1: \documentclass[showpacs,preprintnumbers,amsmath,amssymb,twocolumn]{revtex4}

2: \usepackage{graphicx}

3: \usepackage{makeidx}

4: \usepackage{chemarr}

5: \usepackage{bm}\newcommand{\C}{{\cal{C}}}

6: \newcommand{\Cp}{{\cal{C^{\prime}}}}

7: \def\prl#1#2#3{{ Phys. Rev. Lett.} {\bf #1}, #2 (#3)}

8: \def\prep#1#2#3{{ Phys. Rep.} {\bf #1}, #2 (#3)}

9: \def\pla#1#2#3{Phys. Lett. A {\bf #1}, #2 (#3)}

10: \def\ijbc#1#2#3{Int J. Bifurcation and Chaos {\bf #1}, #2 (#3)}

11: \def\pre#1#2#3{Phys. Rev. E {\bf #1}, #2 (#3)}

12: \def\prb#1#2#3{Phys. Rev. B {\bf #1}, #2 (#3)}

13: \def\pra#1#2#3{Phys. Rev. A {\bf #1}, #2 (#3)}

14: \def\pnas#1#2#3{Proc. Natl. Acad. Sci. (USA) {\bf #1}, #2 (#3)}

15: \def\lanl{{\it LANL archives}}

16: \def\pr#1#2#3{Phys. Rev. {\bf #1}, #2 (#3)}

17: \def\jpa#1#2#3{J. Phys. A {\bf #1}, #2 (#3)}

18: \def\jcp#1#2#3{J. Chem. Phys. {\bf #1}, #2 (#3)}

19: \def\jpc#1#2#3{J. Phys. Chem. {\bf #1}, #2 (#3)}

20: \def\jsp#1#2#3{J. Stat. Phys. {\bf #1}, #2 (#3)}

21: \def\jmb#1#2#3{J. Math. Biol. {\bf #1}, #2 (#3)}

22: \def\cmp#1#2#3{Comm. Math. Phys. {\bf #1}, #2 (#3)}

23: \def\physd#1#2#3{Physica D {\bf #1}, #2 (#3)}

24: \def\physa#1#2#3{Physica A {\bf #1}, #2 (#3)}

25: \def\pram#1#2#3{Pramana J. Phys. {\bf #1}, #2 (#3)}

26: \def\prtp#1#2#3{Prog. Theor. Phy. {\bf #1}, #2 (#3)}

27: \def\epj#1#2#3{Eur. Phys. J. {\bf #1}, #2 (#3)}

28: \def\jct#1#2#3{J. Combin. Theor. {\bf #1}, #2 (#3)}

29: \def\ap#1#2#3{Adv. Phys. {\bf #1}, #2 (#3)}

30: \def\csf#1#2#3{Chaos, Solitons \& Fractals {\bf #1}, #2 (#3)}

31: \def\tree#1#2#3{Trends Ecol. Evol. {\bf #1}, #2 (#3)}

32: \def\ch#1#2#3{Chaos {\bf #1}, #2 (#3)}

33: \def\nat#1#2#3{Nature {\bf #1}, #2 (#3)}

34: \def\mol{\mbox{mol}}

35: \def\hr{\mbox{h}}

36: \def\p{^{\prime}}

37: \def\h{\mbox{h}}

38: \def\DA{D_{A}}

39: \def\DAP{D_{A}^{\prime}}

40: \def\DR{D_{R}}

41: \def\DRP{D_{R}^{\prime}}

42: \def\DS{D_{S}}

43: \def\DSP{D_{S}^{\prime}}

44: \def\MA{M_{A}}

45: \def\MR{M_{R}}

46: \def\MS{M_{S}}

47: \def\ta{\theta_{A}}

48: \def\tr{\theta_{R}}

49: \def\ts{\theta_{S}}

50: \def\ga{\gamma_{A}}

51: \def\gr{\gamma_{R}}

52: \def\gs{\gamma_{S}}

53: \def\aa{\alpha_{A}}

54: \def\apa{\alpha_{A}^{\prime}}

55: \def\ar{\alpha_{R}}

56: \def\apr{\alpha_{R}^{\prime}}

57: \def\as{\alpha_{S}}

58: \def\aps{\alpha_{S}^{\prime}}

59: \def\ba{\beta_{A}}

60: \def\br{\beta_{R}}

61: \def\bs{\beta_{S}}

62: \def\dma{\delta_{MA}}

63: \def\dmr{\delta_{MR}}

64: \def\dms{\delta_{MS}}

65: \def\da{\delta_{A}}

66: \def\dr{\delta_{R}}

67: \def\ds{\delta_{S}}

68: \def\gc{\gamma_{C}}

69: \def\gcp{\gamma_{C}^{\prime}}

70: 

71: \def\etc{ etc.}

72: \def\epsilon{\varepsilon}

73: \def\yin{y_{\mbox{in}}}

74: \def\sgn{\mbox{sgn}}

75: \def\half{\frac{1}{2}}

76: \def\ie{i.e.}

77: \def\l{\ell}

78: \def\x{{\bf{x}}}

79: \def\f{{\bf{f}}}

80: \def\i{\indent}

81: \def\noi{\noident}

82: \def\beqr{\begin{eqnarray}}

83: \def\eqnr{\end{eqnarray}}

84: \def\beq{\begin{equation}}

85: \def\bc{\begin{center}}

86: \def\ec{\end{center}}

87: \def\eqn{\end{equation}}

88: \topmargin=-1.5cm

89: \textheight=23cm

90: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

91: \begin{document}

92: \title{The synchronization of stochastic oscillators}

93: \author{Amitabha Nandi$^1$, Santhosh G.$^1$, R. K. Brojen Singh$^2$,

94: and Ram Ramaswamy$^{1,2}$}

95: \affiliation{$^1$School of Physical Sciences, Jawaharlal Nehru University,

96: New Delhi 110067, India\\

97: $^2$Center for Computational Biology and Bioinformatics, School of

98: Information Technology,

99: Jawaharlal Nehru University, New Delhi 110067, India}

100: \date{\today}

101: \begin{abstract}

102: We examine microscopic mechanisms for coupling stochastic oscillators so

103: that they display similar and correlated temporal variations. Unlike 

104: oscillatory motion in deterministic dynamical systems, complete 

105: synchronization of stochastic oscillators does not occur, but appropriately 

106: defined oscillator phase variables coincide. This is illustrated in model 

107: chemical systems and genetic  networks that produce oscillations in the 

108: dynamical variables, and we show that  suitable coupling of different 

109: networks can result in their {\it phase} synchronization.

110: \end{abstract}

111: \pacs{02.50.Fz, 05.40.-a, 05.45.Xt, 82.20.Fd, 87.19.Jj}

112: \keywords{Synchronization, stochasticity}

113: \maketitle

114: The {concerted} behaviour  of different stochastic processes can be of

115: considerable importance in a variety of situations. Examples from a

116: biological context \cite{glass} range from the synchronized firing of groups 

117: of neurons \cite{neuron} to the occurence of circadian or ultradian 

118: cycles \cite{church,tu} in groups of metabolic and cellular regulatory 

119: processes, each of which is individually stochastic. Similar phenomena also 

120: occur in other areas of study, such as weather modeling or the study of 

121: coupled populations \cite{lloyd}. It is therefore a moot question whether 

122: there is a sense in which two or more independent (or unrelated) stochastic 

123: phenomena can become temporally synchronized. In this Letter we consider 

124: mechanisms through which such synchronization can be effected, and examine 

125: measures through which the synchronization can be detected.

126: 

127: The synchronization of coupled nonlinear oscillators has been studied 

128: extensively in the past two decades. Both chaotic and nonchaotic oscillator 

129: systems can show complete synchronization, when all variables of the systems 

130: coincide, as well as more general forms such as phase, generalized, or lag 

131: synchronization \cite{kurths}. Studies have examined a variety of 

132: different coupling schemes and topologies, and the robustness of 

133: synchronization to noise. In the typical cases examined, the noise 

134: is external and thus appears largely as additional stochastic terms in 

135: otherwise deterministic dynamical equations \cite{a,hanggi}.

136: 

137: For the systems we study here the evolution is {\underline {intrinsically}} 

138: stochastic. Such systems do not follow deterministic equations of motion: 

139: a master equation  describes the evolution of configurational 

140: probabilities \cite{master},

141: \beq

142: \label{me}

143: \frac{d}{dt}P(\C,t)  = -\sum_{\Cp} P(\C,t) W_{\C \to\Cp} + \sum_{\Cp}  P(\Cp,t) W_{\Cp \to\C}

144: \eqn

145: where in standard notation \cite{osw}, $P(\C,t)$ is the probability of 

146: configuration $\C$ and the $W$'s are transition probabilities. The 

147: configurations that are realized as a function of time give a probabilistic 

148: description of the system as it traverses the state space of the problem.

149: 

150: Consider now two such independent systems, each of which can be described by 

151: a (reduced) master equation of the above form.  Starting with similar 

152: configurations, $\C$, the subsequent evolution of each of the subsystems 

153: will typically be quite distinct. The concern here is to (a) examine 

154: conditions under which the two subsystems synchronize, namely follow very 

155: similar paths in the state space, and to (b) detect this phenomenon.

156: 

157: Our main result can be stated in general terms as follows. When the two 

158: systems are coupled by employing a mediating process, then variables of the 

159: two subsystems {\it phase--synchronize}, namely they vary  in unison. Such 

160: synchronization depends on the nature and strength of the coupling and can 

161: persist even when fluctuations are large.

162: 

163: It is simplest to illustrate this through an example. The stochastic dynamics 

164: of the Brusselator model, which has been studied extensively \cite{brussel} 

165: derives from consideration of the following ``chemical'' reactions 

166: \cite{gillespie}

167: 

168: \beqr

169: \label{cb}

170: A_1 &\xrightarrow{c_1}& X \\

171: A_2 + X &\xrightarrow{c_2}& Y + A_3 \\

172: 2X + Y &\xrightarrow{2c_3}& 3X \\

173: X &\xrightarrow{c_4}& A_4.

174: \eqnr

175: 

176: A master equation formalism is exact for this system in the gas phase and in 

177: thermal equilibrium \cite{megillespie}, and indeed is necessary when 

178: the number of molecules is small. On the other hand, in 

179: the thermodynamic limit, when fluctuations are negligible, one can derive a 

180: set of deterministic reaction--rate equations  for the concentrations of 

181: species $X$ and $Y$,

182: \beqr

183: \label{deter1}

184: \dot x &=& c_1 - c_2 x + c_3  x^2y -c_4 x \equiv f_x(x,y)\\

185: \dot y &=& c_2 x -  c_3  x^2y \equiv f_y(x,y).

186: \label{deter2}

187: \eqnr

188: These rate equations usually have to be integrated numerically to determine

189: the orbits (for the parameters here, a limit cycle). Similarly, the 

190: corresponding master equation cannot be solved analytically and recourse must be made to  Monte Carlo simulations

191: \cite{gillespie}. The number of molecules of species $X$ or $Y$

192: (taking the volume to be unity) will vary in a stochastic manner, giving a

193: (noisy)  limit--cycle solution to Eqs.~(2-5).

194: 

195: Another Brusselator system (denoted by similar equations, but with primed variables

196: and parameters, say) would also naturally give a noisy limit cycle solution,

197: the characteristics of which would depend on the parameters of the problem.

198: Thus if the parameters $\{c_i\}$ and $\{c^{\prime}_i\}$ are different

199: the evolution of the two subsystems

200: will be independent and uncorrelated.

201: 

202: We consider two scenarios where the subsystems are coupled at the microscopic

203: level. In the ``direct'' case, we take the species $X$ and $X^{\prime}$ to be identical:

204: the subsystems share a common drive. We find that for any finite volume this

205: coupling results in the temporal variation of species $Y$ and $Y^{\prime}$ rapidly becoming

206: highly correlated, even if the fluctuations are large.

207: 

208: Alternately one may consider an ``exchange'' process, when species $X$ and

209: $X^{\prime}$ can interconvert. This introduces an additional reaction,

210: \beq

211: X\xrightleftharpoons[c\p]{c} X\p

212: \eqn

213: that serves to couple the subsystems, and depending on the rate of interconversion

214: (governed by $c$ and ${c^{\prime}}$), species $Y$ and $Y^{\prime}$ show

215: synchronization.

216: 

217: Although both direct as well as exchange coupling mechanisms  effect synchronization,

218: they differ in the details of how they act. Fig.~\ref{fig1} shows the

219: variation of species  $Y$ and $Y\p$  as a function of  time,  from stochastic

220: simulations as well as the so--called chemical Langevin approach \cite{lange},

221: which effectively adds stochastic noise to the

222: deterministic dynamics, Eqs.~(\ref{deter1}-\ref{deter2}). Clearly the two

223: concentrations vary in unison. However, due to intrinsic stochastic

224: fluctuations the two solutions do not become identical but only

225: phase--synchronize \cite{rpk} as we now show.

226: 

227: \begin{figure}

228: \scalebox{0.5}{\includegraphics{fig1.eps}}

229: \caption{ (Colour online)

230: Species $Y$ and $Y^{\prime}$ as a function of  time for the coupled

231: Brusselator system. The  parameter $c_2=50$ differs from

232: $c^{\prime}_{2}= c_2+5$. The other parameters are $c_1=c^{\prime}_{1}=5000$,

233: $c_3=c^{\prime}_{3}=0.000025$ and $c_4=c^{\prime}_{4}=5$. (a) Stochastic

234: simulation results for  $Y$ and $Y^{\prime}$ using direct coupling.

235: (b) As in (a), but for the case of exchange coupling, Eq.~(8) with

236: $c=c\p=0.6$. (c) Results obtained by solving the chemical Langevin equation

237: \cite{lange} with  $\gamma=0.01$.}

238: \label{fig1}

239: \end{figure}

240: 

241: To judge the phase synchronization of two stochastic oscillators, it is

242: necessary to first obtain the phase for a single oscillator. The

243: Hilbert phase of a signal $s(t)$ is obtained \cite{rpk} by first computing its Hilbert transform,

244: \beq

245: \pi\bar s(t) = \mbox{P. V.~}\int_{-\infty}^{\infty}\frac{s(\tau)}{t-\tau}d\tau

246: \eqn

247: (P. V. denotes principal value) and defining the instantaneous amplitude $A(t)$ and phase $\phi(t)$ through

248: \beq

249: A(t) e^{i\phi(t)} = s(t)+ i \bar s(t).

250: \eqn

251: 

252: The difference in the Hilbert phases $\phi$ and $\phi^{\prime}$ of the two

253: stochastic oscillators is shown in Fig.~\ref{fig2}. When the subsystems are

254: uncoupled, the phase difference $\vert \phi - \phi^{\prime}\vert$ increases

255: linearly in time on average since the oscillators evolve independent of one

256: another. With coupling the two oscillators

257: phase--synchronize and the phase difference fluctuates around  a constant

258: value.

259: 

260: \begin{figure}

261: \scalebox{0.5}{\includegraphics{fig2.eps}}

262: \caption{(Colour online)

263: Phase difference with parameter mismatch for the Brusselator model for

264: (a) direct coupling of species $X$ and $X^{\prime}$, and

265: (b) exchange coupling, for different strengths.}

266: \label{fig2}

267: \end{figure}

268: 

269: The coupling schemes discussed here can bring about stochastic

270: synchronization in a very general setting. Consider, for example, a model

271: circadian oscillator that has been quantitatively studied in some detail

272: recently \cite{genosc}.  Shown in

273: Fig.~\ref{bioch} is the biochemical network for two such oscillators coupled

274: with a common drive, namely two genetic circuits with a common activator.

275: In effect a single activator binds to two promoter sites

276: for repressor proteins $R$ and $S$, a fairly commonplace situation.

277: Each individual circuit gives stochastic oscillations in the number of

278: repressor molecules. When the two systems are coupled the  stochastic

279: oscillations of the two subsystems rapidly phase--synchronize. The

280: synchronization is robust to parameter variation: in the examples shown in

281: Fig.~\ref{fig4}, the corresponding parameters of the two subsystems differ by

282: as much as 10\%; nevertheless the variables of the two systems oscillate in

283: phase in a stable and sustained manner.

284: 

285: \begin{figure}

286: \scalebox{0.25}{\includegraphics{fig3.eps}}

287: \caption{

288: Biochemical network of the extended circadian oscillator model. $\DA$ and $\DAP$

289: denote the number of activator genes with and without $A$ bound to its

290: promoter respectively, and  $\DR$, $\DRP$ and $\DS$, $\DSP$, refer to

291: the two repressors driven by the common promoter $A$. $\MA$, $\MR$ and $\MS$

292: denote mRNA corresponding to the activator  $A$, and the repressors $R$ and $S$.

293: $C$ and $C^{\prime}$ corresponds to the inactivated complexes formed by $A$ and

294: $R$, and $A$ and $S$ respectively. The constants $\alpha$ and

295: $\alpha^{\prime}$ denote the basal and activated rates of transcription,

296: $\beta$ the rates of translation, $\delta$ the rates of spontaneous

297: degradation, $\gamma$ the rates of binding of $A$ to other components, and

298: $\theta$ denotes the rates of unbinding of $A$ from those components. 

299: See \cite{param}.}

300: \label{bioch}

301: \end{figure}

302: The temporal behaviour of the two repressors and their phase difference for

303: both direct  and exchange coupling are shown in Fig.~\ref{fig4}. In (a) the

304: two systems are initially uncoupled and therefore evolve independently. The

305: coupling is switched on for $t \ge 2000$, leading rapidly to a constant phase

306: difference, indicative of the phase synchronization (Fig. \ref{fig4}(b)). In the case

307: of   exchange coupling the genetic circuit differs somewhat from Fig.~\ref{bioch}: the

308: circuit of Ref. \cite{genosc} is essentially doubled, and there is an

309: additional activator $A^{\prime}$. Through the equivalent of Eq.~(8)  the activators of

310: the two circuits are allowed to interconvert, and the two subsystems

311: synchronize as can be seen in Fig. \ref{fig4}(c-d).

312: 

313: \begin{figure}

314: \scalebox{0.4}{\includegraphics{fig4.eps}}

315: \caption{(Colour online)

316: Temporal behaviour and phase difference of the repressors in the

317: circadian oscillator model. (a) The repressors are initially  uncoupled

318: and the coupling is switched on (vertical arrow) at time $t \ge 2000$.

319: (b) The corresponding phase difference, which

320: becomes constant for  $t \ge 2000$. (c) Time-series of the  repressors

321: oscillating in unison for  diffusive coupling with  $c=c\p= 0.55$. (d) Phase

322: differences for various $c$.}

323: \label{fig4}

324: \end{figure}

325: 

326: Examination of the macroscopic dynamics offers some clues as to how the

327: stochastic oscillators synchronize. While the master equation

328: description is formally exact at the microscopic level, the kinetic equations

329: that describe the  system at a macroscopic level can be derived from 

330: Eq.~(\ref{me}) by first obtaining the generating function representation, 

331: followed by a cumulant expansion \cite{nicolis} to give a set of ordinary 

332: differential equations,

333: \beq

334: \label{nde}

335: \dot \x  = \f (\x,t)+ O(\frac{1}{V})+ O(\frac{1}{V^2})+ \ldots

336: \eqn

337: in the variables $x_i={\langle X_i\rangle}/{V}$,

338: namely the average concentrations,  with ``noise'' corrections that depend on

339: the  system volume $V$. Coupling two such systems as discussed above would give a set of

340: coupled stochastic differential equations, and the circumstances under which these will phase synchronize have been studied to some extent \cite{kurths,neiman}.  

341: The direct coupling, in the deterministic limit \cite{gillespie,cle} leads 

342: to dynamical equations which are similar (but not identical) to those 

343: that derive from the coupling scheme proposed by Pecora and Carroll 

344: \cite{pecora} for the synchronization of chaotic oscillators.

345: Although it is not possible to define Lyapunov exponents for stochastic

346: systems, analysis of the deterministic limit gives indications of what

347: coupling schemes could be effective. Making species

348: $Y$ the common drive between the two Brusselator subsystems does not serve

349: to synchronize them but instead leads to the stochastic analogue of oscillator death.

350: 

351: Similarly, the exchange process, Eq.~(8) results in

352: diffusive coupling \cite{boccaletti} in the kinetic equations for the species $X$

353: and $X^{\prime}$, namely terms of the form $c(x^{\prime}-x)$. The conditions

354: under which this form of the coupling leads to synchronization have also

355: been studied \cite{boccaletti}. As can be seen from the examples

356: presented here, the systems synchronize when the effective coupling exceeds a

357: threshold strength (Figs.~2(b) and 4(d)). The rate of growth of the phase

358: difference vanishes at this threshold as a power \cite{wang}.

359: 

360: Stochastic oscillators appear to effectively synchronize independent of the size of fluctuations,  

361: in a manner similar to chaotic synchronization \cite{pecora}. However,

362: even when the parameters are identical, exact synchronization is not possible

363: and the systems can only phase--synchronize.

364:  

365: The synchronization  of stochastic oscillators thus has similarities to

366: the analogous process in deterministic dynamical systems (both with and

367: without added noise), but also important differences. The mechanisms that we

368: have described here pertain to systems with intrinsic stochasticity, and are

369: therefore in a nonperturbative limit.  

370: 

371: The coupling schemes that we have proposed here can find application in the

372: design and control of synthetic biological networks where synchronous

373: oscillation may be a desirable feature. McMillen et al.

374: \cite{kopell} have shown that intercell signaling via a diffusing molecule

375: can couple genetic oscillators and effect  synchrony.

376: The present results indicate that such phase synchrony can emerge under very

377: general conditions with high levels of ambient noise.  In a related vein, one can

378: speculate that similar mechanisms underlie the synchrony that is so dramatically

379: evident in cellular processes. As recent time--resolved microarray

380: experiments of yeast have revealed, the multitude of variable gene

381: expression patterns classify into a small number of groups, all genes of a

382: given group having very similar temporal variation profiles \cite{tu}. We have

383: observed that the above coupling schemes are effective in synchronizing

384: ensembles of stochastic oscillators \cite{nandi}.   Other physical situations where

385: such mechanisms  may be relevant are in the study of coupled ecosystems or

386: coupled weather systems: individual subsystems show stochastic dynamics,

387: which however have phase synchrony \cite{lloyd}.

388: 

389: Although we have discussed the explicit case of stochastic oscillators,

390: there is reason to believe that similar microscopic intersystem couplings can bring

391: about temporal correlations in more general stochastic systems. Investigations of such

392: phenomena are currently under way \cite{nandi}.

393: 

394: %\noindent {\bf Acknowledgment:}

395: This research is supported by grants from the DBT,

396: India (to RR) and the CSIR, India through the award of Senior Research

397: Fellowships (to AN and SG)  and a Research Associateship (to RKBS). We thank M Bennett, T Gross, and J K Bhattacharjee for helpful correspondence.

398: 

399: \begin{thebibliography}{abc99}

400: 

401: \bibitem{glass} L. Glass, \nat{410}{277}{2001}.

402: 

403: \bibitem{neuron} A. Neiman et al., %, X. Pei, D. Russel, W. Wojtenek, L. Wilkens, F. Moss, H. A. Braun, M. T. Huber, and K. Voigt,

404:  \prl{82}{660}{1999}.

405: 

406: \bibitem{tu} B. P. Tu, A. Kudlicki, M. Rowicka and S. L. McKnight,

407: Science, {\bf 310}, 1152 (2005).

408: 

409: \bibitem{church} See e. g. S. Tavazoie, J. D. Hughes, M. J. Campbell, 

410: R. J. Cho and G. M. Church, Nature Genetics  {\bf 22}, 281 (1999).

411: 

412: \bibitem{lloyd} A. L. Lloyd and R. M. May, \tree{14}{417}{1999}.

413: 

414: \bibitem{kurths} A. Pikovsky, M. Rosenblum and J. Kurths,

415: {\it Synchronization: A universal concept in nonlinear science}, (Cambridge

416: University Press, Cambridge, 2001).

417: 

418: \bibitem{a} V. S. Afraimovich, N. N. Verichev, and M. I. Rabinovich,

419: Radiophysics and Quantum Electronics, {\bf 29}, 795 (1986).

420: 

421: \bibitem{hanggi} J. A. Freund, L. S. Geier and P. H\"anggi, \ch{13}{225}{2003}.

422: 

423: \bibitem{master} D. A. McQuarrie, J. Appl. Prob. {\bf 4}, 414 (1967).

424: 

425: \bibitem{osw} I. Oppenheim, K. E. Shuler and G. H. Weiss, {\it Stochastic

426: Processes in Chemical Physics: The Master Equation}, (The MIT Press, 1977).

427: 

428: \bibitem{brussel} I. Prigogine and R. Lefever, \jcp{48}{1695}{1968};

429: J. J. Tyson, \jcp{58}{3919}{1973}.

430: 

431: \bibitem{gillespie} D. T. Gillespie, \jpc{81}{2340}{1977}.

432: 

433: \bibitem{megillespie} D. T. Gillespie, \physa{188}{404}{1992}.

434: 

435: \bibitem{cle} D. T. Gillespie, \jcp{113}{297}{2000}; C. W. Gardiner,

436: {\it Handbook of Stochastic Processes}, (Springer-Verlag, Berlin, 1985).

437: 

438: \bibitem{pecora} L. M. Pecora and T. L. Carroll, \prl{64}{821}{1990}.

439: 

440: 

441: \bibitem{lange} In practice, at the end of each integration step, a term

442: $\gamma\eta X^{1/2}_{i}$ is added to the solution of the deterministic

443: equations of motion, $\gamma$ being the strength of the Gaussian noise

444: $\eta$ \cite{cle}.

445: 

446: \bibitem{rpk} M. G. Rosenblum, A. S. Pikovsky, and J. Kurths,

447: \prl{76}{1804}{1996}.

448: 

449: \bibitem{genosc} M. G. Vilar, H. Y. Kueh, N. Barkai and S. Leibler,

450: \pnas{99}{5988}{2002}.

451: 

452: \bibitem{nicolis} G. Nicolis and I. Prigogine, {\it Self-Organization in

453: Nonequilibrium Systems} (Wiley, New York, 1977).

454: 

455: 

456: %\bibitem{kapral} X. G. Wu and R. Kapral, \prl{70}{1940}{1993}.

457: 

458: %\bibitem{crs} B. Schafer, M. G. Rosenblum, and J. Kurths, Nature (London)

459: %{\bf 392}, 239 (1998).

460: 

461: \bibitem{param} We take the following values for the reaction rates

462: \cite{genosc}:

463: $\aa = 50 \h^{-1}$, $\apa = 500 \h^{-1}$, $\ar = 0.01 \h^{-1}$,

464: $\as = \ar + 0.001 \h^{-1}$, $\apr = 50 \h^{-1}$,

465: $\aps = \apr + 5 \hr^{-1}$, $\ba = 50 \h^{-1}$, $\br = 5 \h^{-1}$,

466: $\bs = \br + 0.5 \hr^{-1}$, $\dma = 10 \h^{-1}$, $\dmr = 0.5 \h^{-1}$,

467: $\dms = \dmr + 0.05 \hr^{-1}$, $\da = 1 \h^{-1}$, $\dr = 0.2 \h^{-1}$,

468: $\ds = \dr + 0.02 \hr^{-1}$, $\ga = 1 \mol^{-1} \hr^{-1}$,

469: $\gr = 1 \mol^{-1} \hr^{-1} = \gs$, $\gc = 1 \mol^{-1} \hr^{-1} = \gcp$,

470: $\ta = 50 \hr^{-1}$, $\tr = 100 \hr^{-1} = \ts$. The initial conditions

471: are  $\DA=\DR=\DS=1 \mol, \DAP=\DRP=\DSP=\MA=\MR=\MS=A=R=C=0$, $S=1,

472: C^{\prime}= 1$. The cell has a single copy of the activator and repressor

473: genes: $\DA + \DAP = \DR + \DRP = \DS + \DSP =1 \mol$, and the volume is

474: assumed to be unity.

475: 

476: \bibitem{boccaletti} S. Boccaletti, J. Kurths, G. Osipov, D. L. Valladares

477: and C. S. Zhou, \prep{366}{1}{2002}.

478: 

479: \bibitem{wang} M. Wang, Z. Hou, H. Xin, \jpa{38}{145}{2005}.

480: \bibitem{neiman} A. Neiman, \pre{49}{3484}{1994}.

481: \bibitem{kopell} D. McMillen, N. Kopell, J. Hasty and J. J. Collins,

482: \pnas{99}{679}{2002}.

483: \bibitem{nandi} A. Nandi, Ph.D Thesis, JNU, 2007.

484: 

485: \end{thebibliography}

486: 

487: \end{document}

488: 

489: