Hellfried Schreiber

The influence of a protein on water dynamics in its vicinity investigated by molecular dynamics simulation

A system containing the globular protein ubiquitin and 4,197 water molecules has been used for the analysis of the influence exerted by a protein on solvent dynamics in its vicinity. Using Voronoi polyhedra, the solvent has been divided into three subsets, i.e., the first and second hydration shell, and the remaining bulk, which is hardly affected by the protein. Translational motion in the first shell is retarded by a factor of 3 in comparison to bulk. Several molecules in the first shell do not reach the diffusive regime within 100 ps. Shell‐averaged orientational autocorrelation functions, which are also subject to a retardation effect, cannot be modeled by a single exponential time law, but are instead well‐described by a Kohlrausch‐Williams‐Watts (KWW) function. The underlying distribution of single‐molecule rotational correlation times is both obtained directly from the simulation and derived theoretically. The temperature dependence of reorientation is characterized by a strongly varying correlation time, but a virtually temperature‐independent KWW exponent. Thus, the coupling of water structure relaxation with the respective environment, which is characteristic of each solvation shell, is hardly affected by temperature. In other words, the functional form of the distributions of single‐molecule rotational correlation times is not subject to a temperature effect. On average, a correlation between reorientation and lifetimes of neighborhood relations is observed.

The influence of temperature on pairwise hydrophobic interactions of methane‐like particles: A molecular dynamics study of free energy

The association of a pair of hydrophobic solutes in water has been investigated by free energy molecular dynamics simulations of a system containing 516 water molecules. Convergence of the calculations is guaranteed by the comparison of data obtained with two independent free energy sampling techniques, which have been optimized for our system. Coulomb interactions have been treated with the Ewald method. Using this computationally expensive approach many of the previously reported discrepancies in the temperature, pressure and interaction parameter dependence of hydrophobic association are clarified. A temperature effect on both the free energy of association and the equilibrium between contact and solvent‐separated species is observed. Raising temperature favors association. The most pronounced temperature dependence occurs in the interval between 300 and 350K.

Parallel molecular dynamics of biomolecules

Abstract The basic principles of a typical sequential Molecular Dynamics (MD) program suitable for the study of solvated biomolecules are described, the inherent parallelism of MD is analysed and strategies for parallelisation are developed. Due to separate treatment of computation and communication a high level of portability is achieved and both tasks can be optimized independently. It is found out that communication trees are highly efficient means for sending, receiving and gathering the computed data, especially for large processor numbers. A current implementation on a transputer system is presented. Due to the tight memory budget slight modifications are necessary. Nevertheless, we get excellent performance with an average degree of parallelization of 82%.

The frequency-dependent conductivity of a saturated solution of ZnBr2 in water: A molecular dynamics simulation

The first part of this paper reviews the theory of the calculation of the frequency-dependent dielectric properties (i.e., conductivity and dielectric constant) of ionic solutions from computer simulations. Based on a 2.2-ns molecular dynamics simulation, the second part presents a detailed analysis of the various contributions to the frequency-dependent conductivity of a saturated solution of ZnBr2 in water. We find evidence for two separate relaxation channels in the frequency-dependent conductivity, and a very low value for the static (i.e., zero frequency) conductivity, which is consistent with the high degree of ion association and the prevalence of electrically neutral ion clusters that we observe in this system.

Static and dynamic structural analysis of a saturated solution of ZnBr2 in water: Anomalous x‐ray diffraction and molecular dynamics simulations

The supposedly very simple system of a saturated solution of ZnBr2 in water exhibits unusually complex and therefore interesting structural behavior. Motivated by this, Mager did a detailed x‐ray diffraction study (Th. Mager, PhD. thesis, Universitat Wurzburg, 1991), and we performed a long molecular dynamics (MD) simulation—using potential parameters from the general purpose GROMOS force field—of such a solution, which can be grossly characterized by the formula ZnBr2⋅3H2O. We start by calculating those properties that are directly accessible through the experiment from the MD simulation, in order to validate the physical relevance of the simulation. Seeing that the simulation delivers results that are compatible with those of the experiment, we proceed by analyzing the MD simulation in much more detail according to the static and dynamic structure of the system, thereby gaining insight into the structural behavior of ZnBr2⋅3H2O that is very difficult, if at all possible, to get from experimental studies. To this end we use the Voronoi algorithm to define coordination shells around atoms and ions in ZnBr2⋅3H2O. We study the time averaged as well as the time‐resolved geometry and composition of these coordination shells and find that octahedral coordination of Zn2+ ions is the dominant geometry in ZnBr2⋅3H2O, and that these octahedra are remarkably stable (after 1 ns only 10% decayed). We further find evidence for polymerlike Zn2+ chains, where O atoms of water and Br− ions connect the Zn2+ ions.

Parallel Biomolecular Simulation: An Overview and Analysis of Important Algorithms

We have presented important algorithms for serial MD simulations of biomedical systems and have analysed their impact on parallel performance. None of these algorithms can be neglected if we are interested in true gains in throughput and not just in good formal scalability numbers. This is especially true for the SHAKE algorithm, which due to its small contribution to the total runtime and due to its inherently serial character is often not included in reports on the parallelisation of MD programs. We have shown clearly that even a very modest speedup in this algorithm is essential for increased overall performance.

Parallel biomolecular simulation: Theory, algorithms and implementation

Abstract The essential goal of this work is the unified treatment of quantum mechanical and classical degrees of freedom in biomolecular simulation on all three levels: Theory, algorithms and implementation. In theory this is done within the framework of the Lagrangian model, which handles electronic coordinates and Cartesian nuclear coordinates in a consistent way. Furthermore, there is a 1:1 correspondence between the various algorithmic substeps of self-consistent-force molecular dynamics (SCF-MD) and classical molecular dynamics (CMD): The overlap criterion corresponds to the cut-off principle, the list of integrals to the pairlist and the computation of interaction matrix elements to the computation of pair forces. (The integration step is identical in both schemes.) This complete analogy can be conserved on the implementation level, for which actual benchmarks on three common architectures are given for a CMD simulation of the hydrated protein ubiquitin.