Michael R. Pear

Molecular dynamics of ferrocytochrome c: Magnitude and anisotropy of atomic displacements

Abstract Two computer simulations of the atomic motion in tuna ferrocytochrome c have been carried out. The average structures and the structural correlations of the magnitudes of the atomic position fluctuations are in substantial agreement with recent X-ray diffraction results, particularly for the protein interior. The simulations show, however, that the atomic displacements are quite anisotropic. The degree of anisotropy and the preferred directions of atomic displacement exhibit correlations with structural features of the protein.

Brownian dynamics study of a polymer chain of linked rigid bodies

A Brownian dynamics model for the backbone chain of a macromolecule is developed as a system of linked rigid bodies so that constraints on valence angles and bond lengths are satisfied exactly. For comparison, a corresponding flexible model is developed in which bond lengths and valence angles are held nearly constant by strong harmonic potentials. Equilibrium properties and barrier crossing rates are examined theoretically and by computer simulation of both models, with differences arising due to the presence of constraints in the rigid case. A compensating potential based on the metric determinant of unconstrained coordinates in the rigid model is found to eliminate the effect of constraints. Barrier crossing rates in the transition state approximation are studied when a force fixed in space is applied to the end atoms of the three‐bond chain. An exact transition state rate formula developed for this case predicts curved Arrhenius plots of barrier crossing rates; this result is confirmed by computer sim...

A generalization of Kramers’ rate formula to include some anharmonic effects

Consideration from the Langevin approach of Brownian‐motion effects on a particle in a parabolic barrier potential leads to a transmission function which gives the probability that the particle will surmount the barrier. When used in conjunction with an approximate low‐temperature normalization condition, the Kramers rate formula, originally derived using the Fokker–Planck approach, is reproduced. The rate formula is then generalized by including anharmonic effects due to the presence of the barrier as they enter in an exact normalization condition. The generalized Kramers formula has a temperature dependence of the frequency factor which is verified by computer simulation for a periodic and double‐well potential. Data from computer experiments are fitted using both the original and generalized formulas. The generalized formula is found to be useful in extracting information on the barrier height and friction coefficient from the experimental data.

Brownian dynamics study of a polymer chain of linked rigid bodies. II. Results for longer chains

As an extension of a previous paper, computer simulation results for a polymer chain of linked rigid bodies are presented for chains with between four and fifteen bonds. Previous theoretical results for a three‐bond chain are found by computer simulation to apply to longer chains. In particular, a Fixman potential for an N‐bond chain is developed as the sum of three‐bond Fixman potentials, and transition state rates measured in computer simulation for a four‐bond chain compare reasonably well with the theory for a three‐bond chain. A further study of the motion of the free end of a chain with one end fixed indicates correlation effects diminish the possibility of a ’’whipping motion’’ of the chain end.

Hydrodynamic interaction effects on local motions of chain molecules

The librational motions of a butane molecule in water are simulated by Brownian dynamics with and without inclusion of hydrodynamic interactions. Inclusion of the hydrodynamic interactions is found to increase the torsional correlation times by 35% to 46%, depending on the isomeric state of the butane molecule, and to decrease the rate constants for gauche–trans isomerization by about 28%. A critical discussion of the approximations involved in such simulations is presented.The librational motions of a butane molecule in water are simulated by Brownian dynamics with and without inclusion of hydrodynamic interactions. Inclusion of the hydrodynamic interactions is found to increase the torsional correlation times by 35% to 46%, depending on the isomeric state of the butane molecule, and to decrease the rate constants for gauche–trans isomerization by about 28%. A critical discussion of the approximations involved in such simulations is presented.

Diffusional correlations in polymer dynamics

The authors present a simple theory which shows how viscous forces influence the character of conformational transitions in a polymer. (AIP)

GUANIDINIUM AS A PROBE OF THE GRAMICIDIN CHANNEL INTERIOR

Guanidinium is a planar trigonal cation which is similar in size to the gramicidin channel pore. We measured the effect of guanidinium on the conductance properties of the gramicidin channel and theoretically evaluated its interactions with the β-6.3 channel interior using an energy minimization and conformational search approach. Guanidinium current (measured in the absence of other permeable ions) could not be detected directly (g(Guan)/g(K) < 0.004). However, guanidinium induces blocks in gramicidin channel potassium currents. The average block duration gets shorter with increased membrane potential suggesting that guanidinium can penetrate the ion channel. Energy minimization calculations indicate that, by reorienting along the pathway, the guanidinium should be able to penetrate the gramicidin channel. This finding is illustrated by a computer graphics animation of the series of minimum-energy orientations. The low permeability of the channel to guanidinium is tentatively ascribed to an entropic barrier resulting from the restrictions on the ion motion in the channel.

A new model for user services: distributed support

During the 1980s, computing has become broadly available on our campuses beyond the traditional technical disciplines. Improvements in networking and data communications have made mainframe computing more accessible to users in the humanities and social sciences. And personal computers have made computing affordable to individuals and resource-poor departments. With the proliferation of computers and access, the user community which we serve has increased manyfold. Gone are the days when user services staff worked with expert users who understood programming and were willing to put up with the idiosyncrasies of a computing environment. Today, most of the computing on campuses relies on commercially available applications — not custom-built programs. While computing systems, particularly on personal computers, have become easier to use, there are still system complexities, advanced application features and new ways of thinking to make use of computing tools which go beyond the experience level (and desire to learn) of most users. Lacking depth of understanding, many computers users only slowly assimilate new methods and discover new applications. How can we deal with this larger and more diverse community effectively? I believe the answer is to establish departmentally-based user services staff balanced by a central staff of experts in areas of general interest on campus. Local staff, with a background and interest in the discipline, bring an understanding of computer technology and the applications to the discipline. In many ways, such a structure is evolving naturally as some people who began using computers as tools in their field became interested in the technology itself. These people are known to others in their departments as the local “expert”, and are consulted informally by others to resolve computing problems. For lack of a common title, we call these people Departmental Computing Coordinators. In this paper, I first discuss an ideal model for distributed support in campus departments. I then discuss how we at Brown are working towards such a model through our Departmental Computing Coordinator Program. I will focus on support for faculty, staff, and graduate students, i.e., those people who are generally associated with a particular unit of a university, and not on student computing.