1: \documentstyle[epsf,12pt,graphicx]{article}
2: \hoffset=-0.5cm
3: \voffset=-2.3cm
4: %\voffset=1cm
5:
6: \textwidth=15cm
7: \textheight=22.5cm
8: \flushbottom
9: \parindent=0.7cm
10: \topmargin=0cm
11: \leftmargin=0cm
12: \renewcommand{\baselinestretch}{1.04}
13: \oddsidemargin=1.0cm
14: \evensidemargin=1.0cm
15:
16:
17: \begin{document}
18:
19: \def\E{{\rm e}}
20: \newcommand{\be}{\begin{enumerate}}
21: \newcommand{\ee}{\end{enumerate}}
22: \newcommand{\beq}{\begin{equation}}
23: \newcommand{\eeq}{\end{equation}}
24: \newcommand{\uu}{\underline}
25:
26: \centerline{\Large \bf Is there a universality}
27: \centerline{\Large \bf of the helix-coil transition}
28: \centerline{\Large \bf in protein models?}
29: \vskip 1.1cm
30: \centerline{\bf Josh P. Kemp\dag , Ulrich H. E. Hansmann\ddag ,
31: Zheng Yu Chen\dag}a
32: \vskip 0.3cm
33: \centerline{i\it \dag
34: Dept.~of Physics, University of Waterloo, Waterloo, Ontario, N2L 3G1,
35: Canada}
36: \vskip 0.3cm
37: \centerline{\it \ddag Dept.~of Physics, Michigan Technological University,
38: Houghton, MI 49931-1291, USA}
39:
40:
41: \begin{abstract}
42: The similarity in the thermodynamic properties of
43: two completely different theoretical models for
44: the helix-coil transition is examined critically.
45: The first model is an all-atomic representation for a poly-alanine chain,
46: while
47: the second model is a minimal helix-forming model that contains no system
48: specifics.
49: Key characteristics of the helix-coil transition, in particular, the
50: effective
51: critical exponents of these two models agree with each other, within
52: a finite-size scaling analysis.
53:
54: \end{abstract}
55: \centerline{Pacs: 87.15.He, 87.15-v, 64.70Cn, 02.50.Ng}
56:
57:
58: \baselineskip=0.8cm
59: The importance of understanding the statistical physics of the
60: protein-folding problem has been stressed recently
61: \cite{DC,SHA4}.
62: For instance, it is now often assumed that the
63: energy landscape of a protein resembles a partially rough funnel.
64: Folding occurs by a
65: multi-pathway kinetics and the particulars of the folding funnel determine
66: the transitions between the different thermodynamic states
67: \cite{DC,Bryngelson87}.
68: This ``new view'' \cite{DC} of folding was
69: derived from studies of minimal protein
70: models which capture only a few, but probably dominant parameters (chain
71: connectivity, excluded volume, etc.) in real proteins.
72:
73: An implicit yet fundamentally crucial assumption is that the basic mechanism
74: of structural transitions in biological molecules depends solely on gross
75: features of the energy function, not on their details, and that a law of
76: corresponding states can be used to explain dynamics and structural
77: properties of real proteins from studies of related minimal models. This
78: assumption needs to be proven. An even stronger notion in statistical
79: physics is the universality hypothesis for critical phenomena. The critical
80: exponents are identical for different theoretical models and realistic
81: systems belonging to the same universality class. Many theoretical concepts
82: in protein folding, such as coil-helix or coil-globular transitions
83: involve phase transition or phase
84: transition-like behavior. Thus, one wonders if physical measurements between
85: two model systems for the same transition would have any
86: ``universal'' properties.
87:
88: The purpose of this article is to examine these questions for the
89: helix-coil transition in homopolymers of amino acids \cite{Jeff,HO98c}.
90: Traditionaly, the coil-helix transition is
91: described by theories such as the Zimm-Bragg model \cite{ZB}
92: in which the homopolymers are regarded as one dimensional systems with only
93: local interactions; as such
94: a true thermodynamic phase transition is impossible. However, recently there
95: have been
96: \cite{Jeff,HO98c} indications that the coil-helix transition near the
97: transition temperature displays phase-transition like behavior.
98: We use here finite-size scaling analysis, a common tool in statistical
99: physics, to examine
100: the question of universality of the helix-coil transition in two
101: completely different, illuminating models. On one hand, we have a
102: detailed, all-atomic representation of a homo poly-alanine chain
103: \cite{OH95b}. On the other hand, we have a simple coarse-grained model
104: describing the general features of helix-forming polymers \cite{Jeff}.
105: In this article, our interest lies in finding out how far the
106: similarity of the two models go.
107: If the two models yield the same key
108: physical characteristics, then we at least have one concrete example of the
109: validity of the corresponding state principle or universality hypothesis in
110: biopolymer structures.
111:
112: Poly-alanine is well-known to have high helix-propensities in proteins, as
113: demonstrated both experimentally and theoretically
114: \cite{HO98c,OH95b}.
115: It has been well tested and
116: generally believed that approximate force fields, such as ECEPP/2\cite{EC}
117: as implemented in the KONF90 program \cite{Konf}, give protein-structure
118: predictions to a surprisingly degree of faithfulness. As our first model, we
119: have ``synthesized'' poly-alanine with $N$ residues, in which the
120: peptide-bond dihedral angles were fixed at the value 180$^\circ$ for
121: simplicity. Since one can avoid the complications of electrostatic and
122: hydrogen-bond interactions of side chains with the solvent for alanine (a
123: non-polar amino acid), we follow earlier work \cite{OH95b} and neglect
124: explicit
125: solvent molecules in the current study.
126:
127: Our second model is a minimalistic view of a helix forming polymer
128: \cite{Jeff} without atomic-level specifics. A wormlike chain is used to model
129: the backbone of the molecule, while a general directionalized interaction, in
130: terms of a simple square well form, is used to capture the essence of
131: hydrogen like bonding. The interaction energy between the residue labeled $i$
132: and $j$ is modeled by,
133: \begin{equation}
134: \label{minpot}
135: V_{ij}({\bf r}) = \left\{ \begin{array}{cl}
136: \infty & r < D \\
137: %-\epsilon [{\bf u}_i\cdot {\hat {\bf r}}_{ij}]^6
138: -v
139: & D\leq r <\sigma \\
140: 0 & \sigma\leq r
141: \end{array}
142: \right.
143: \end{equation}
144: where
145: $v = \epsilon [{\bf \hat{u}}_i\cdot {\hat {\bf r}}_{ij}]^6
146: + \epsilon [{\bf \hat{u}}_j\cdot {\hat {\bf r}}_{ij}]^6$, ${\bf
147: \hat{u}}_i=({\hat{
148: \bf r}}_{i+1,i})\times({\hat{\bf r}}_{i,i-1})$,
149: ${\hat{ \bf r}}_{ij}$ is the unit vector between monomer $i$ and $j$, $D=
150: 3/2a$ is the diameter of a monomer, $\sigma = \sqrt{45/8}a$ is the bonding
151: diameter,
152: and $a$ is the bond length while bond angle is fixed at $60^\circ$.
153:
154:
155: To obtain the thermodynamic properties, we have conducted multicanonical
156: Monte Carlo simulations for both models. In the low-temperature region where
157: most of the structural changes occur, a typical thermal energy of the order
158: $k_BT$ is much less than a typical energy barrier that the polymer has to
159: overcome. Hence, simple canonical Monte Carlo or molecular dynamics
160: simulations cannot sample statistically independent configurations separated
161: by energy barriers within a finite amount of available CPU time, and usually
162: give rise to bias statistics. One way to overcome this problem is the
163: application of {\it generalized ensemble} techniques \cite{Review}, such as
164: the {\it multicanonical algorithm} \cite{MU} used here, to the protein
165: folding problem, as has recently been utilized and
166: reported\cite{HO}.
167:
168: In a multicanonical algorithm \cite{MU} conformations with energy $E$ are
169: assigned a weight $ w_{mu} (E)\propto 1/n(E)$, $n(E)$ being the density of
170: states. A simulation with this weight generates a random walk in the energy
171: space; since a large range of energies are sampled,
172: one can use the re-weighting techniques \cite{FS} to calculate
173: thermodynamic quantities over a wide range of temperatures by
174: \begin{equation}
175: \left< {\cal{A}}\right>_T ~=~ \frac{\displaystyle{\int
176: dx~{\cal{A}}(x)~w_{mu}^{-1}(E(x))~e^{-\beta E(x)}}}
177: {\displaystyle{\int dx~w_{mu}^{-1}(E(x))~e^{-\beta E(x)}}}~,
178: \label{eqrw}
179: \end{equation}
180: where $x$ stands for configurations and $\beta$ is the inverse temperature.
181:
182:
183: In the case of poly-alanine chains, up to $N=30$ alanine residues were
184: considered. The multicanonical weight factors were determined by the
185: iterative procedure described in Refs.~\cite{MU} and we needed between
186: $4\times 10^5$ sweeps (for $N=10$) and $5\times 10^5$ sweeps (for $N=30$)
187: for estimating
188: the weight factor approximately. All thermodynamic quantities were measured
189: from a subsequent production run of $M$ Monte Carlo sweeps, where
190: $M$=$4\times 10^5$, $5\times 10^5$, $1\times 10^6$, and $3 \times 10^6$
191: sweeps for $N=10$, 15,20, and
192: 30, respectively. In the minimal model, chain lengths up to 39 monomers were
193: considered. In this model a single sweep involves a rotation of a group of
194: monomers via the pivot algorithm\cite{pivot}. For the weight factors the
195: similar
196: number of iterative procedure was used, and for the production run
197: $1\times 10^8$ sweeps was used in all cases.
198:
199: We obtain the temperature dependence of the specific heat, $C(T)$,
200: by calculating
201: \begin{equation}
202: C(T)={\beta}^2 \ \frac{\left< E_{\rm tot}^2\right> - {\left<E_{\rm
203: tot}\right> }^2}{N}~,
204: \label{eqsh}
205: \end{equation}
206: where $E_{\rm tot}$ is the total energy of the system. We also analyze the
207: order parameter $q$ which measures the helical content of a polymer
208: conformation and the susceptibility
209: \begin{equation}
210: \chi (T) = \frac{1}{N-2} (\langle q^2\rangle - \langle q\rangle^2
211: )~.
212: \end{equation}
213: associated with $q$. For poly-alanine $q$ is defined as
214: \begin{equation}
215: q = \tilde{n}_H
216: \end{equation}
217: where $\tilde{n}_H$ is the number of residues (other than the terminal ones)
218: for which the dihedral angles ($\phi,\psi$) fall in the
219: range ($-70 \pm 20^{\circ},-37 \pm 20^{\circ}$). For our worm-like chain
220: model the order parameter $q$ is defined as
221: \begin{equation}
222: \label{wormorder}
223: q = \sum_{i=2}^{N-1} {\bf u}_i \cdot {\bf u}_{i+1}
224: \end{equation}
225: In both cases the first and last residues, which can move more freely, are
226: not counted in the procedure.
227:
228: From a finite-size scaling analysis of the heights and width of specific
229: heat
230: and susceptibility we can extract a set of effective critical exponents which
231: characterize the helix-coil transition in these two models \cite{fukugita}.
232: For instance,
233: with $C_{\rm MAX}$ defined to be the maximum peak in the specific heat,
234: we have
235: \begin{equation}
236: C_{\rm MAX} \propto N^{\displaystyle \frac{\alpha}{d\nu}}~.
237: \label{alpha}
238: \end{equation}
239: In a similar way, we find for the scaling of the maximum of the
240: susceptibility
241: \begin {equation}
242: \chi_{\rm MAX} \propto N^{\displaystyle \frac{\gamma}{d\nu}}~.
243: \label{gamma}
244: \end{equation}
245: %
246: For both quantities we can also define the temperature gap
247: $\Gamma=T_2-T_1$ (where $T_1 < T_{\rm MAX} < T_2$) chosen such that $C(T_1)
248: =b C_{\rm MAX} = C(T_2)$, and $\chi (T_1)=b \chi (T_c) =\chi (T_2)$
249: where $b$ is a fraction.
250: The temperature gap obeys
251: \begin{equation}
252: \Gamma = T_2 - T_1 \propto N^{\displaystyle -\frac{1}{d\nu}},
253: \label{nu1}
254: \end{equation}
255: as has been suggested in Ref.~\cite{fukugita} . The analysis should be
256: insensitive to the actual fraction, $b$, of $C_{\rm MAX}$ ($\chi_{\rm MAX}$)
257: considered
258: for defining $T_1$ and $T_2$
259: which was verified from our numerical data
260: fitting of poly-alanine chains.
261:
262: The scaling exponents, $\alpha, \nu$, and $\gamma$, have their usual meaning
263: in critical phenomena; however, the above scaling relations also hold
264: formally
265: for the case of a first-order transition, with effective scaling
266: exponents $d\nu = \alpha = \gamma = 1$~\cite{fukugita,binder}.
267: Note that $d$ is the dimensionality of the system, and it always appears in
268: the combination $d\nu$. Without knowing further the effective dimensionality
269: of our systems, we use the combination $d\nu$ as a single parameter in the
270: fit.
271:
272: It then becomes straightforward to use the above equation and the values
273: given in Table~\ref{tab1} to estimate the critical
274: exponents. We
275: obtain for poly-alanine from the scaling of the width of the specific heat
276: $1/d\nu=1.02(11)$ with a goodness of fit $(Q=0.9)$ (see Ref.~\cite{NR} for
277: the definition of $Q$), for chains of length $N=15$ to $N=30$. Inclusion of
278: $N=10$ leads to $1/d\nu=0.84(7)$, but with a less acceptable fit $(Q=0.1)$.
279: Similarly, we find from the scaling of the width of the susceptibility
280: $1/d\nu = 0.98(11)$ $(Q=0.5)$ for chains of length $N=15$ to $N=30$ and
281: $1/d\nu=0.81(7)$ $(Q=0.2)$ when the shortest chain $N=10$ is included in the
282: fit. Hence, we present as our final estimate for the correlation exponent of
283: poly-alanine $d\nu=1.00(9)$. This value is in good agreement with the
284: estimate $d\nu =0.93(5)$ obtained from the partition function zero analysis
285: in Ref.~\cite{AH99b}.
286:
287: The results for the exponent $\alpha$ give $\alpha = 0.89(12)$ (Q=0.9) when
288: all chains
289: are considered, and $\alpha = 0.86(10)$ $(Q=0.9)$ when the shortest chain is
290: excluded from the fit. Analyzing the peak in the susceptibility we find
291: $\gamma = 1.06(14)$ $(Q=0.5)$ for chain lengths $N=15-30$ and $\gamma =
292: 1.04(11)$ $(Q=0.5)$ for chain lengths $N=10-30$.
293: We summarize our final estimates for the critical exponents in Table
294: \ref{tab5}. The scaling plot for the susceptibility
295: is shown in Fig.~1: curves for all lengths of
296: poly-alanine chains collapse on each other indicating the validity of finite
297: size scaling of our poly-alanine data.
298:
299: The same procedure can be applied to analyze the data from the minimal model.
300: All calculation has been done with the omission of the
301: shortest chain. Using the widths of the specific heat a $b=80\%$ of the peak
302: height we obtain $1/d\nu~=~1.03(7)$, $(Q=0.2)$. The
303: width of the
304: peak at half maximum is more unreliable in this case as the coil-helix
305: transition is complicated by the additional collapsing transition to a
306: globular state in the vicinity of the coil-helix transition\cite{Jeff}. This
307: exponent agrees with that calculated from the susceptibility widths,
308: $1/d\nu~=~0.89(9)$, $(Q=0.3)$. Hence, our final estimate for this critical
309: exponent in
310: our second model is $d\nu = 0.96(8)$. These values are in good agreement
311: with those of the poly-alanine model.
312:
313:
314: From the $C_{\rm MAX}$ data in Table~\ref{tab1} and using the above given
315: value for
316: the exponent $d\nu$ we find $\alpha~=~0.70(16)$ ($Q=0.3$) which is somewhat
317: smaller than that of the poly-alanine model. The susceptibility exponent as
318: calculated from the data in Table~\ref{tab1} yields a value of
319: $\gamma~=~1.3(2)$
320: ($Q=0.5$), which agrees with the previous estimation within the error bar.
321: The scaling plot for the susceptibility is shown in
322: Fig.~2. While curves corresponding to large polymer
323: sizes collapse into the same curve, the $N=13$ case shows small disagreement,
324: indicating that the finite size scaling are valid only for longer chain
325: lengths in the minimal model.
326:
327: Comparing the critical exponents of our two models as summarized in Table
328: \ref{tab5} we see that the estimates for the correlation exponent $d\nu$
329: agrees well for the two models. Within the error bars, the estimates for the
330: susceptibility exponent $\gamma$ also agree. The estimates for the specific
331: heat exponent $\alpha$ seem disagree within the error ranges. However, in
332: view of the fact that both analyses are based on small system size the true
333: error ranges could be actually larger than the ones quoted here. Using these
334: rather crude results, we have already demonstrated a striking similarity in
335: finite-size scalings of the two model. Therefore, we can convincingly make
336: the conjecture that minimal model can be used to represent the structural
337: behavior of real helix-forming proteins.
338:
339: Our analysis should tell us also whether the helix-coil
340: transition in our models is of first or second order.
341: In the former case we
342: would expect $d\nu = \alpha = \gamma = 1$
343: which seems barely supported by our
344: data due to the rather large error bars associated with the estimate of the
345: exponents.
346: We have further explored the nature of the transition from another
347: perspective, by considering the change in energy crossing a small temperature
348: gap (taken to be within 90\% of $C_{\rm MAX}$) from the original data,
349: \begin{equation}
350: \label{delE}
351: \Delta E = (E_{\rm tot}(T_2) - E_{\rm tot}(T_1))/N
352: \end{equation}
353: This value should approach either a finite value or zero as
354: $N^{-1}$ goes to zero. A finite value would indicate a first order transition
355: while a zero value a second order transition. In the case of a first order
356: transitions the intercept would indicate the latent heat. Now, the
357: assumption is that this energy change scales linearly as $N^{-1}$ goes to
358: zero. Figure 3 shows a plot of the data from
359: both the atomic-level and minimal models, where nonzero intercepts can be
360: extrapolated at $N^{-1}$=0.
361: Hence, our results seem to indicate and finite latent heat and a
362: first-order like helix-coil transition. However, we can not
363: exclude the possibility that the true asymptotic limit of
364: $|E|$ is zero, and some of the results of Ref.~\cite{HO98c} point for
365: the case of poly-alanine rather towards a second-order transition.
366: Further simulations of larger chains seem to be necessary to determine the
367: order of the helix-coil transition without further doubts.
368:
369: In summary, we conclude that in view of the similarity of the two models
370: examined here, a corresponding state principle can be established for the
371: coil-helix transition. Examining the finite size scaling analysis allows
372: us to calculate estimators for critical exponents %and latent heat
373: in the two models which indicate ``universality'' of helix-coil transitions.
374:
375:
376: \noindent
377: {\bf Acknowledgments}:
378: Financial supports from Natural Science and Engineering Research Council of
379: Canada and the National Science Foundation (CHE-9981874)
380: are gratefully acknowledged.
381:
382:
383: \begin{thebibliography}{99}
384: \bibitem{DC} K.A. Dill and H.S. Chan, Nature Structural Biology {\bf 4}, 10
385: (1997).
386: \bibitem{SHA4} E.I. Shakhnovich, Curr. Opin. Struct. Biol.{\bf 7}, 29
387: (1997); T. Veitshans, D. Klimov and D. Thirumalai, Fold.Des. {\bf 2},
388: 1 (1997).
389: \bibitem{Bryngelson87} J.D. Bryngelson and P.G. Wolynes, Proc. Natl. Acad.
390: Sci. U.S.A. {\bf 84}, 524 (1987);J.N. Onuchic, Z. Luthey-Schulten,
391: P.G. Wolynes, Annual Reviews in Physical Chemistry {\bf 48}, 545
392: (1997).
393: \bibitem{Jeff} J.P. Kemp and Z.Y. Chen, Phys. Rev. Lett.{\bf 81}, 3880
394: (1998).
395: \bibitem{HO98c} U.H.E. Hansmann and Y. Okamoto, J. Chem. Phys. {\bf 110},
396: 1267 (1999); {\bf 111} (1999) 1339(E).
397: \bibitem{ZB} B.H. Zimm and J.K. Bragg, J. Chem. Phys. {\bf 31}, 526 (1959).
398: \bibitem{OH95b} Y. Okamoto and U.H.E. Hansmann, J. Phys. Chem. {\bf 99},
399: 11276 (1995).
400: \bibitem{AH99b} N.A.Alves and U.H.E. Hansmann, Phys. Rev. Let. {\bf 84}
401: (2000) 1836.
402: \bibitem{EC} M.J. Sippl, G. N{\'e}methy, and H.A. Scheraga, J. Phys.
403: Chem. {\bf 88}, 6231 (1984), and references therein.
404: \bibitem{Konf} H. Kawai et al., Chem. Lett.{\bf 1991}, 213 (1991); Y.
405: Okamoto et al., Protein Engineering {\bf 4}, 639 (1991).
406: \bibitem{Review} U.H.E. Hansmann and Y. Okamoto, in: Stauffer, D. (ed.)
407: {\it Annual Reviews in Computational Physics VI},(Singapore: World
408: Scientific), p.129. (1998).
409: \bibitem{MU} B.A. Berg and T. Neuhaus, Phys. Lett. {\bf 267}, 249 (1991).
410: \bibitem{HO} U.H.E. Hansmann and Y. Okamoto, J. Comp. Chem. {\bf 14}, 1333
411: (1993);{\bf 18}, 920 (1997).
412: \bibitem{FS} A.M. Ferrenberg and R.H. Swendsen, Phys. Rev. Lett. {\bf 61},
413: 2635 (1988); Phys. Rev. Lett. {\bf 63}, 1658(E) (1989), and
414: references given in the erratum.
415: \bibitem{pivot} N. Madras and A.D. Sokal, J. Stat. Phys. {\bf 50}, 109
416: (1988).
417: \bibitem{fukugita} M. Fukugita, H. Mino, M. Okawa and A. Ukawa J. Stat. Phys.
418: {\bf 59}, 1397 (1990), and references given therein.
419: \bibitem{binder}K. Binder, D.W. Heermann, {\it Monte Carlo Simulation in
420: Statistical Physics} Springer-Verlag, Berlin, 1988
421: \bibitem{NR} W.H. Press, S.A. Teukolsky, W.T. Vetterling and
422: B.P. Flannery, {\it Numerical Recipes}, 2nd ed. (Cambridge University
423: Press, New York, 1992) p. 657.
424: \end{thebibliography}
425:
426: \newpage
427: {\Large Figure Captions}\\
428: \begin{enumerate}
429: \item Scaling plot for the susceptibility $\chi(T)$
430: as a function of temperature $T$, for poly-alanine
431: molecules of chain lengths $N = 10, 15, 20$ and $30$.
432: \item Scaling plot of $\chi(T)$
433: as a function of temperature $T$, for the minimum model
434: of chain lengths $N = 13, 19, 26, 33$ and $39$.
435: \item Scaling of energy gap and transition width at 80\% and 90\%
436: of $C_{MAX}$. Here we have used $\Delta E_{80\%}$($\bigtriangleup$ for
437: all-atom model,$\Diamond$ for minimal model), $\Delta E_{90\%}$($\Box$ for
438: all-atom model, $\bigcirc$ for minimal model).
439: \end{enumerate}
440:
441:
442: \newpage
443: \begin{table}
444: \caption{Shown are the location of
445: the specific heat maximum $T_{\rm MAX}$, the maximum of specific heat $C_{\rm
446: MAX}$, susceptibility $\chi_{\rm MAX}$, the width of the half peak in
447: specific
448: heat $\Gamma_{C}$, and width of the half peak of susceptibility
449: $\Gamma_{\chi}$ for various chain lengths.}
450: \begin{center}
451: \begin{tabular}{|c|c|c|c|c|c|}
452: \hline
453: $N$&$T_{\rm MAX}$& $C_{\rm MAX}$ & $\Gamma_{C}$ & $\chi_{\rm MAX} $ &
454: $\Gamma_{\chi}$\\
455: \hline
456: \multicolumn{6}{|c|}{All-Atomic Model}\\\hline
457: 10& 427(7)& 8.9(3) & 160(7) & 0.49(2) & 140(7)\\
458: 15& 492(5)& 12.3(4)& 119(5) & 0.72(3)& 110(5)\\
459: 20& 508(5)& 16.0(8)& 88(5) & 1.08(3)& 78(5)\\
460: 30& 518(7)& 22.8(1.2)&58(4)& 1.50(8)& 56(3)\\
461: %\hline
462: \hline
463: \multicolumn{6}{|c|}{Minimal Model}\\\hline
464: 13& 1.25(1) & 1.088(2) & 1.22(2) & 0.232(2) & 2.20(2) \\
465: 19& 1.17(1) & 1.424(5) & 1.12(2) & 0.353(3) & 0.81(2) \\
466: 26& 1.16(1) & 1.789(8) & 0.89(2) & 0.553(8) & 0.57(2) \\
467: 33& 1.13(1) & 2.08(1) & 0.73(2) & 0.78(1) & 0.45(2) \\
468: 39& 1.12(1) & 2.27(2) & 0.61(2) & 0.96(2) & 0.41(2) \\\hline
469: \end{tabular}
470: \end{center}
471: \label{tab1}
472: \end{table}
473:
474: % Table 3
475: \begin{table}
476: \caption{Summary of the critical exponents obtained for the two models.}
477: \begin{center}
478: \begin{tabular}{|c|c|c|}
479: \hline
480: & All-atomic & Minimal \\
481: \hline
482: $d\nu$ & 1.00(9) & 0.96(8)\\
483: $\alpha$ & 0.89(12) & 0.70(16)\\
484: $\gamma$ & 1.06(14) & 1.3(2)\\\hline
485: \end{tabular}
486: \end{center}
487: \label{tab5}
488: \end{table}
489:
490:
491:
492:
493: %FIGURE 1 Image
494: \pagebreak
495: \begin{center}
496: \includegraphics[width=6.0in]{Fig1.eps}
497: \end{center}
498: \vspace{2in}
499: \begin{center}
500: {\Large Fig.~1 }
501: \end{center}
502:
503:
504:
505: %FIGURE 2 Image
506: \pagebreak
507: \begin{center}
508: \includegraphics[width=6.0in]{Fig2.eps}
509: \end{center}
510: \vspace{1.5in}
511: \begin{center}
512: {\Large Fig.~2 }
513: \end{center}
514:
515: %FIGURE 3 Image
516: \pagebreak
517: \begin{center}
518: \includegraphics[width=6.0in]{Fig3.eps}
519: \end{center}
520: \vspace{1in}
521: \begin{center}
522: {\Large Fig.~3 }
523: \end{center}
524:
525: \end{document}
526: