponding rows of kinetic orders in F are identical, i.e., F_{ih}=F_{jh} for h=1, \dots,m. Otherwise, a PLK system has non-reactant-determined kinetics (of type PL-NDK).

Definition

Definition 2.2.3

DH9Z

The species formation rate function of a chemical kinetic system is the vector field

f(c)=NK(c)=\displaystyle \sum_{y_i\rightarrow y'_i\in \mathscr R}K_i(c)(y'_i-y_i), where c\in \mathbb R^{\mathscr S}_{\geq 0},

where N is the m \times r matrix, called stoichiometric matrix, whose columns are the reaction vectors of the system. The equation dc/dt=f(c(t)) is the ODE or dynamical system of the chemical kinetic system. An element c^* of \mathbb R^{\mathscr S}_{>0} such that f(c^*)=0 is called a positive equilibrium or steady state of the system. We use E_+(\mathscr N,K) to denote the set of all positive equilibria of a CKS.

Used by Definition 2.2.4, Theorem 2.2.1, Example 5.1.1
Definition

Definition 2.2.4

U9SR

The complex formation rate function g: \mathbb R^{\mathscr S}_{>0}\rightarrow \mathbb R^{\mathscr C} of a chemical kinetic system is the vector field

g(c)=I_aK(c)=\displaystyle \sum_{y_i\rightarrow y'_i\in \mathscr R}K_i(c)(\omega_{y'_i}-\omega_{y_i}), where c\in \mathbb R^{\mathscr S}_{\geq 0}.

where I_a is the incidence map. A CKS is complex balanced if it has complex balanced steady state, i.e., there is a composition c^{**}\in \mathbb R_{>0}^{\mathscr S} such that g(c^{**} )=0. We denote by Z_+(\mathscr N,K) the set of all complex balanced steady states of the system.

Uses Definition 2.2.3 · Used by Theorem 2.2.1
Definition

Definition 3.1.1

QWZI

Let \mathscr N=\left(\{A\}, \mathscr C, \mathscr R \right) be a one-species network with |\mathscr R|\geq 1. Let every reaction of \mathscr N be of the form aA\rightarrow bA, where a,b\geq 0 and a\neq b. Suppose \mathscr N has m distinct reactant complexes with a_1<a_2<\dots<a_m as their stoichiometric species. The arrow diagram of \mathscr N, denoted by \rho=(\rho_1,\rho_2,\dots,\rho_m), is the element of \{\rightarrow,\leftarrow, \longleftrightarrow\}^m given by:

\rho_i = \left\{ \begin{array}{clrc@{\qquad}l} \rightarrow & \textnormal{if for all reactions\;} a_iA \rightarrow bA \textnormal{\;in\;} G, \textnormal{\;it is the case that\;} b > a_i \\ \leftarrow & \textnormal{if for all reactions\;} a_iA \rightarrow bA \textnormal{\;in\;} G, \textnormal{\;it is the case that\;} b < a_i\\ \longleftrightarrow & \textnormal{otherwise} \end{array} \right.
Uses Definition 2.1.1 · Used by Example 3.1.1, Theorem 3.1.1, Example 3.2.1
Example

Example 3.1.1

RP3S

Consider the network determined by \{B\rightarrow A, 2A+B \rightarrow A+2B\}. After removing species A, the embedded network has the following reaction set \{0\leftarrow B \rightarrow 2B\} which has (\longleftrightarrow) as arrow diagram. On the other hand, \{0\rightarrow A, A\rightarrow 2A\} is the embedded network obtained after removing B with arrow diagram (\rightarrow,\leftarrow).

Uses Definition 2.1.1, Definition 3.1.1
Theorem

Theorem 3.1.1

75NU

Let \mathscr N be a rank-one network with species A_1, A_2, \dots, A_m. Then, the following are equivalent:

  1. \mathscr N has stable ACR and admits a positive state.

  2. There is a species A_{i^*} such that the following holds:

    1. for the embedded network of \mathscr N obtained by removing all species except A_{i^*}, the arrow diagram has one of the following forms:

      \begin{array}{rl} (\longleftrightarrow,\leftarrow,\leftarrow,\dots,\leftarrow), & (\rightarrow, \rightarrow, \dots, \rightarrow,\longleftrightarrow) \\ (\rightarrow, \rightarrow, \dots, \rightarrow,\longleftrightarrow, \leftarrow,\leftarrow,\dots,\leftarrow), & (\rightarrow, \rightarrow, \dots, \rightarrow, \leftarrow,\leftarrow,\dots,\leftarrow) \end{array}

    2. the reactant complexes of \mathscr N differ only in species A_{i^*} (i.e., if y and \hat{y} are both reactant complexes, then y_i=\hat{y_i} for all i\neq i^*).

Uses Definition 2.1.1, Definition 3.1.1
Example

Example 3.2.1

JPIH

Consider \mathscr N=(\{A,B\},\mathscr C, \mathscr R), where \mathscr R consists of the following reactions and rate constants:

\begin{array}{cl} R_1:& B {\xrightarrow[r_1]{}} A \\ R_2:& 2A+B {\xrightarrow[r_2]{}} 3A\\ R_3:& 3A+B {\xrightarrow[r_3]{}} 2A+2B\\ R_4:& 4A+B {\xrightarrow[r_4]{}} 3A+2B\\ \end{array}

This is a rank-one network with S=\textnormal{span\;}\{A-B\} as its stoichiometric subspace. Suppose that it is endowed with power law kinetics and has the following kinetic order matrix:

F=\begin{blockarray}{ccc} A & B \\ \begin{block}{[cc]c} p_1 & q & R_1 \\ p_2 & q & R_2 \\ p_3 & q & R_3 \\ p_4 & q & R_4\\ \end{block} \end{blockarray}

where p_i’s are integers. We obtained the following embedded network after removing B.

\begin{array}{c} 0\rightarrow A\\ 2A\leftrightarrow 3A\leftarrow 4A \end{array}

With the four distinct reactant complexes in this network, here is the corresponding arrow diagram (\rightarrow, \rightarrow, \leftarrow, \leftarrow). Now, observe that (\mathscr N, K) has the following ODEs:

\begin{array}{cc} \dot{A}= & {r_1}A^{p_1}B^q + {r_2}A^{p_2}B^q - {r_3}A^{p_3}B^q - {r_4}A^{p_4}B^q \\ \dot{B}= & -{r_1}A^{p_1}B^q - {r_2}A^{p_2}B^q + {r_3}A^{p_3}B^q + {r_4}A^{p_4}B^q \end{array}

Solving for the positive equilibria, one of these equations is equated to zero and get

\begin{array}{cc} B^q({r_1}A^{p_1} + {r_2}A^{p_2} - {r_3}A^{p_3} - {r_4}A^{p_4} & = 0\\ \Leftrightarrow {r_1}A^{p_1} + {r_2}A^{p_2} - {r_3}A^{p_3} - {r_4}A^{p_4} & = 0 \end{array}

Observe that when p_1<p_2<p_3<p_4, the polynomial

{r_1}A^{p_1} + {r_2}A^{p_2} - {r_3}A^{p_3} - {r_4}A^{p_4}

has exactly one positive root by the Descarte’s Rule of Signs. This means that the system has ACR in A. On the other hand, suppose that

\begin{bmatrix} p_1 \\ p_2 \\ p_3 \\ p_4 \end{bmatrix} = \begin{bmatrix} 0 \\ 3 \\ 1 \\ 2 \end{bmatrix} \; and \begin{bmatrix} r_1 \\ r_2 \\ r_3 \\ r_4 \end{bmatrix} = \begin{bmatrix} 4 \\ 1 \\ 2 \\ 3 \end{bmatrix} \;

Then, the polynomial in (?) becomes

4+A^3-2A-3A^2=4-2A-3A^2+A^3

This polynomial has two positive roots namely, 1 and 1 + \sqrt{5}. This means that the system does not have an ACR in A in this case.

Uses Definition 2.2.1, Definition 2.1.1, Definition 3.1.1
Definition

Definition 3.3.1

UHD5

Two kinetics K, K' in \mathscr{K}_\Omega (\mathscr{N}) are positive function factor equivalent (PPF-equivalent) if for all x\in \mathbb{R}^{\mathscr{S}}_{>} and every reaction q, \dfrac{K_q(x)}{K'_q(x)} is a positive function of x only, i.e. independent of q.

Used by Proof
Definition

Definition 3.3.2

EB6L

A homogeneous PL quotient of a MAK system (\mathscr{N}, K) is a PLK system (\mathscr{N}, K_{PLK}) with the same rate constants and a kinetic order matrix F_{PLK} = F_{MAK} - F_\beta.

Used by Proof
Proposition

Proposition 3.3.2

S9H4

The stable ACR criterion holds for homogeneous PL-quotients of rank-one mass action systems.

Proof of Proposition 3.3.2

Let (\mathscr{N}, K_{PLK}) be a homogeneous PL quotient of a rank-one MAK system (\mathscr{N}, K) where (\mathscr{N}, K) satisfies the stable ACR criterion. For each species, write X_i^{\alpha_i} = X_i^{\beta_i}(X_i^{\alpha_i - \beta_i}). For a reaction q, each K_q(x) = k_q \prod X_i^{\beta_i} \prod (X_i^{\alpha_i - \beta_i}). Note that \prod X_i^{\beta_i} is independent of the reaction q and hence, from Definition 3.3.1, K_{PLK} and K are PFF-equivalent. From Proposition 3.3.1, E_+(\mathscr{N}, K) = E_+(\mathscr{N}, K_{PLK}). Thus, their ACR coincide and, hence the stable ACR criterion also holds for (\mathscr{N}, K_{PLK}).

Uses Definition 3.3.1, Proposition 3.3.1, Definition 3.3.2, Definition 2.1.1
Theorem

Theorem 4.1.1

SWU3

If \mathscr N has rank one and \left (\mathscr N, K\right ) is multistationary, then (the line) S lies in the species hyperplane of every ACR species X. In other words, for any basis vector v of S, its X-coordinate is 0.

Uses Definition 2.1.1 · Used by Proof
Proof of Theorem 4.1.1

Since \left (\mathscr N, K\right ) is multistationary, there is a stoichiometric class that contains two distinct positive equilibria x_1 and x_2. In other words, x_1 - x_2 \in S and x_1 - x_2 \neq 0. For any ACR species X of \left (\mathscr N, K\right ), it follows that the X-th coordinate of x_1 - x_2 is 0. Since x_1 - x_2 \neq 0, it is a basis vector for the one-dimensional subspace S of the rank-one system, proving the claim.

Uses Definition 2.1.1
Corollary

Corollary 4.1.1

6N9Q

Let m_{\textnormal{ACR}} be the number of \textnormal{ACR} species of a rank-one multistationary system \left (\mathscr N, K\right ). Then m_{\textnormal{ACR}} \leq m - |\textnormal{supp\;}v|, where \textnormal{supp\;}v is the support of any basis vector of S.

Uses Definition 2.1.1
Corollary

Corollary 4.1.2

7GJW

If a rank-one and co-conservative network is multistationary, then it has no \textnormal{ACR} species.

Uses Definition 4.1.1, Definition 2.1.1
Proof of Corollary 4.1.2

Any non-zero vector v in S has only positive or negative coordinates, i.e., |\textnormal{supp\;}v|=m. Hence, m_{\textnormal{ACR}}=0.

Example

Example 4.2.1 (Class 1-alt^c)

UI99

The rank-one network below was the focus of Example 4.2 in [voitiuk]. It was shown in that paper that it has the capacity for multiple positive and non-degenerate equilibria. Moreover, the network has stoichiometric subspace generated only by the vector v=(1,1,1,0,-1,-1).

\begin{array}{cl} R_1:& B+2C+2E \rightarrow A+2B+3C+E \\ R_2:& 2A+2B+C+2D+E \rightarrow A+B+2D+2E\\ R_3:& A+3C+D+2E \rightarrow 2A+B+4C+D+E\\ R_4:& 3A+3B+C+E \rightarrow 2A+2B+2E\\ \end{array}
Example

Example 4.2.2 (Class 2-alt)

7YOJ

The rank-one network below is a 2-alternating network with stoichiometric subspace generated by the vector v=(2,1,0). From Corollary 4.3 in [voitiuk], we conclude that it has the capacity for multiple positive and non-degenerate equilibria.

\begin{array}{cl} R_1:& 2X + 2Y \rightarrow Y \\ R_2:& 3X + Y \rightarrow 5X + 2Y\\ R_3:& 4X + 2Y + Z \rightarrow Z \\ R_4:& 4X + 2Y + 2Z \rightarrow 2X + Y + 2Z\\ \end{array}
Definition

Definition 5.1.1

GYCX

The (positive) equilibria variation of a kinetic system (\mathscr N, K) is the number of non-ACR species divided by the number of species, i.e.,

v_+(\mathscr N, K) = \dfrac{m-m_{ACR}}{m}.
Uses Definition 2.1.1
Proposition

Proposition 5.1.1

YCLC

Let E_+(\mathscr N, K) be a kinetic system. Then,

  1. v_+(\mathscr N, K) = 0 \Leftrightarrow |E_+(\mathscr N, K)| = 1 and

  2. v_+(\mathscr N, K) = 1 \Leftrightarrow (\mathscr N, K) has no ACR species

Uses Definition 2.1.1
Corollary

Corollary 5.1.1

8WTN

Let (\mathscr N, K) be a kinetic system.

  1. If (\mathscr N, K) is multistationary, then v_+(\mathscr N, K) \geq \dfrac{1}{m}. In particular, if v_+(\mathscr N, K) = 0, then (\mathscr N, K) is monostationary.

  2. If \mathscr N is open and v_+(\mathscr N, K) > 0, then (\mathscr N, K) is multistationary.

Proof of Corollary 5.1.1

For (a): Multistationarity implies at least two distinct positive equilibria, hence m_{ACR} \leq m - 1, and the claims follow. For (b): if the network is open, then there is only one stoichiometric class. Hence, multistationarity is equivalent to the occurrence of at least two distinct equilibria in species space.

Uses Definition 2.1.1
Example

Example 5.1.1

BHXG

Schmitz’s model of the earth’s pre-industrial carbon cycle system was analyzed by Fortun et al. in [fort3.5]. Below is its corresponding reaction network.

\begin{tikzcd} M_5 \arrow[dd, "R_1 "'] \arrow[rd, "R_2", shift left] & & M_2 \arrow[ld, "R_5"', shift right] \arrow[rd, "R_{11}"', shift right] \arrow[dd, "R_9"'] & \\ & M_1 \arrow[lu, "R_3", shift left] \arrow[ru, "R_6"', shift right] \arrow[rd, "R_8"', shift right] & & M_4 \arrow[lu, "R_{10}"', shift right] \arrow[ld, "R_{12}"', shift right] \\ M_6 \arrow[ru, "R_4"'] & & M_3 \arrow[lu, "R_7"', shift right] \arrow[ru, "R_{13}"', shift right] & \end{tikzcd}

In this network, M_1, M_2, M_3, M_4, M_5, and M_6 stand for atmosphere, warm ocean surface waters, cool ocean surface waters, deep ocean waters, terrestrial biota, and soil and detritus, respectively. Important numbers of the network as well as its kinetic order matrix are given in the following.

The paper has a figure here, which this page does not show.

Fortun and Mendoza [fort3.5] showed that the system has no ACR species, i.e., v_+(\mathscr N, K) = 1 > \dfrac{1}{6}. Since the system has a conservative, concordant, and weakly reversible network and weakly monotonic kinetics, by Theorem 6.6 of [shinar], it has a unique positive equilibrium in each stoichiometric class and consequently monostationary. Note also that the network has rank 5 < 6, hence it is a closed network.

Uses Definition 2.1.1, Definition 2.2.3
Proposition

Proposition 5.1.2

5CCX

Let (\mathscr N, K) be a multistationary rank-one kinetic system and v a basis vector of S. Then, v_+(\mathscr N, K) \geq \dfrac{|supp(v)|}{m}.

Proof of Proposition 5.1.2

According to Theorem 4.1.1, m_{ACR} \leq m - |supp(v)| \Leftrightarrow |supp(v)| \leq m - m_{ACR}, leading to the new lower bound. Note that since v is a basis vector, |supp(v)|\geq 1, confirming the improvement.

Uses Theorem 4.1.1
Definition

Definition 5.1.2

FJY1

For any species X, the (m-1)-dimensional subspace

H_X:=\{x\in \mathbb R^{\mathscr S}| x_X=0\}

is called the species hyperplane of X.

Uses Definition 2.1.1
Definition

Definition 5.1.3

X8GR

For a subset Y of E_+(\mathscr N, K), the set

\Delta_{\phi}Y:=\{\phi(x)-\phi(x')| x,x'\in Y\}

is called the difference set of \phi-transformed equilibria in Y, and its span \langle \Delta_{\phi}Y \rangle the difference space of \phi-transformed equilibria in Y.

Proposition

Proposition 5.1.3

93M6

Let (\mathscr N, K) be a kinetic system. Then, \dfrac {\dim \langle \Delta_{\phi}E_+\rangle}{m} \leq v_+(\mathscr N, K).

Example

Example 5.2.1

N06V

Fortun and collaborators studied variants of the Anderies et al. [anderies] model of the earth’s pre-industrial carbon cycle (see the reaction network below) in [noel] and [fort3.5].

Proposition

Proposition 5.3.2

5ZVM

Let m'=|\mathscr S'| and v_+(\mathscr N', K'), \Tilde{v}_+(\mathscr N', K') be the equilibria variations of the non-embedded and embedded subnetworks. Then,

  1. m-m'_{ACR}=(m'-m'_{ACR})+(m-m') and

  2. 0\leq \Tilde{v}_+(\mathscr N', K')-v_+(\mathscr N', K')\leq \dfrac{m-m'}{m}

Proof of Proposition 5.3.2

(i.) should be read as: the number of non-ACR species in an embedded network is the number of non-ACR species in the non-embedded network plus the number of non-occurring species, and (ii.): dividing the equation in (i.) by m and using the inequality \dfrac{m'-m'_{ACR}}{m}\leq \dfrac{m'-m'_{ACR}}{m'}, we obtain \Tilde{v}_+(\mathscr N', K')-v_+(\mathscr N', K')\leq \dfrac{m-m'}{m}. On the other hand, the left-hand side, after forming a common denominator, is now equal to \dfrac{m'_{ACR}(m-m')}{mm'}\geq 0, since all factors are non-negative.

Uses Definition 2.1.1
Proposition

Proposition 5.3.3

KFVB

v_+(\mathscr N, K)\leq \Tilde{v}_+(\mathscr N', K') for any independent subnetwork \mathscr N'.

Proof of Proposition 5.3.3

It follows that m-m_{ACR}\leq m-m'_{ACR} and dividing both sides with m results in the claim.

Source and licence

The text on this page is from arXiv:2304.03611v1 by Eduardo R. Mendoza, Dylan Antonio SJ. Talabis, Editha C. Jose and Lauro L. Fontanil, under the licence CC BY 4.0. The corpus made these changes: it read the items out of the LaTeX source, it expanded the macros of the paper, it put each proof after its statement, and it converted the text to HTML. The formulas are the TeX of the paper, and KaTeX renders them. A formula or a reference can look different from the PDF. The page is published under CC BY 4.0.

The record of each item is in analysis/objects/2304/2304.03611.jsonl.