Research activities

Program P1-0002

Dr. Dušanka Janežič

Computer simulation methods have been developing primarily in the areas of increasing the simulation lengths and the size of modeled systems, which allows a greater understanding of the relationship between the structure and function in biological macromolecules. Due to the heterogeneous nature of these macromolecules, an average of several simulations of the same system must be performed. Including a solvent in simulations greatly increases the scope of the simulated system. The length of simulations has to be greater than on the nanosecond scale to study chemically and biologically interesting processes, which occur on the microsecond scale. We expect that the development of new methods for molecular modeling and their implementation on parallel computers will increase the speed of computer simulations for several orders of magnitude. This will allow a more detailed examination of known problems as well as allow the examination of new problems. The program has the following goals:

a) Further development and use of the SISM (Split Integration Symplectic Method) and HANA (Hydrogens ANAlytically) symplectic methods for molecular dynamics simulations of macromolecules that will allow integrating equations of motion with a long time step while remaining stable, computationally economical and being efficiently parallelizable. The methods are based on the factorization of the Liouville operator and differ from other methods using a split scheme in their analytical treatment of high frequency oscillations. Their benefit will be demonstrated on some biologically interesting examples, especially proteins.

b) Further development and use of the combination of molecular dynamics methods, normal mode vibrational analysis, and quasiharmonic analysis of proteins in solutions for studying protein hydration.

c) Further development and use of QM/MM methods, which allow computer simulations using a combination of a classical and quantum potential, allowing for more accurate simulations of large biological molecules at the ab initio level.

d) Further development of computationally efficient methods for determining the time-dependent electronic structure of molecules based on the Kohn-Sham formulation of the density functional theory in which the electron density is calculated using single-electron Green's functions. Since the most computationally demanding operations are only computed locally, the speed of calculating the electronic structure of molecules is significantly increased.

e) Further development and application of quantum chemical and classical approaches for calculating reaction mechanisms, especially calculating the ionic reactions of isocyanides. We will determine whether it is generally true that ionic reactions of isocyanides proceed as multicomponent chemical reactions. No such studies have been performed using computational methods and it is not possible to study this problem experimentally.

f) Further development and use of the RISM formalism, which is based on the theory of integral equations, which will, in combination with Monte Carlo simulations, enable a deeper understanding of the relationship between the properties of molecules and the macroscopic properties of the matter they form. These relationships are given by distribution functions that are obtained by using the theory of integral equations.

g) Further development of new and effective network topologies for connecting personal computers into clusters and that will enable fast parallel performance for molecular dynamics programs and allow for an effective implementation of the newly developed methods on parallel computers.