ANALYSIS OF INTRAMOLECULAR MOTIONS BY FILTERING MOLECULAR-DYNAMICS TRAJECTORIES

A novel method for analyzing molecular dynamics trajectories has been developed, which uses digital signal processing techniques to eliminate unwanted motion and retains only motions of interest. In particular, it is possible to filter out the high-frequency motions and focus on the low-frequency collective motions of molecules. The trajectories of each of the atoms in the system (or any subset of interest) are Fourier transformed to the frequency domain, a filtering function is applied, and then an inverse transformation back to the time domain yields the filtered trajectory. The validity and merits of the filtering method were studied in detail for acetamide and N-acetylalanine-N′-methylamide as models for peptides and proteins. Initially the technique was tested for fluctuations around one local minimum. The normal modes obtained by diagonalizing the mass-weighted second-derivative matrix were combined to generate a well-characterized “normal-mode trajectory”. The frequency distribution of this trajectory approximates a δ function with a peak for each of the frequencies in the normal-mode analysis. By use of a filtering function that retains only one peak of the spectral distribution, a single mode was extracted. This filtered mode had all the characteristics of the original normal mode. Technical aspects such as the effects of simulation length and sampling frequency were also examined by using normal-mode trajectories. The method was then applied to “real” molecular dynamics trajectories. We have shown that information about the structural mobility at the vicinity of a minimum, traditionally obtained from normal-mode analysis, can also be extracted from molecular dynamics simulations by using the filtering technique. In addition, molecular dynamics simulations that include conformational transitions, such as αR → C7eq, were used for evaluating the merits of the filtering technique in anharmonic regions. We concluded that with this technique it is possible to characterize the important motions of a molecule in a way analogous to normal-mode analysis, without confining the study to harmonic oscillations and one local minimum energy conformation. The filtering technique is very flexible and can easily be applied to sections of a molecule or whole molecular systems, and various types of motion can be selected by designing the appropriate filtering function. Since it is also not very demanding computationally, it can serve as a powerful tool for the characterization of the dynamic behavior of small and large molecular systems. © 1990, American Chemical Society. All rights reserved.