Agent-based modelling for molecular self-organization

Sara Fortuna* and Alessandro Troisi

University of Warwick

If we consider chemical systems, such as the crystal structure [1] or the supramolecular assembly [2] of an ensemble of molecules, a common question will arise: given N molecules, what is the lowest-energy organized structure that they can form?

This problem can be solved with the decomposition of the system in N agents. An agent is a system capable of exchanging information with other agents and its environment, taking decisions and performing autonomous actions.

The agent-based technique, previously used to study social phenomena, was first applied to a chemical system by Troisi et al. [3]. This technique allows a system of rigid shapes to evolve to the lowest-energy ordered structure on a 2-D lattice following a combination of stochastic, deterministic and adaptive rules with less computational effort than comparable Monte Carlo simulations.

The aim of our research is to extend the model to the study of the multiscale assembly of realistic systems [1,2]. We therefore extended our model to off-lattice systems, comparing its performance with a standard Monte Carlo simulation.

[1] G.M.Day, W.D.S.Motherwell, H.L.Ammon, et al., Acta Crystall.B 61, 511 (2005)

[2] Y.Kato, N.Mizoshita and K.Kishimoto, Angew.Chem.Int. 45, 38 (2006)

[3] A.Troisi, V.Wong and M.A.Ratner, PNAS 102, 255 (2005).

Molecular dynamics simulation of the mesophase behaviour of a model bolaamphiphilic liquid crystal with a lateral flexible chain

Andrew J. Crane, Francisco Martinez-Veracoechea, Fernando Escobedo and Erich A. Müller

Department of Chemical Engineering and Chemical Technology, Imperial College London

We report coarse-grained molecular dynamic simulations of a model bolamphiphile molecule with a grafted flexible side chain, inspired by the recently reported synthesis and characterisation of a series of T-shaped polyphilic liquid crystal molecules. The model consists of six fused rigid particles, where the end particles have strong self-attraction, the lateral chain particles have weak self-attraction, and the remaining particle pairs are softly repulsive. A parametric study of the phases of these molecules at temperatures in the fluid range up to the onset of crystallization was made. The molecules with no side chain exhibit a smectic liquid crystal behaviour. Upon increasing the number of particles in the lateral chains from 1 to 12, the liquid system exhibits a rich variety of self-assembled structures; columnar arrays of different cross sections (triangular, square, pentagonal, hexagonal, rectangular) and for longer chains lamellar structures of different interlayer spacing. We showcase the global phase behaviour of the model, determining order parameters and apparent diffusivities of the several regimes encountered.

Visualisation Techniques for Molecular Dynamics and Related Computer Simulations

M. Robinson, R. Smith and S. D. Kenny

Loughborough University

Extracting useful and relevant information from masses of raw data is a vital requirement when conducting computer simulation. This is undeniably true for simulations based on the atomic scale, where fundamental mechanisms can be overshadowed by the sheer quantity of outputted data. This work highlights various visualisation techniques that enable in depth and intuitive analysis through full 3D rendering and user-based interactions. Numerous examples are presented that demonstrate the flexibility and applicability of the visualisation suite, reiterating its importance as an essential tool alongside simulation.

Nanoscale fluid transport and the slip problem

V. Sokhan and N. Quirke

National Physical Laboratory, Teddington, U.K. TW11 0LW

We present the results of our recent work on calculation of transport properties of fluids in nanoscale flow including the expression for the slip coefficient applicable at this scale. Our results show that for wide pores in thermodynamic equilibrium with the bulk reservoir the slip coefficient is independent of the Knudsen number, in contrast to the linear scaling predicted by Maxwell's theory of slip.

Monte Carlo Simulations of Polymers

Adam Swetnam

University of Warwick

Very simple models are sufficient to capture the behaviour of polymer molecules. This allows advanced Monte Carlo techniques such as the Wang-Landau algorithm to be applied. This algorithm determines the density of states of a system, which allows its properties to be determined at all temperatures. Particularly useful is the specific heat capacity as peaks in this indicate phase transitions. For polymers near surfaces the density of states can be determined as a function of self interaction energy and surface interaction energy, from this the specific heat capacity can be determined as a function of two variables which allows a phase diagram to be plotted.

Coarse grained MD simulations of transmembrane peptides in a model raft lipid membrane.

Djurre H.de Jong, Lars V. Schäfer, Siewert J. Marrink

Rijks Universiteit Groningen

There is more and more evidence for a large degree of lateral heterogeneity in biological membranes, both from experiments in vitro and in vivo.

Accordingly, membrane rafts, that is, small and highly dynamic yet distinct regions in the membrane are supposed to play important roles for a number of biological processes, such as signalling as well as membrane protein structure and clustering.

Here we present the results from multi-microsecond coarse grained molecular dynamics simulations of a number of alpha-helical WALP transmembrane peptides in a model raft that consists of a liquid-disordered (Ld) and a liquid-ordered (Lo) domain.

We compare both, WALP23 and WALP31 to study low and high hydrophobic mismatch conditions, respectively. The peptides leave the Lo domain and prefer the Ld domain, irrespective of the hydrophobic mismatch. This agrees well with recent optical microscopy experiments on model vesicles.

Once in the Ld phase, the WALPs form clusters of different size: WALP23 forms predominantly dimers and trimers, which re-organize and exchange individual peptides in a dynamic equilibrium. In contrast, the large hydrophobic mismatch of WALP31 leads to the formation of one stable peptide cluster.

Our simulations yield detailed insights into the structural dynamics of the lateral organization of transmembrane peptides at near-atomic resolution.

Simulation of microstructure and charge transport in polymer semiconductors

David L. Cheung, David P. McMahon, and Alessandro Troisi

University of Warwick

Polymer semiconductors are becoming increasingly investigated as materials for a disparate range of electronic applications. The ability of polymer semiconductors to self-assemble into a range of ordered structures makes them significantly easier to process than traditional solid-state semiconductors and lends them to applications such as lighting, displays, and photovoltaics which require low-cost, large-area applications. The mechanism of charge transport and the relation between charge carrier motion and polymer microstructure and morphology are, however, poorly understood [1] and this limits our ability to develop novel materials and devices.

This poster reports a combined classical/quantum investigation of poly(3-hexylthiophene) [P3HT], the de facto standard polymer semiconductor. Atomistic molecular dynamics simulations, performed at a range of temperatures, have been performed to characterise the microstructure of the polymer. Quantum chemical calculations have been used to calculate the charge transfer integrals [2], which have play a central role in the theoretical description of charge transport using Variable Range Hopping and similar models, and to study the localisation of charge carriers [3], with the correlation between the location of traps and specific structural features in the polymer being investigated.

[1] David L. Cheung and Alessandro Troisi, Phys. Chem. Chem. Phys.,10, 5941 (2008)

[2] David L. Cheung, David P. McMahon, and Alessandro Troisi, J. Phys. Chem. B., submitted

[3] David L. Cheung, David P. McMahon, and Alessandro Troisi, J. Am. Chem. Soc., submitted

Describing of mechanical properties of graphene based on particle dynamics models

I. E. Berinskiy, A. M. Krivtsov

Institute for Problems in Mechanical Engineering, Russian Academy of Sciences

The purpose of this investigation is to construct a stable model of graphene (monolayer of graphite) using microstructure analysis and to connect parameters of this model with the macro parameters such as Young modulus and Poisson ratio. It is well-known that usage of pair potentials usually leads to the instability of the crystal lattice because of the lack of transverse stiffness in C-C bonds. There are three models of graphene considered in this work that contain additional stiffnesses in the potentials of interaction. The first one is a force model. It includes the forces interacting between the atoms of the first coordination sphere (nearest neighbours) and the atoms situated on the second coordination sphere (far neighbours). The second model includes a moment interaction between the carbon atoms in addition to the force interaction. The third model contains the terms describing the central and angular interaction. The relations between the model parameters and macro characteristics are found. Parameters of potential are determined using experimental data. Comparison with Tersoff-Brenner and AMBER force fields is provided. It is planned to use the models considered above for description of the deformation processes in ghaphene layers.

Confinement Effects on the Kinetics of the Isomerization of n-butane

S.J. Brookes^(a), D.J. Searles^(a) and K.P. Travis^(b)

Griffith University, Nathan QLD 4111, Australia^(a) University of Sheffield, Sheffield S1 3JD, UK^(b)

We examined the effect of confinement on the torsional kinetics in n-butane by conducting equilibrium Molecular Dynamics simulations of the adsorbed phase in a slit-pore geometry. Torsional rate constants were obtained with high precision using the powerful relaxation function method devised by Brown and Clarke [D. Brown and J. H. R. Clarke, J. Chem. Phys. 92, 3062 (1990)].

Values of the three rate constants; k(GT), k(TG) and k(GG) were determined as a function of temperature, density and porewidth for 2 different slit pore models: one employing explicit moving atoms, the other employing a smooth potential, both models being simulated at 2 different wall surface densities. We observed that direct gauche-gauche transitions taking place via the trans well (i.e. double well jumps) are significant under all conditions. All rate constants show a linear variation with temperature when plotted in Arrhenius form but the pre-exponential factor has very different density dependence for the double well kinetics compared with the single well transitions (monotone decreasing and monotone increasing respectively in the bulk fluid). We also find clear differences between behaviour of the kinetics under confinement compared with the bulk condensed phase results with k(GG) generally higher in confined system than the corresponding bulk systems.

Thermo-mechanical effects in perfect crystals with arbitrary multibody potential

Vitaly Kuzkin, Anton Krivtsov

Institute for Problems in Mechanical Engineering RAS

Derivation of equivalent thermo-mechanical parameters for perfect crystals in the case of arbitrary interatomic potential is conducted. The approach based on the averaging of equations of motion is considered. Long wave approximation is used to make link between the discrete system and equivalent continuum. Macroscopic thermo-mechanical parameters such as Piola and Cauchy stress tensors, heat flux are represented via microscopic parameters. Expansion of the expression for heat flux into series with respect to small thermal parameter is conducted. Connection between the heat flux and temperature is discussed. The influence of thermal motion on wave propagation is investigated. Nonlinear wave equation in adiabatic assumption is obtained. Linearization of this equation is considered. Equation of propagation of linear waves in uniformly deformed crystal which takes thermal motion into account is obtained.

Structure and Dynamics of Anti-p17 Single Chain Fv

Vannajan Sanghiran Lee¹, Panthip Tue-ngeun¹, Kuntida Kitidee², Sawitree Nangola², Jitrayut Jitonom¹, Piyarat Nimmanpipug¹, Supat Jiranusornkul³, and Chatchai Tayapiwatana^{2, 4}

¹Computational Simulation and Modeling Laboratory (CSML), Department of Chemistry and Center for Innovation in Chemistry, Faculty of Science; ²Division of Clinical Immunology, Department of Medical Technology, Faculty of Associated Medical Sciences; ³Department of Pharmaceutical Sciences, Faculty of Pharmacy, ⁴Biomedical Technology Research Unit, National Center for Genetic Engineering and Biotechnology, National Science and Technology Development Agency at the Faculty of Associated Medical Sciences, Chiang Mai University, Chiang Mai, 50200, Thailand

The potential applications of antibody fragment scFv have been reported in many fields including research tools, diagnostics and therapy. In this study, the model of scFv anti-p17 scfv structure were built from there x-ray structure homolog and validated. The complex structures between the single chain Fv antibody and its natural peptide epitopes at c-terminal of p17 were then constructed by flexible docking algorithm using potential mean force (PMF) scoring. The selected complex models were further investigated through molecular dynamics simulations. Molecular Mechanics Poisson-Boltzmann Surface Area (MM-PBSA) methodology and decomposition energy were calculated using the set of snapshots. The essential dynamics and the structural analysis of the antibody-antigen interface with the different functional behaviors between the wild-type and its mutants have been reported in detail. These results may give useful implications for structure-based design of antibody combining site.

Theoretical Modelling of Hydrogen-Bonded Supramolecular Networks

N. Martsinovich, L. Kantorovich, A. Troisi

Department of Chemistry, University of Warwick and Department of Physics, King's College London

The phenomenon of self-assembly of molecular superstructures is attracting an increasing interest in nanotechnology due to possible applications in materials science and molecular electronics. Self-assembly of molecules is controlled by specific intermolecular interactions. In particular, hydrogen bonds are among the strongest intermolecular interactions, they are directional and are known to give rise to complex supramolecular networks on a variety of substrates.

This abstract presents theoretical studies of self-assembly of organic molecules, specifically, benzenedicarboxylic acids and cyanuric acid, into hydrogen-bonded supramolecular structures. Density functional theory is used to explore the stabilities of a variety of possible structures formed by cyanuric acid and its sulfur analogue, trithiocyanuric acid. Force field molecular dynamics and Monte Carlo calculations are used to model the formation of two-dimensional networks of benzenedicarboxylic acids from starting disordered structures.

Modelling of the interactions between peptides and graphitic surfaces using a polarisable force-field: The role of mutations

Susana de Miranda Tomasio and Tiff R. Walsh

University of Warwick

The affinity of several peptide sequences interacting with graphitic surfaces is investigated using molecular dynamics (MD) simulations with a polarisable force-field[1]. Phage-display experiments have demonstrated the important role of tryptophan (W) in the strong affinity to graphitic surfaces[2]. Our previous work recovered these observations[1]. Mutation studies have been done to explore if the importance of tryptophan is due to its aromaticity or if intramolecular stablilisation between residues is also a factor. To this end, we mutate two strong-binding tryptophan-rich sequences, such that each tryptophan is replaced by either tyrosine (Y) or phenylalanine (F). The effect of the curvature of the surface is investigated by studying those sequences interacting with a carbon nanotube (CNT) and a graphite sheet.

Efficient evaluation of accuracy of quantum molecular dynamics using dephasing representation

Baiqing Li, Cesare Mollica, and Jiri Vanicek

Ecole Polytechnique Federale de Lausanne (EPFL)

Ab initio methods for electronic structure of molecules have reached a satisfactory accuracy for calculation of static properties, but remain too expensive for quantum dynamical calculations. We propose an efficient semiclassical method for evaluating the accuracy of a lower level quantum dynamics, as compared to a higher level quantum dynamics, without having to perform any quantum dynamics. The method is based on dephasing representation of quantum fidelity and its feasibility is demonstrated on the photodissociation dynamics of CO₂. We suggest how to implement the method in existing molecular dynamics codes and describe a simple test of its applicability.