Overview

People of LSCP

Computer Facilities

Publications

Collaborators

Activities


ConferencesWorkshops ( 2000~2011 ) / Seminars and Group Meetings

¡@

2005 NCTS July Workshop on Macromolecules and Biological Systems

 

 

 

 

¡@

Date :

 27-29 July 2005

¡@

¡@

¡@

Place :

 Room 312, Department of Physics, National Taiwan University

¡@

¡@ Organized by :

 National Center for Theoretical Sciences (Critical Phenomena and Complex Systems focus group)

 Institute of Physics of Academia Sinica (Taipei)

¡@

¡@ ¡@
¡@ Contact Info. :

 Miss Chia-Chi Liu (Secretary, Physics Division, NCTS)
 Tel:(886)-2-33665566; Fax:(886)-2-33665565; E-mail: ccliu@phys.ntu.edu.tw

¡@ ¡@
¡@

          Speaker :¡@

Professor R. Abagyan

The Scripps Research Institute, 10550 N Torrey Pines, La Jolla 92037, USA

abagyan@scripps.edu

¡@ ¡@
¡@ ¡@       Title: Accurate prediction of partial protein structure by stochastic global optimization in internal coordinates ¡@ ¡@
¡@ ¡@ Abstract:

Despite many years of efforts to predict local protein structure at the accuracy level of crystallographic structures, the problem is not solved yet. Predicting conformations of protein loops is an important task in several areas, e.g.homology modeling and simulating the association-induced conformational changes. We introduced a new conformational sampling procedure based on the original stochastic global optimization method in internal coordinates which employed concerted movements of groups of coupled variables and a certain search strategy. The new modification lowered the energy barriers for sampling of the loop conformations. This method predicted loops of up to 15 residue long with high crystallographic resolution accuracy (under 0.5 for the backbone atoms and under 1A for the side chains). The protocol was tested on a benchmark of protein loops. This result for the first time establishes that if and when the simulation runs to full convergence, the prediction will be accurate in most cases.

¡@ ¡@
¡@
¡@ ¡@

Professor Iksoo Chang

Department of Physics, Pusan National University, 30, Jangjeon-Dong, Gumjeong-Gu, Busan, KOREA

chang@random.phys.pusan.ac.kr

¡@ ¡@
¡@ ¡@       Title:

Is RMSD-criterion good enough for an identification of native structures form structural ensembles of protein in NMR experiment? 

(by Suhkmann Kim, Muyoung Heo, Mookyung Cheon, Kwanghoon Chung, Eun-Joung Moon, Haejin Kim, and Iksoo Chang)

¡@ ¡@
¡@ ¡@ Abstract:

NMR experiment for resolving a protein structure provides structural ensembles which have a different feature compared to a unique one from X-ray experiment. Currently, the conventional way to select a good candidate for native structure from NMR relies on choosing structures with a smaller (and minimal) RMSD value in the ensembles of protein structures, which suppose to result in the good clustering of structures for the native state. We, however, show several examples where this is not always true, and introduce a thermodynamic criterion to select a set of good native structures in NMR experiment. We provide a systematic and physical approach based on the equilibrium statistical mechanics to verify a set of good and bad ensemble structures for a native state of protein from NMR. Using energy parameters obtained from perceptron learning and protein threading of 1006 representative proteins, we develop an approximate partition function to calculate the unfolded fraction, the energy, and the specific heat as a function of temperature for each ensemble structure. We show how one can systematically determine/cluster a set of good/bad native structures from ensembles of protein structure in NMR experiment, which then can be compared to that from X-ray experiment. Our result suggests that the thermodynamic quality for a set of native structures from NMR is equivalent to and no worse than that from X-ray experiment. Having tested on various proteins for their structural ensembles from NMR, we conclude that RMSD-criterion is not always consistent nor thermodynamic to provide a proper native-state structure for protein. We suggest to employ a new approach based the statistical mechanics to determine/cluster a set of native structures.

¡@

¡@ ¡@
¡@ ¡@       Title:

Identification of the elusive transition state in the protein folding kinetics

(by Iksoo Chang, Marek Cieplak, Jayanth R. Banavar, and Amos Maritan)

¡@ ¡@
¡@ ¡@ Abstract:

Protein Engineering has offered a new tool to probe properties of proteins. This method involves substituting amino acids in different positions of a protein with other amino acids and monitoring the resulting changes in the folding and unfolding rates. Another way is to probe the effect of a denaturant on folding/unfolding kinetics through a Chevron plot. The effect of these is characterized by the folding/unfolding phi values of the transition state, which are measures of the changes in the kinetics as determined relative to the changes in the stability. We present a kinetic analysis for the notion of the transition state for a simple model of proteins using a master equation, and also provide the physical meaning of phi value determined in the protein engineering experiment. We benchmark our findings to various theoretical approaches and interpretations for the identification and characterization of the elusive transition state.

¡@

¡@ ¡@
¡@

Doctor Kwanghoon Chung

National Research Laboratory for Computational Proteomics and Biophysics, Department of Physics, Pusan National University, Busan 609-735, KOREA

av15@pusan.ac.kr  

¡@ ¡@       Title: Folding and Unfolding Kinetics of CI2 and BBL Proteins : Monte Carlo Simulation using Munoz-Eaton Energy Function  (by Kwanghoon Chung, Wookyung Yu, Mookyung Cheon, Muyoung Heo, and Iksoo Chang)
¡@
¡@ ¡@
¡@ ¡@ Abstract:

Folding and unfolding behavior of 3CI2 and 1BBL protein is studied by Munoz-Eaton energy function. After simplifying the state of amino acid as a native state or non-native state and transforming the conformational ensembles of a protein to the configurations of spin variable in Ising model, we performed Monte Carlo simulation for a short time kinetics to keep track a time evolution Pk(t), Pr(t) of the onset for a nativeness of k-th pairwise-contact and r-th residue in 3CI2 and 1BBL protein. We also calculated the native state probability Peq(k), Peq(r) as a function of k and r by the equilibrium Monte Carlo simulation. Folding and unfolding events of 3CI2 and 1BBL proteins are described by Pk(t), Pr(t) and are explained by the concept of the local contact order of pairwise-contact along the sequence of amino acids and in terms of the entropic barrier which is evident from the topology of a protein. Also the differences emerged in thermodynamics and kinetics between 3CI2 and 1BBL protein are discussed, and the free energy landscape distinguishes the 2-state folding(3CI2) from the barrierless folding(1BBL).

¡@ ¡@
¡@
¡@ ¡@

Professor Michael W. Deem

John W. Cox Professor, Rice University, 6100 Main Street - MS 142, Houston, TX 77005-1892, USA

mwdeem@king.rice.edu

¡@ ¡@
¡@ ¡@       Title: Quantifying Influenza Vaccine Efficacy and Antigenic Distance ¡@ ¡@
¡@ ¡@ Abstract:

The adaptive vertebrate immune system is a wonder of modern evolution. Under most circumstances, the dynamics of the immune system is well-matched to the dynamics of pathogen growth during a typical infection.  Some pathogens, however, have evolved escape mechanisms that interact in subtle ways with the immune system dynamics.  In addition, negative interactions the immune system, which has evolved over 400 000 000 years, and vaccination, which has been practiced extensively for only 200 years, are possible.  For example, vaccination against the flu can actually increase susceptibility to the flu in the next year. 

In this talk, I present a physical theory of original antigenic sin and immunodominance.  How localization in the immune system leads to the observed phenomena is discussed.

¡@ ¡@
¡@ ¡@ ¡@ ¡@ ¡@
¡@ ¡@       Title: Evolvability is a Selectable Trait ¡@ ¡@
¡@ ¡@ Abstract:

Concomitant with the evolution of biological diversity must have been the evolution of mechanisms that facilitate evolution, due to the essentially infinite complexity of protein sequence space. I describe how evolvability can be an object of Darwinian selection, emphasizing the collective nature of the process. I quantify our theory with computer simulations of protein evolution. These simulations demonstrate that rapid or dramatic environmental change leads to selection for greater evolvability. The selective pressure for large scale genetic moves, such as DNA exchange, becomes increasingly strong as the environmental conditions become more uncertain. These results demonstrate that evolvability is a selectable trait and allow for the explanation of a large body of experimental results.

¡@ ¡@
¡@ ¡@ ¡@ ¡@ ¡@
¡@ ¡@       Title: Glassy Dynamics in the Adaptive Immune Response Prevents Autoimmune Disease ¡@ ¡@
¡@ ¡@ Abstract:

The immune system normally protects the human host against death by infection. However, when an immune response is mistakenly directed at self antigens, autoimmune disease can occur. I describe a model of protein evolution to simulate the dynamics of the adaptive immune response to antigens. Computer simulations of the dynamics of antibody evolution show that different evolutionary mechanisms, namely gene segment swapping and point mutation, lead to different evolved antibody binding affinities. Although a combination of gene segment swapping and point mutation can yield a greater affinity to a specific antigen than point mutation alone, the antibodies so evolved are highly cross-reactive and would cause autoimmune disease, and this is not the chosen dynamics of the immune system. I suggest that in the immune system a balance has evolved in the mechanism for searching the amino acid sequence space of antibodies between binding affinity and specificity.

¡@ ¡@
¡@

Doctor Muyoung Heo

National Research Laboratory for Computational Proteomics and Biophysics, Department of Physics, Pusan National University, Busan 609-735, KOREA

myheo@pusan.ac.kr  

¡@ ¡@       Title:

Perceptron learning of the pairwise-contact energies of proteins incorporating amino acids' environment

(by Muyoung Heo, Suhkmann Kim, Eun-Joung Moon, Mookyung Cheon, Kwanghoon Chung, and Iksoo Chang)
¡@

¡@ ¡@
¡@ ¡@ Abstract:

Although a coarse-grained description of proteins is a simple and convenient way to attack the protein folding problem, the construction of a global pairwise energy function which can simultaneously recognize the native folds of many proteins has resulted in partial success. We have sought the possibility of a systematic improvement of this pairwise-contact energy function as we extended the parameter space of amino acids, incorporating local environments of amino acids, beyond a 20x20 matrix. We have studied the pairwise-contact energy functions of 20x20, 60x60, and 180x180 matrices depending on the extent of parameter space, and compared their effect on the learnability of energy parameters in the context of a gapless threading, bearing in mind that a 20x20 pairwise-contact matrix has been shown to be too simple to recognize the native folds of many proteins.

In this paper we show that the construction of a global pairwise energy function was achieved using 1,006 training proteins of a homology of less than 30% which include all representatives of different protein classes. After parameterizing the local environments of the amino acids into 9 categories depending on three secondary structures and three kinds of hydrophobicity (desolvation), the 16,290 pairwise-contact energies (scores) of the amino acids could be determined by perceptron learning and protein threading. These could simultaneously recognize all the native folds of the 1,006 training proteins.

When these energy parameters were tested on the 382 test proteins of a homology of less than 90%, 370 (96.9%) proteins could recognize their native folds. We set up a simple thermodynamic framework in the conformational space of decoys to calculate the unfolded fraction and the specific heat of real proteins. The different thermodynamic stabilities of E. coli ribonuclease H (RNase H) and its mutants were well described in our calculation, agreeing with the experiment.
¡@

¡@ ¡@
¡@

Doctor Hsiao-Ping Hsu

John von Neumann Institute for Computing, Forschungszentrum Juelich D-52425, Juelich, GERMANY

h.p.hsu@fz-juelich.de

¡@ ¡@       Title: Simulations of Star Polymers: Scaling of Single Star Polymers and Effective Interactions between Two Star Polymers ¡@ ¡@
¡@ ¡@ Abstract:

Three-dimensional star polymers are studied by using a very efficient chain growth algorithm with resampling, the pruned-enriched Rosenbluth method (PERM). Our simulations are performed for the soft core Domb-Joyce model on the simple cubic lattice, to minimize corrections to scaling and to allow for an unlimited number of arms.In this model polymers are described by lattice random walks where monomers are located on lattice sites. Multiple visits to the same site are allowed, but a self-repulsion with Boltzmann factor v<1 is taken into account. In our work, we choose the 'magic' value v=0.6 and simulate 3-dimensional single star polymers with up to f=80 arms and with the number of monomers per arm up to 4000 for small values of f. PERM is modified such that all arms grow simultaneous. We measure the swelling of single star polymers, i.e. the center-to-end distance and gyration radius, and also estimate the partition sum directly with very high precision, giving very precise estimates of the critical exponents gf [1]. The effective pair potential between two star polymers with equal arm lengths and equal number f of arms are also studied numerically. We find that the potential is much less soft at large r than claimed in previous papers, in particular for f>>1. While we verify the logarithmic divergence of V(r), with r being the distance between the two cores, predicted by Witten and Pincus, we find for f>20 that the Mayer function is hardly distinguishable from that for a Gaussian potential [2].

¡@

References:

[1] Hsiao-Ping Hsu, Walter Nadler, and Peter Grassberger, Scaling of Star Polymers with one to 80 Arms, Macromolecules 37,

     4658 (2004).
[2] Hsiao-Ping Hsu and Peter Grassberger,
Effective Interactions between Star Polymers, Europhysics Lett. 66, 874 (2004).

¡@

¡@ ¡@
¡@ ¡@ ¡@ ¡@ ¡@
¡@ ¡@

Professor Ching-Hwa Kiang

Department of Physics and Astronomy, Rice University, 6100 Main Street - MS 61, Houston, TX 77005-1892, USA

chkiang@rice.edu

¡@ ¡@
¡@ ¡@       Title:

Protein Folding Pathways Studied with Single-Molecule Atomic Force Microscopy

¡@ ¡@
¡@ ¡@ Abstract:

Atomic force spectroscopy has been used to probe the mechanical properties of proteins at the single-molecule level. Forced protein unfolding provides valuable information about protein stability as well as the unfolding pathways of these molecules. However, questions remains over how pulling data be properly interpreted and compared to bulk measurements. We studied forced unfolding of the I27 domain of the protein titin under varying solution conditions, such as exposure to various concentrations of chemical denaturant, and at different pulling speeds in the hope to resolve the question of whether mechanical and chemical unfolding are directly comparable.

¡@ ¡@
¡@ ¡@ ¡@ ¡@ ¡@
¡@ ¡@

Professor Jung-Hsin Lin

School of Pharmacy, National Taiwan University, Taipei 100 & Institute of Biomedical Sciences, Academia Sinica, Taipei 115, TAIWAN

jlin@rx.mc.ntu.edu.tw

¡@

¡@ ¡@
¡@ ¡@       Title:

Dynamics, Conformational Transition, and Drug Interactions of P-glycoprotein upon ATP-binding

¡@ ¡@
¡@ ¡@ Abstract:

The overexpression of P-glycoprotein (Pgp) is one of the major causes of multidrug resistance (MDR) in cancer chemotherapies. Until now, it remains very challenging to obtain the high resolution Pgp structure and currently only low resolution electron microscopy structure is available. The recent determination of the X-ray crystallographic structure of a bacterial lipid A transporter, MsbA, has provided good structure templates for homology modeling, and in our lab this model is used to further construct a full-length Pgp structure.

We have conducted molecular dynamics (MD) simulations of the full-length Pgp in an excessively hydrated POPC bilayer. The whole system was simulated in an environment closer to physiological condition with the salt concentration of 0.15 M. Both free and ATP-bound forms of Pgp were simulated. The total simulation time was 9 ns for both simulations.

It has been shown from our MD simulations that the overall architecture of the Pgp is stable in a realistic lipid bilayer environment, and the simulation results have allowed us to investigate the conformational changes to Pgp upon ATP binding in the efflux process. Not only the binding of ATP indeed drives the conformational changes in the transmembrane domains (TMDs), but also the helix rearrangement reduce the space of TMD center pore to mediate drug transportation. The mechanistic details of the transport cycle upon ATP binding will be helpful for interpretating remote connections between the ATP binding and the subsequent release of substrates. The refined structure model of Pgp by our MD simulations may be used as the basis for designing Pgp inhibitors or "Pgp-ignoring" drugs.

¡@

¡@ ¡@
¡@
¡@ ¡@

Doctor Phuong H. Nguyen

Institute of Physical and Theoretical Chemistry, J. W. Goethe University, Marie-Curie-Str. 11, D-60439 Frankfurt, GERMANY

phuong@theochem.uni-frankfurt.de

¡@ ¡@
¡@ ¡@       Title: Nonequilibrium Dynamics in Biomolecules ¡@ ¡@
¡@ ¡@ Abstract:

Femtosecond experiments on photoswitchable peptides provide a new and promising way to study the folding and unfolding of biomolecules in real time and unprecedented detail. To obtain an appropriate theoretical description of these experiments, we suggest to extend well-established molecular dynamics simulation techniques to the description of photoinduced conformational dynamics in peptides. The goal is to perform true nonequilibrium molecular dynamics simulations in which the laser-induced initial state of the molecule is represented by a suitable semiclassical phase-space distribution, from which initial conditions of nonequilibrium trajectories are sampled. Performing a time-dependent ensemble average over these trajectories, the observables of interest are calculated, including the time-dependent transfer of vibrational energy as well as the photoinduced conformational rearrangements of the peptide. The results are direct compare to experiment.

¡@ ¡@
¡@ ¡@ ¡@ ¡@ ¡@

Doctor Ruben Ghulghazaryan

Institute of Physics, Academia Sinica, Nankang, Taipei 11529, TAIWAN

ghulr@phys.sinica.edu.tw

¡@ ¡@       Title:

Monte Carlo Simulation of Small Peptides with Wang-Landau Algorithm

(by R. G. Ghulghazaryan, Sh. Hayryan, Chin-Kun Hu)

¡@

¡@ ¡@
¡@ ¡@ Abstract:

The knowledge of the density of states solves the problem of studying the thermodynamic properties of the system. The partition function of the system, the free energy, specific heat and other thermodynamic quantities may be computed from the density of states.


Recently Wang and Landau proposed a Monte Carlo algorithm (WL algorithm) for numerical calculation of the density of states [1,2]. The algorithm performs a random walk in energy space by accepting new states with probability inversely proportional to the density of states. The advantage of WL algorithm is that it updates the density of states at each Monte Carlo step, instead of accumulating histogram and then updating the density of states. Initially, the WL algorithm was designed for discrete energy spectrum systems and was successfully applied to the 2D Ising, Potts and Edwards-Anderson models [2].

In this work we use WL algorithm for simulation of small peptides. As an example we studied thermodynamical and structural properties of 13-residue synthetic peptide, P1, from the C-terminal
a-helix part of the human Prion Protein (PrP) [4], 16-residue alanine polypeptide and other.

All-atom energy potential function is used for simulation of peptides in a vacuum and in aqueous solution with normal PH factor, PH=7, using Solvent Accessible Surface Area method for calculation of peptide-solvent interaction energy, as implemented in SMMP package [3] Temperature dependencies of thermodynamic quantities such as average energy, specific heat, relative free energy and entropy are calculated from the density of states.

Our results shows that WL algorithm may be successfully applied for finding density of states of small peptides. Nevertheless, the WL algorithm badly converges to the true density of states for large peptides and freeze errors in the density of states.


References:

[1] F. G. Wang, D. P. Landau, Phys. Rev. Lett. 86, 2050 (2001).
[2] F. G. Wang, D. P. Landau, Phys. Rev. E 64, 056101 (2001).
[3] F. Eisenmenger, U. H.E. Hansmann, Sh. Hayryan, Chin-Kun Hu, Comp. Phys. Comm. 138, 192 (2001).
[4] R. G. Ghulghazaryan, Sh. Hayryan, Chin-Kun Hu, N. Poklar Ulrih, Monte Carlo Simulation of Peptide Fragment PrP

     (214-226) of Cellular Human Prion Protein (PrP), in preparation.
¡@

¡@ ¡@

Professor Chung-Kang Peng

Beth Israel Deaconess Medical Center / Harvard Medical School, USA

peng@physionet.org

¡@

¡@ ¡@       Title: Measuring Complexity of Complex Biological Systems
¡@
¡@ ¡@
¡@ ¡@ Abstract:

One of the great challenges of contemporary biomedical science is to understand more fully the dynamics of living systems in health and disease. The importance of this challenge is highlighted by headlines announcing unexpected, life-threatening side effects of once-promising drugs, as well as the serendipitous discoveries deriving from ¡§outside the box¡¨ approaches to major public health problems, for example, in heart disease. The basis of such unexpected findings, both negative and positive, is the extraordinary complexity of physiologic systems, which exceeds that of the most challenging systems in the physical world. These systems defy understanding based on traditional mechanistic models and conventional biostatistical analyses.

Central to this enterprise are quantitative measurements that best reflect the emergent properties of the integrative system. To identify system-level behaviors that are critical to our understanding of a healthy system and its pathological perturbations, the following hypotheses are necessary: i) The complexity of a biological system reflects its ability to adapt and function in an ever-changing environment. ii) Biological systems need to operate across multiple spatial and temporal scales, and hence their complexity is also multi-scale. iii) A wide class of disease states, as well as aging, appear to degrade biological complexity and reduce the adaptive capacity of the system. Thus, loss of complexity may be a generic feature of pathologic dynamics.

Recently, we developed a multiscale information approach to address this challenge. Traditional complexity measurements are often based on the concept of entropy, which quantifies the regularity (orderliness) of fluctuating signals. Entropy increases with the degree of disorder and is maximal for completely random systems. However, an increase in entropy may not always be associated with an increase in dynamical complexity. For instance, a randomized (scrambled) time series has higher entropy than the original time series, although the process of scrambling the data destroys correlations and degrades the meaningful information content of the original signal. We recently showed that this inconsistency is due to the fact that widely used entropy measures are based on single-scale analyses. Instead, biological systems operate on a wide range of temporal and spatial scales. This multiscale, hierarchical feature is critical to biological systems and needs to be taken into account in the analysis of complex systems.

To illustrate the basic concept underlying the multiscale information approach, I will apply this quantitative measurement to several biological time series.

¡@

¡@ ¡@

Doctor O. Rozanova

Faculty of mathematics and mechanics, Moscow State University, RUSSIA

rozanova@mech.math.msu.ru

¡@ ¡@       Title: Solution of a population genetics model with two wild alleles: the basin of attractions for different steady states
¡@
¡@ ¡@
¡@ ¡@ Abstract:

We solve exactly the diploid evolution model with two wild alleles. The steady state depends from the original distribution of alleles. We calculate the basins of attractions and error thresholds. The selective ability of different peaks is defined for the diploid evolution model.
¡@

¡@ ¡@

Doctor David B. Saakian

Theoretical Physics Division, Yerevan Physics Institute, ARMENIA

saakian@phys.sinica.edu.tw

¡@ ¡@       Title: Exact solution of Diploid evolution model with a single peak fitness ¡@ ¡@
¡@ ¡@ Abstract:

We solve exactly both the dynamics and statics for the model with single peak fitness function, suggested by Wiehe, Baake and Schuster (WBS).  For the dominance of wild allele we rigorously derive the error threshold given by  WBS. For the dominance of defective alleles there is a bi-stability and epidemic threshold phenomenon: if original concentration of defective alleles is higher than the threshold value, the wild allele disappears during the evolution. The threshold phenomenon has been found also for the case of flat peak fitness and in a model suggested by J. Clark. We assume that it is a general theorem for the population genetics.

¡@ ¡@
¡@

Doctor Jie-Jun Tseng

Institute of Physics, Academia Sinica, Nankang, Taipei 11529, TAIWAN

gen@phys.sinica.edu.tw  

¡@ ¡@       Title:

Reconstructing Gene Regulatory Networks from Time-Series Microarray Data

(by S.P. Li, J.J. Tseng and S.C. Wang)
¡@

¡@ ¡@
¡@ ¡@ Abstract:

A gene regulatory network depicts which genes turn on which and at what moment. Knowledge of such gene networks is key to an understanding of the biological process. We propose here to use a statistical method for the reconstruction of gene regulatory networks based on Bayesian networks from microarray data. We describe a nonlinear model for the rate of gene transcription in which levels of gene expression are continuous. The reconstruction becomes an optimization problem where optimization algorithms are employed to search for optimal solutions. We apply the methodology to reconstruct the regulatory network of 41 yeast cell cycle genes from a real microarray data set. The result obtained is promising: more than 70% (31 out of 43 arcs) of the reconstructed regulations are consistent with experimental findings.

¡@ ¡@

         


[Academia Sinica]  [Institute of Physics]  [Library of the Institute of Physics]  [Electronic Journals]  [Related Links]

Laboratory of Statistical and Computational Physics