A parallel realization of the NDDO-WF technique for semi-empirical quantum-chemical calculations on large molecular systems in the spd-basis is described. The technological aspects of designing scalable parallel calculations on super computers (by using MPI library) are discussed. The scaling of individual algorithms and entire package was carried out for two model systems with a number of atomic orbitals of 894 and 2014, respectively. The speedup was determined in computer experiments with the RM600 E60 and Cluster Intel PIII multi-processor systems. The effect of communication rate on the package performance is discussed.
The computer modelling and simulation methods are widely used in rational drug design to obtain information necessary for understanding interactions between a ligand (drug molecule) and its cellular macromolecular target on molecular level. The determination of free energies of binding for ligand-target complexes is one of the crucial points in those studies. In recent years several methods have been proposed to solve this problem. The majority of them use molecular dynamics (MD) simulations. Two, most popular methods: (i) a free energy perturbation method (FEP), and (ii) a linear response (LR) method, are shortly presented in this paper together with their limitations and advantages. In this work I presented the first attempt to use LR approach to 10 anti-tumour agents able to intercalate into DNA. The LR relationship obtained in the present study indicated that in the system studied the electrostatic term has no influence on the free energy of binding. The relationship is now successfully used in our research group in further molecular modelling studies concerning DNA intercalators with similar structure.
Molecular dynamics simulations were carried out on tyrosine and phenylalanine and their derivatives with various terminal groups to determine the populations of side-chain rotamers. The obtained populations were compared with those calculated from fluorescence-decay lifetime distributions and NMR studies. It was found that theoretically calculated populations do not match the experimental ones, which suggests that the static rotamer model is inadequate to explain the dynamics of tyrosine and phenylalanine side chain in fluorescence and NMR experiments.
Using the Kirkwood-Buff theory of solutions, the preferential solvation of the N-methylacet-amide (NMA), N-methylformamide, (NMF), and N,N-dimethylformamide, (DMF), molecule has been investigated in the binary {amide+methanol} mixtures at 313.15 K. Moreover, for the {amide+methanol} mixtures, where amide = NMF, DMF, and NMA, the molecular dynamics calculations at x_{amide}=0.518 were performed. From the obtained molecule-molecule radial distribution functions, (rdf), and atom-atom rdf, it was possible to estimate the local mole fractions around the amide molecule, the orientation effects of molecules within the solvation shell, and a possibility of the formation of complexes. The general picture obtained from analysis of the molecular dynamics results is consistent with the deductions derived from thermodynamic data.
Structural and elastic properties of the densest known solid phase of two-dimensional (2D) system of hard cyclic pentamers (each pentamer is composed of five discs which centres are placed at vertices of a perfect pentagon of sides equal to the disc diameter, σ) are studied by Monte Carlo simulations. The present study confirms that at high densities the pentamers form a 2D solid structure of rectangular lattice with two pentamers (which librate, without rotation, around their mean orientations) in the unit cell. Elastic constants calculated for this structure show that, in contrast to densely packed 2D hard cyclic heptamers (composed of seven discs of centres forming a perfect heptagon of sides equal to the disc diameter σ), the pentamers do not exhibit anomalous Poisson's ratios.
In this paper we discuss applications of molecular dynamics in modeling of nonequilibrium effects in chemical systems. We focus our attention on simulations, which use the "reactive" hard spheres technique. It is demonstrated that information on nonequilibrium rate constant in a system with a thermally activated reaction can be easily obtained from such simulations. We also present results for a wavefront propagation in a system with an autocatalitic reaction: A + B --> A + A.
The dispersion of the agglomerating fluid process involving colloids has been investigated at the mesoscale level by a discrete particle approach - the hybrid fluid particle model (FPM). Dynamical processes occurring in the granulation of colloidal agglomerate in solvents are severely influenced by coupling between the dispersed microstructures and the global flow. On the mesoscale this coupling is further exacerbated by thermal fluctuations, particle-particle interactions between colloidal beds and hydrodynamic interactions between colloidal beds and the solvent. Using the method of FPM, we have tackled the problem of dispersion of a colloidal slab being accelerated in a long box filled with a fluid. Our results show that the average size of the agglomerated fragments decrease with increasing shearing rate Γ, according to the power-law A*Γ^{k}, where k is around 2. For larger values of Γ, the mean size of the agglomerate S__{avg} increases slowly with Γ from the collisions between the aggregates and the longitudinal stretching induced by the flow. The proportionality constant A increases exponentially with the scaling factor of the attractive forces acting between the colloidal particles. The value of A shows a rather weak dependence on the solvent viscosity. However, A increases proportionally with the scaling factor of the colloid-solvent dissipative interactions. These results may be applied to enhance our understanding concerning the nonlinear complex interaction occurring in mesoscopic flows such as blood flow in small vessels.
The paper presents results of a few preliminary simulation experiments of a self-reproduction system realized in a programming environment for individual-based modelling of physical systems and discusses the advantages and difficulties of such modelling. The programming environment named an abstract universe, is aimed at modelling of complex systems which manifest the self-organization, self-modification, growing and emergent behavior. The central idea of the universe are specific interactions of its entities in a two dimensional space. The entities move and collide according to rules like those of classical mechanics, and interact between themselves according to function encoded in them modifying their structures and functions.
1997-2020 (C) CI TASK quarterly@task.gda.pl