cond-mat0012093/a.tex
1: % DRAFT!  VERSION OF June 13, 2000    PREPRINT-Nr.: MTU-PHY-HA/00-2
2: 
3: %\documentstyle[aps,prl,twocolumn,graphicx,floats]{revtex}
4:  \documentclass[eps,12pt]{article}
5:  \usepackage{graphicx}
6: %%%%%%%%%%%%%%%%%%%%%%%%%%%
7: % alves
8:  \setlength\textwidth{15cm}
9:  \setlength\textheight{22.5cm}
10: \voffset=1.8cm
11: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
12: 
13: \renewcommand{\thesection}{\arabic{section}}
14: %%\renewcommand{\thesubsection}{\arabic{subsection}}
15: \renewcommand{\thetable}{\arabic{table}}
16: 
17: \begin{document}
18: 
19: % SETS A DOUBLE SPACE
20: \baselineskip=18pt
21: 
22: %{\wideabs{
23: \begin{center}
24: {\Large\bf Yang-Lee zeros and the helix-coil transition}
25: {\Large\bf in a continuum model of polyalanine} 
26: \vskip 1.1cm
27: {\bf Nelson A. Alves\footnote{E-mail: alves@quark.ffclrp.usp.br}} 
28: \vskip 0.1cm
29: {\it Departamento de F\'{\i}sica e Matem\'atica, FFCLRP 
30:      Universidade de S\~ao Paulo. \\
31:       Av. Bandeirantes 3900.  CEP 014040-901 \, 
32:       Ribeir\~ao Preto, SP, Brazil}
33: \vskip 0.3cm
34: {\bf Ulrich H.E. Hansmann \footnote{E-mail: hansmann@mtu.edu}}
35: \vskip 0.1cm
36: {\it Department of Physics, Michigan Technological University,
37:          Houghton, MI 49931-1291, USA}
38: 
39: \vskip 0.3cm
40: \today
41: \vskip 0.4cm
42: \end{center}
43: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
44: \begin{abstract}
45: We calculate the Yang-Lee zeros for characteristic temperatures of 
46: the helix-coil transition in a continuum model of polyalanine. 
47: The distribution of these zeros differs from  predictions of the 
48: Zimm-Bragg theory and supports recent claims that polyalanine
49: exhibits a true phase transition. New estimates for critical exponents
50: are presented and the relation of our results to the Lee-Yang theorem is
51: discussed.
52: 
53: \vskip 0.1cm
54: {\it Keywords:} Yang-Lee zeros, Lee-Yang theorem, Helix-coil transition,
55:                 Protein folding, Generalized ensemble simulations
56: 
57: 
58: \end{abstract}
59: \vskip 0.1cm
60: {\it PACS-No.: 87.15.Aa, 64.60.Cn, 64.60.Fr }
61: %} 
62: 
63: %\newpage
64: %\vskip 1cm
65: 
66: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
67: %\begin{multicols}{2}
68: %\twocolumn
69: 
70: \section{Introduction} 
71: \vspace*{-0.5pt}
72: %\noindent
73: \indent
74:   
75: Recently, there has been an increased interest in the statistical physics of
76: biological macromolecules. For instance, the conditions under which
77: \hbox{$\alpha$-helices}, a  common structure in proteins, 
78: are formed or disolved, have been extensively
79: studied \cite{Poland}. Traditionary, the characteristics of the 
80: observed sharp transition between random coil state and (ordered) helix,
81: have been described in the framework of  Zimm-Bragg-type theories \cite{ZB}.
82: In these theories, the molecules are  approximated by a one-dimensional
83: Ising model with the residues as ``spins'' taking values ``helix'' or ``coil'',
84: and solely local interactions. Hence, thermodynamic phase transitions are
85: not possible in these theories. 
86:     However, in previous work 
87: \cite{HO98c,AH99b,KHC99d} evidence was presented that 
88:  polyalanine exhibits a phase transition between the ordered 
89: helical state and the disordered random-coil state when interactions
90: between all atoms in the molecule are taken into account. 
91: Investigating the finite size scaling of various
92: quantities such as specific heat, susceptibility and the Fisher zeros,
93: we studied the nature of this phase transition and presented estimates for 
94: the critical exponents $\nu$, $\alpha$ and $\gamma$. 
95: Here, we  continue our previous investigation by exploring the
96: helix-coil transition in polyalanine from a different point of view. 
97: We intend to demonstrate in greater detail
98: the differences in the critical behaviour of our model and the Zimm-Bragg
99: theory, and we will present additional  critical exponents.
100: 
101:   Yang and Lee have established long ago \cite{Yang-L} that the statistical
102: theory of phase transitions can also be described by means
103: of the distribution of the
104: zeros of the grand partition function in the complex fugacity  plane.
105: In the case of the Ising model this description corresponds 
106: to  study this model with an external magnetic field $H$ 
107: mathematically extended from real to complex values.
108: Although complex values for the fugacity are not physical, 
109: the characteristics of the phase transition  can be extracted from
110: the distribution of the zeros.
111:   As the volume of the
112: given (finite) system increases, the number of complex zeros grows 
113: and they will close onto the positive real fugacity axis.  
114: In the thermodynamic limit, those zeros 
115: circumscribe closed regions on this plane, thus defining thermodynamic 
116: phases which are themselves free of zeros. 
117: In Ref.~\cite{Lee-Y} Lee and Yang  proved 
118: the famous circle theorem, namely that in the Ising model with 
119: ferromagnetic couplings the complex zeros lie on the unit circle
120: in the complex $y-$plane,  $y = e^{-2 \beta H}$.
121: 
122: The goal of our paper is to investigate whether the helix-coil transition
123: in biological macromoelecules can also be described in the frame work
124: of the Yang-Lee zeros.  For polyalanine, the analogous of the magnetization
125: in the Ising model is  the  number of helical residues $M$. 
126: Formally we can  introduce a (non physical) 
127: external field $H$ as the conjugate variable to this
128: order parameter $M$ (which plays the role of a magnetization), and study the
129: Yang-Lee zeros. Such an investigation allows us not only to check our 
130: previous results with an independent method, it opens also the
131: possibility to calculate new critical exponents and to point out 
132: much clearer the differences between our all-atom model of polyalanine
133: and the Zimm-Bragg theory. If the helix-coil transition can indeed be described
134: by the Zimm-Bragg model, then the distribution of the Yang-Lee zeros 
135: in function of the temperature
136: should
137: resemble that of the one dimensional Ising model. On the other hand, a 
138: substantially different distribution would demonstrate that the helix-coil
139: transition in polyalanine is not accurately 
140: described by the Zimm-Bragg theory.
141: 
142: 
143: Before proceeding, we give the outline of the paper. In the next
144: section we describe the numerical evaluation of the partition function.
145: Our approach and numerical results are presented in the third Section, 
146: which is followed by our conclusions.
147: 
148: 
149: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
150: \section{Methods}
151: \noindent
152: Our investigation of the helix-coil transition for polyalanine is
153: based on a detailed, all-atom representation of that homopolymer. 
154: Since one can avoid the complications of electrostatic and 
155: hydrogen-bond interactions of side chains with the solvent for alanine 
156: (a nonpolar amino acid), explicit solvent molecules were neglected. 
157: The interaction between the atoms  was
158: described by a standard force field, ECEPP/2,\cite{EC}  (as implemented 
159: in the KONF90 program \cite{Konf}) and is given by:
160: \begin{eqnarray}
161: E_{tot} & = & E_{C} + E_{LJ} + E_{HB} + E_{tor},\\
162: E_{C}  & = & \sum_{(i,j)} \frac{332q_i q_j}{\epsilon r_{ij}},\\
163: E_{LJ} & = & \sum_{(i,j)} \left( \frac{A_{ij}}{r^{12}_{ij}}
164:                                 - \frac{B_{ij}}{r^6_{ij}} \right),\\
165: E_{HB}  & = & \sum_{(i,j)} \left( \frac{C_{ij}}{r^{12}_{ij}}
166:                                 - \frac{D_{ij}}{r^{10}_{ij}} \right),\\
167: E_{tor}& = & \sum_l U_l \left( 1 \pm \cos (n_l \chi_l ) \right).
168: \end{eqnarray}
169: Here, $r_{ij}$ (in \AA) is the distance between the atoms $i$ and $j$, and
170: $\chi_l$ is the $l$-th torsion angle. Note that with the electrostatic
171: energy term $E_{C}$ our model contains a long range interaction neglected
172: in the Zimm-Bragg theory \cite{ZB}.  It was  conjectured  that it is 
173: this long range interaction in our model and  the fact that it is not  
174: one-dimensional which allows the existence of the observed phase
175: transition in Refs.~\cite{HO98c,AH99b,KHC99d}. 
176: We remark that  the 1D Ising model with long-range interactions
177: also exhibits  a phase transition at finite $T$ if the interactions
178: decay like $1/r^{\sigma}$ with $1\le \sigma < 2$ \cite{1dlr}.
179: 
180: Simulations of detailed models of biological macromolecules are 
181: notoriously difficult. This is because the various competing interactions 
182: within the polymer lead to an energy landscape characterized by a 
183: multitude of local minima.  Hence, in the low-temperature region, 
184: canonical Monte Carlo or molecular dynamics 
185: simulations will tend to get trapped in one of these
186: minima and the simulation will not thermalize within the available
187: CPU time. Only recently, with the introduction of new and sophisticated 
188: algorithms such as  {\it multicanonical sampling} \cite{MU} and other 
189: {\it generalized-ensemble} techniques \cite{Review} was it
190: possible to alleviate this problem in  protein simulations \cite{HO}. 
191: For polyalanine,  both the failure of standard Monte Carlo techniques 
192: and the superior performance of the multicanonical algorithm are 
193: extensively documented  in earlier work \cite{OH95b}. For this reason,
194: we use again this simulation technique for our project  where we
195: considered polyalanine chains of up to $N=30$ monomers. 
196:  
197: In the multicanonical algorithm \cite{MU}
198: conformations with energy $E$ are assigned a weight
199: $  w_{mu} (E)\propto 1/n(E)$. Here, $n(E)$ is the density of states.
200: A  simulation with this weight
201: will  lead to a uniform distribution of energy:
202: \begin{equation}
203:   P_{mu}(E) \,  \propto \,  n(E)~w_{mu}(E) = {\rm const}~.
204: \label{eqmu}
205: \end{equation}
206: This is because the simulation generates a 1D random walk in the 
207: energy space,
208: allowing itself to escape from any  local minimum.
209: Since a large range of energies are sampled, one can
210: use the reweighting techniques \cite{FS} to  calculate thermodynamic
211: quantities over a wide range of temperatures $T$ by
212: \begin{equation}
213: <{\cal{A}}>_T ~=~ \frac{\displaystyle{\int dx~{\cal{A}}(x)~w^{-1}(E(x))~
214:                  e^{-\beta E(x)}}}
215:               {\displaystyle{\int dx~w^{-1}(E(x))~e^{-\beta E(x)}}}~,
216: \label{eqrw}
217: \end{equation}
218: where $x$ stands for configurations.
219: 
220: Note that unlike in the case of canonical simulations the weights are
221: not a priori known. 
222: We needed between 40,000 sweeps ($N=10$)  and 500,000 sweeps ($N=30$) for
223: the weight factor calculations by the iterative  procedure described in 
224: Refs.~\cite{MU,HO94c}. 
225: All thermodynamic quantities were  estimated from one
226: production run of $N_{sw}$ Monte Carlo sweeps starting from a random
227: initial conformation, i.e. without introducing any bias.
228: We chose $N_{sw}$=2\,000\,000, 2\,000\,000, 4\,000\,000, and 3\,000\,000 
229: sweeps for $N=10$, 15, 20, and 30, respectively. 
230: 
231: 
232: It follows from Eq.~(\ref{eqmu}) that the multicanonical algorithm allows us
233: to calculate estimates for the spectral density:
234: \begin{equation}
235:   n(E) = P_{mu} (E) w^{-1}_{mu} (E)~.
236: \end{equation}
237: We can therefore construct
238: the corresponding partition function for our all-atom model of polyalanine
239: keeping track of the corresponding number of helical residues $M$ at a 
240: given energy,  
241: \begin{equation}
242: Z(u,y)\,=\, \sum_{M=0}^{N-2} \sum_E n(E,M) u^E y^M ,     \label{eq:z1}
243: \end{equation}
244: where 
245: $u = e^{-\beta}$ with $\beta$ the inverse temperature,
246: $\beta=1/k_BT$ and $y=e^{-H}$. 
247: We have absorbed the factor $\beta$ in the 
248: definition of the  field $H$ which is the conjugate variable to our
249: order parameter $M$. The later
250: quantity is defined  by the condition that for a helical residue 
251: the dihedral angles ($\phi,\psi$) fall in the
252: range ($-70 \pm 20^{\circ},-37 \pm 20^{\circ}$).
253: Note, that for convenience we have discretized the energy in Eq.~(\ref{eq:z1}).
254: In previous work 
255: %%%\cite{AH99b}
256: we found that the partition function zeros depend only
257: weakly on the energy bin size, and for the present study we choose 
258: energy bins of length $0.5$ kcal/mol.  
259: 
260: 
261: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
262: \section{Yang-Lee Zeros of Polyalanine}
263: \noindent
264: In order to analyse the helix-coil transition in our all-atom model
265: of polyalanine we  first determine the (pseudo)critical temperatures.
266: This can be done in various ways. For instance, in Ref.~\cite{AH99b}
267: these temperatures where calculated from the distribution of the
268: Fisher zeros \cite{Itzykson}. Here, we  check our previous results
269: by calculating estimates for the pseudo-critical temperatures in a 
270: different way.
271:   
272:  Our temperature-driven transition model can be analysed by considering
273: the coexistence of two phases: a disordered (coil) phase  and
274: an ordered (helix) phase at the critical temperature $T_c$.
275: This behaviour can be described by the existence of an equilibrium
276: probability distribution $P_N(M)$ for the corresponding order 
277: parameter $M$ \cite{BinderB30,ChallaB34}. 
278:   This approach leads to a definition of a finite size critical temperature
279: by approximating the equilibrium distribution by a sum of two
280: distributions, each one characterizing the corresponding phase.
281:   This ansatz gives origin  to a two-peak like distribution separated by a
282: minimum and has led to the finite size scaling study
283: of the coexistence of bulk phases \cite{Lee-K90}.
284:   Following these works, we consider for our  purpose
285: the histograms of the helicity for all chain
286: sizes $N$, which is our order parameter (and 
287: the equivalent to the magnetization in the Ising model), 
288: at different temperatures $\beta$. 
289:    It is straightforward to calculate the histograms of 
290: helicity distribution at these temperatures
291: from our multicanonical estimates of the density of states,
292: \begin{equation}
293:                      w(M)\,=\, \sum_E n(E,M) u^E .     \label{eq:hist}
294: \end{equation}
295:    As expected,
296: these histograms exhibit a clear double peak structure at 
297: what we call critical temperature, while at high (low) temperatures one 
298: observes solely a predominant single peak at low (high) values of the 
299: helicity. 
300:   Hence, we define our critical temperatures by the condition that
301: the reweighted histogram of the magnetization have two peaks of equal heights.
302:   As an example we show in Fig.~1 the histogram normalized to area one,
303: for $N=30$ at $T_c=518$ K.
304:   
305: 
306:   Table 1 summarizes the so obtained set of critical temperatures by this
307: method. The quoted values and errors were obtained by reweighting 
308: independent binned data to temperatures $T_c^i$, $i=1,2, ..., nbins$, 
309: where one obtains equal heights in the histograms for our order 
310: parameter $M$.  Our statistics were based on $nbins = 19, 20, 
311: 16 \,{\rm and}\, 3$ for  $N=10,15, 20\, {\rm and}\, 30$, respectively. 
312: for comparison we also list our previous estimates \cite{AH99b}
313: which were derived from an analysis of the Fisher zeros and the position
314: of the maximum of specific heat $C_v$.  We note, that our new results 
315: agree within the errors with our previous estimates for the 
316: pseudo-critical temperatures which demonstrates the validity
317: of our approach. 
318: 
319: 
320: After having demonstrated the reliability of our data we calculate now 
321: at the so obtained critical temperatures the distribution of the Yang-Lee 
322: zeros by means of Eq.~(\ref{eq:z1}).  Our results are displayed in 
323: Fig.~2 for the case of chain lengths $N=30$. In that figure
324:  we draw in addition a  
325: unit circle to point out the relative position of our zeros around it. 
326:   Similar distributions are obtained for smaller systems.
327:   The presented error bars for the zeros in this figure, 
328: were estimated by considering the new partition functions  at 
329: the limiting temperatures  $T_c(N) \pm \Delta T_c(N)$ where $\Delta T_c(N)$
330: is the error in our pseudo-critical temperatures $T_c(N)$. 
331:   Our error bars  clearly 
332: exclude the possibility that the complex zeros are distributed
333: on the unit circle. 
334:   On the other hand, the Yang-Lee zeros for the 
335: 1D Ising model (which can be calculated exactly) fall on the unit circle. 
336: This is shown in Fig.~3 for chain lengths 
337: $N=5, 15$ and $30$.\footnote{As a further illustration for the  behaviour 
338:       of the continuous loci of zeros for the one dimensional 
339:       Ising partition function, we refers to Ref.~\cite{Katsura},
340:       and Ref.~\cite{Glumac94} for $q=2$ states.} 
341: 
342: After having demonstrated the difference in the two models
343: at the critical point, we extend our study to the behaviour 
344: of our polyalanine
345: zeros as function of the temperature.
346:   In Fig. 4 we show for our smallest polyalanine chain how the zeros 
347: change with temperature $u(\beta)$, measured in  multiple 
348: of the critical value $u_c$. Again, we draw for illustration a unit circle
349: to call attention to the relative position of our zeros. 
350:   We exhibit the smallest chain since the density
351: of zeros increases with growing chain length making it difficult to
352: the eyes to
353: trace the movement of the zeros in the complex $y$-plane. 
354:  As the temperature increases,  $u(\beta) > u_c$, 
355: we observe that the zeros move away from the unit circle and
356: consequently the edge zero, the zero closest to the positive real $y-$axis,
357: moves away from the real axis.
358:      A similar behaviour is known for the $q$-state 
359:      Potts model with $q>2$ on finite lattices for temperatures 
360:      larger than the critical one \cite{Kim-Cres1}.
361:  For $u(\beta) < u_c$, we find again that
362: the zeros do not fall all on the unit circle. 
363: Hence, the Lee-Yang theorem does not apply for our model demonstrating again 
364: that polyalanine should not be described by  Ising-type models. 
365: 
366:  Figure 4 indicates that the angle $\theta_0$, the angle at
367: which the  edge zeros deviate from the real $y$-axis, grows as 
368: the temperature is increased  above $T_c$, i.e. $u(\beta) >  u_c$. 
369:  For larger lengths $N$, the number of zeros increases and the
370: edge zeros move towards the real axis. 
371: Note  that the movement for the edge zero seems to follow an
372: unit arc, leading to $Re(y)=1$ and  the real value $H=0$ in the 
373: thermodynamic limit.  The movement of the edge zeros towards the 
374: real axis itself is a signal for the 
375: possibility of a true phase transition in the thermodynamic limit
376: if the locus of zeros cut the positive real axis at finite temperature.
377: In that case one would expect the following scaling relation for the 
378: edge zeros \cite{Kim-Cres2}:
379: \begin{equation}
380: \theta_0(t, N) \,=\, N^{-y_h/d} \theta_0(t N^{y_t/d})\, ,   \label{eq:t0}
381: \end{equation}
382: where $t= (T-T_c)/T_c$, $y_h$ is the magnetic scaling exponent
383: and $y_t = 1/\nu$ is the thermal scaling exponent.
384: Note, that we have introduced the scaling variable $N^{1/d}$ 
385: in place of the usual linear length $L$. This is because the number
386: of monomers $N$ is the natural quantity to describe polyalanine chains.
387:   Since we have no theoretical indication to assume  a particular
388: integer geometrical dimension $d$ for our polypeptide, we present our
389: results as a function of this parameter $d$.
390: 
391: Now, assuming that the above finite size scaling relation holds for polyalanine
392: chains we can calculate the ``magnetic'' scaling exponent \cite{Itzykson}
393: \begin{equation}
394:                   y_h/d = (\beta + \gamma)/d\nu
395: \label{eq:yhb}
396: \end{equation}
397:  if we take $t=0$, i.e.  
398: %%\longleftrightarrow 
399: $u(\beta) = u_c$. For this reason
400: we display in Fig. 5 the linear regression for ${\rm ln}\,(\theta_0(N))$. 
401:  The least-square fit gives us
402: \begin{equation}
403:                     y_h/d = 0.97(4) \, ,             \label{eq:hy}
404: \end{equation}
405: with the goodness of fit $Q=0.67$. This value for $Q$ means
406: the fit is acceptable
407: for the present precision data and corroborates our expectation.
408: 
409: %It is interesting to compare our estimate with some values for $y_h/d$
410: %in Ising models.
411: %We have $y_h/d = 15/16 = 0.9375$ for 2D Ising model
412: %and, for $\nu = 0.6294(5)(5)$, $\beta = 0.3265(3)(1)$, 
413: %and $\gamma =1.2353(11)(14)$ (see Ref.~\cite{Parisi99}, Table 4) we obtain 
414: %$y_h/d = 0.8271$ for 3D Ising model, with an error of $0.0019$, estimated
415: %from a simple error propagation.
416: 
417: 
418: If we take our estimate in Eq.~(\ref{eq:hy}) and the reported values
419: for $\gamma$ and $d\nu$ from Ref.~\cite{AH99b}, $\gamma= 1.06(14)$,
420: $d\nu = 0.93(5)$, we can evaluate the exponent $\beta$ either 
421: by the defining equation Eq.~(\ref{eq:yhb}) or from the relation
422: $ 2 \beta \,+\, \gamma = d \nu$. The first option is less precise since
423: it envolves three parameters 
424: ($\beta=-0.16(15)$),  while the second one, relying on two previous estimates,
425:  gives $\beta=-0.065(74)$. As one should expect, both estimates agree
426: with each other within the error bars demonstrating the consistency of
427: this study with earlier investigations \cite{HO98c,AH99b,KHC99d}. 
428: Note also that our estimates for the critical exponent $\beta$ do not
429: help us to answer a question we had  to leave unanswered in
430: the earlier work, namely, whether the helix-coil transition is a
431: first order phase transition ($\beta = 0$) or a (strong) second order
432: phase transition ($\beta \ne 0$). This is not surprising since clarifying 
433: the  nature of a phase transition can be a challenging issue 
434: even for
435: simpler systems (such as Potts models \cite{LandauB39}, 
436: to name only one example)
437: and requires dedicated large-scale simulations.
438: The important point here is  rather that our model of polyalanine, 
439: despite its complexity, can be described in terms of a critical theory. 
440: 
441: 
442: 
443: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
444: \section{Conclusion}
445: \indent
446: We have extended the Yang-Lee zeros analysis to an all-atom model 
447: of polyalanine polymers of various chain lengths. 
448:  Our results were compared with 
449: the Zimm-Bragg model which is often used to describe helix-coil
450: transitions in Biopolymers.
451:   Our analyses of the Yang-Lee zeros for both models 
452:   give clear evidence for our claim that the helix-coil 
453:   transition in an all-atom model of polyalanine is not 
454:   adequately described by the Zimm-Bragg theory.
455:  Our results confirm
456: earlier work where  we pointed out the possibility of a thermodynamic
457: phase transition in polyalanine which is forbidden in the Zimm-Bragg model.
458: 
459: 
460: 
461: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
462: \vspace{0.4cm}
463: {\bf Acknowledgements}
464: \vspace{0.4cm}
465: 
466:  U. Hansmann gratefully acknowledges support by  a research grant 
467: of the National Science Foundation (CHE-9981874). This article was
468: finished while U.H. was visitor at the Department of Physics at
469: the Bielefeld University. He thanks F.~Karsch for his hospitality
470: during his stay in Bielefeld.
471: 
472: 
473: 
474: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
475: 
476: \begin{thebibliography}{99}
477: 
478: \bibitem{Poland} D.~Poland and H.A.~Scheraga, {\it Theory of Helix-Coil
479:          Transitions in Biopolymers} (Academic Press, New York, 1970).
480: 
481: \bibitem{ZB} B.H. Zimm and J.K. Bragg,  J. Chem. Phys. {\bf 31} (1959) 526. 
482: 
483: \bibitem{HO98c} U.H.E.~Hansmann and Y.~Okamoto,  J. Chem.~Phys. {\bf 110}
484:                 (1999) 1267. 
485: 
486: \bibitem{AH99b} N.A. Alves  and U.H.E. Hansmann, 
487:                 {Phys. Rev. Lett.}  {\bf 84} (2000) 1836.
488: 
489: \bibitem{KHC99d} J.P.~Kemp, U.H.E.~Hansmann and Zh.Y.~Chen,\ Eur.~Phy.~J.~B
490:               {\bf 15} (2000) 371. 
491: 
492: \bibitem{Yang-L} C.N. Yang and T.D. Lee, Phys. Rev. {\bf 87} (1952) 404.
493: 
494: \bibitem{Lee-Y} T.D. Lee and C.N. Yang, Phys. Rev. {\bf 87} (1952) 410. 
495: 
496: \bibitem{EC} M.J. Sippl, G. N{\'e}methy, and H.A. Scheraga,
497:              {J. Phys. Chem.} {\bf 88} (1984) 6231, and references therein.
498: 
499: \bibitem{Konf} H.~Kawai, Y.~Okamoto, M.~Fukugita, T.~Nakazawa, and
500:                T.~Kikuchi,   Chem. Lett. {\bf 1991} (1991) 213; 
501:                Y.~Okamoto, M.~Fukugita, T.~Nakazawa, and H.~Kawai,
502:                 Protein Engineering {\bf 4} (1991) 639. 
503: 
504: \bibitem{1dlr} F.J.~Dyson, Commun.~Math.~Phys. {\bf 12} (1996) 91; 
505:                J.F.~Nagle and J.C.~Bonner, J.~Phys.~C {\bf 3} (1970) 352; 
506:                E.~Bayong, H.T.~Diep and V.~Dotsenko, Phys.~Rev.~Lett. {\bf 83},
507:                (1999) 14.
508: 
509: \bibitem{MU} B.A. Berg  and T. Neuhaus,
510:              {Phys. Lett.} B {\bf 267} (1991) 249.
511: 
512: \bibitem{Review} U.H.E.~Hansmann and Y.~Okamoto, 
513:          in: Stauffer, D. (ed.) ``{\it Annual Reviews in Computational 
514:          Physics VI}''
515:          (Singapore: World Scientific), p.129 (1998).
516: 
517: \bibitem{HO} U.H.E. Hansmann and Y. Okamoto,  J.~Comp.~Chem.
518:              {\bf 14} (1993) 1333. 
519: 
520: \bibitem{OH95b} Y.~Okamoto and U.H.E.~Hansmann,\  J.~Phys.~Chem.
521:                 {\bf 99} (1995) 11276. 
522: 
523: \bibitem{FS}  A.M. Ferrenberg  and R.H. Swendsen, 
524:               {Phys. Rev. Lett.} {\bf  61} (1988) 2635;  
525:               {Phys. Rev. Lett.} {\bf 63} (1989)
526:               1658(E) and references given in the erratum.
527: 
528: \bibitem{HO94c}  U.H.E.~Hansmann and Y.~Okamoto, 
529:                  Physica A {\bf 212} (1994) 415.
530: 
531: \bibitem{Itzykson} C. Itzykson, R.B. Pearson and J.B. Zuber,  
532:                    {Nucl. Phys.} B {\bf 220} [FS8] (1983) 415.
533: 
534: \bibitem{BinderB30} K. Binder and D.P. Landau,
535:                     {Phys. Rev. B} {\bf 30} (1984) 1477.
536: 
537: \bibitem{ChallaB34} M.S.S. Challa, D.P. Landau and K. Binder, 
538:                     {Phys. Rev.} {\bf B34}  (1986)  1841. 
539: 
540: \bibitem{Lee-K90} J. Lee  and J.M. Kosterlitz,
541:                   {Phys. Rev. Lett.} {\bf  65} (1990) 137;
542:                   Phys. Rev. {\bf B43} (1991) 3265.
543: 
544: \bibitem{Katsura} S. Katsura and M. Ohminami,
545:                   {J. Phys. A} {\bf 5} (1972) 95.
546: 
547: \bibitem{Glumac94} Z. Glumac and K. Uzelac,
548:                    {J. Phys. A} {\bf 27} (1994) 7709.
549: 
550: \bibitem{Kim-Cres1} S.-Y. Kim  and R.J. Creswick, 
551:                     {Phys. Rev. Lett.} {\bf 81} (1998) 2000. 
552: 
553: \bibitem{Kim-Cres2} R.J. Creswick and S.-Y. Kim, 
554:                     {Phys. Rev. E} {\bf 56} (1997) 2418. 
555: 
556: \bibitem{LandauB39} P. Peczak and D.P. Landau,
557:                     {Phys. Rev. B} {\bf 39} (1989) 11932.
558: 
559: 
560: \end{thebibliography}
561: 
562: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
563: 
564: \newpage
565: {\huge Table:}\\
566: 
567: \noindent
568: % TABLE 1
569: \begin{table}[ht]
570: \renewcommand{\tablename}{Table}
571: \caption{\baselineskip=0.8cm  Critical temperatures $T^{hist}_c$ as
572: obtained by our analysis for polyalanine chain lengths $N$. For comparison,
573: we show also previous estimates obtained from both an analysis of
574: Fisher zeros ($T_c^{\rm Fisher}$) and the location of the peak in
575: the specific heat ($T_c^{C_v}$).}
576: \begin{center}
577: \begin{tabular}{l c c c c }\\
578: %%%\hline
579: \\[-0.3cm]
580: %~~$N$ &~~$\beta_c^{\rm \,hist}$ &~~$T_c^{\rm \,hist}$  
581: ~~$N$  &~~$T_c^{\rm \,hist}$  
582:       &~~$T_c^{\rm \,Fisher}$   &~~$T_c^{\, C_v}$ \\
583: \\[-0.35cm]
584: \hline
585: \\[-0.3cm]
586: %~~10 & ~$1.1617(11)$   & ~$433(1)$  & ~442(8)  & ~427(7) \\
587: %~~15 & ~$1.0136(48)$   & ~$496(3)$  & ~498(4)  & ~492(5) \\
588: %~~20 & ~$0.9717(65)$   & ~$518(4)$  & ~511(5)  & ~508(5) \\
589: %~~30 & ~$0.9713(44)$   & ~$518(3)$  & ~520(4)  & ~518(7) \\
590: ~~10 &  ~$433(1)$  & ~442(8)  & ~427(7) \\
591: ~~15 &  ~$496(3)$  & ~498(4)  & ~492(5) \\
592: ~~20 &  ~$518(4)$  & ~511(5)  & ~508(5) \\
593: ~~30 &  ~$518(3)$  & ~520(4)  & ~518(7) \\
594: %%%\hline\\
595: \end{tabular}
596: \end{center}
597: %%\end{tabular}}
598: \end{table}
599: \vfill
600: 
601: \newpage
602: 
603: {\huge Figure Captions:} \\
604: 
605: {\bf Figure 1.} Histogram for the number of helical residues in conformations
606: for chain length $N=30$. \\
607: 
608: {\bf Figure 2.} Distribution of complex partition function zeros in the
609: $y-$plane with error bars, at $T_c(N=30)$. 
610:  We draw an inexistent unit circle to show
611: how the zeros are distributed around it.
612:  The zeros do not circumscribe an unit circle as expected
613: for the Lee-Yang theorem in Ising like systems with positive interactions. \\
614: 
615: 
616: {\bf Figure 3.} The Yang-Lee zeros for the Ising chain with lengths 
617: $N=5, 15$ and 30. \\
618: 
619: {\bf Figure 4.} The behaviour of the partition function zeros for polyalanine
620: with chain length $N=10$ in function of the temperature $u(\beta)$. \\
621: 
622: {\bf Figure 5.} Linear regression for ${\rm ln}\,(\theta_0)$,
623:        Eq.~(\ref{eq:t0}), at the critical temperature. 
624: \vfill
625: 
626: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
627: 
628: \newpage
629: %\cleardoublepage
630: 
631: %FIGURE 1
632: \begin{figure}[t]
633: %\begin{figure}[!ht]
634: \begin{center}
635: \begin{minipage}[t]{0.95\textwidth}
636: \centering
637: %\includegraphics[angle=-90,width=0.72\textwidth]{hist30_01.eps}
638: \includegraphics[angle=-90,width=0.72\textwidth]{fig1.eps}
639: \renewcommand{\figurename}{Fig.}
640: \caption{Histogram for the number of helical residues in conformations
641: for chain length $N=30$.}
642: \label{fig1}
643: \end{minipage}
644: \end{center}
645: \end{figure}
646: \vfill
647: \newpage
648: %\cleardoublepage
649: 
650: %FIGURE 2
651: %\begin{figure}[t]
652: \begin{figure}[!ht]
653: \begin{center}
654: \begin{minipage}[t]{0.95\textwidth}
655: \centering
656: %\includegraphics[angle=-90,width=0.72\textwidth]{aver_y_30.eps}
657: \includegraphics[angle=-90,width=0.72\textwidth]{fig2.eps}
658: \renewcommand{\figurename}{Fig.}
659: \caption{Distribution of complex partition function zeros in the
660: $y-$plane with error bars, at $T_c(N=30)$. 
661:  We draw an inexistent unit circle to show
662: how the zeros are distributed around it.
663:  The zeros do not circumscribe an unit circle as expected
664: for the Lee-Yang theorem in Ising like systems with positive interactions.}
665: \label{fig2}
666: \end{minipage}
667: \end{center}
668: \end{figure}
669: \vfill
670: 
671: \newpage
672: \cleardoublepage
673: 
674: %FIGURE 3
675: \begin{figure}[t]
676: %\begin{figure}[!ht]
677: \begin{center}
678: \begin{minipage}[t]{0.95\textwidth}
679: \centering
680: %\includegraphics[angle=-90,width=0.72\textwidth]{fig_5_15_30.eps}
681: \includegraphics[angle=-90,width=0.72\textwidth]{fig3.eps}
682: \renewcommand{\figurename}{Fig.}
683: \caption{The Yang-Lee zeros for the Ising chain with lengths 
684: $N=5, 15$ and 30.}
685: \label{fig3}
686: \end{minipage}
687: \end{center}
688: \end{figure}
689: \vfill
690: 
691: \newpage
692: %\cleardoublepage
693: 
694: %FIGURE 4
695: %\begin{figure}[t]
696: \begin{figure}[!ht]
697: \begin{center}
698: \begin{minipage}[t]{0.95\textwidth}
699: \centering
700: %\includegraphics[angle=-90,width=0.72\textwidth]{zeros_N10.eps}
701: \includegraphics[angle=-90,width=0.72\textwidth]{fig4.eps}
702: \renewcommand{\figurename}{Fig.}
703: \caption{The behaviour of the partition function zeros for polyalanine
704: with chain length $N=10$ in function of the temperature $u(\beta)$.}
705: \label{fig4}
706: \end{minipage}
707: \end{center}
708: \end{figure}
709: \vfill
710: 
711: 
712: \newpage
713: \cleardoublepage
714: 
715: %FIGURE 5
716: %\begin{figure}[t]
717: \begin{figure}[!ht]
718: \begin{center}
719: \begin{minipage}[t]{0.95\textwidth}
720: \centering
721: \includegraphics[angle=-90,width=0.72\textwidth]{fig5.eps}
722: \renewcommand{\figurename}{Fig.}
723: \caption{Linear regression for ${\rm ln}\,(\theta_0)$,
724:        Eq.~(\ref{eq:t0}), at the critical temperature.}
725: \label{fig5}
726: \end{minipage}
727: \end{center}
728: \end{figure}
729: \vfill
730: 
731: 
732: \end{document}
733: 
734: 
735: 
736: 
737: 
738: 
739: 
740: 
741: