The objective of this research is to develop a petascale cyberinfrastructure for the efficient calculation of free energy landscapes for complex macro- and biomolecular systems. The approach is to develop a cyberinfrastructure for performing Wang-Landau calculations of free energy distributions on petaflop architectures and, to demonstrate its effectiveness, apply it to two outstanding science problems (conformations of a linker protein in solution and self-assembly of lipids). Free energy is a measure of the energy of the system that takes into account entropic and thermal effects. Understanding the relative free energies of, and barriers between, various conformations of macromolecular and biological molecules, including proteins, and of phase transitions between disordered and self-assembled structures, is an extremely important area of chemistry, biochemistry and biophysics and is an area where increased computational capabilities hold the promise of significant progress.
The intellectual merit of the proposed research is that it has the potential to result in new algorithms for performing free energy calculations on petascale architectures, in addition to yielding insights into conformational equilibria and self-assembly in macro- and biomolecular systems. The broader impacts of the proposed research include the development of a robust open source petascale cyberinfrastructure for performing free energy calculations made possible only through a strong synergy between physics, chemistry, biology, engineering, and computer science. Through the research performed and educational and outreach programs at the participating institutions, the project exposes high school, undergraduate and graduate students and postdoctoral researchers to the frontiers of computational physics, chemistry and materials research.
