Quantum Monte Carlo Study of Photoprotection via Carotenoids in Photosynthetic Centers

A 2004 INCITE Project

This project, led by William A. Lester, Jr. of LBNL and UC Berkeley, was awarded 1,000,000 processor hours. This project aims to increase understanding of the complex processes which occur during photosynthesis, the process by which plants and bacteria convert the sun's light into energy, taking in carbon dioxide and producing oxygen in the process. This project is important on several levels. First, plants and bacteria are the world's foremost means of "carbon sequestration," or storing carbon from the atmosphere — a process which has enormous implications for climate change and global warming. Additionally, photosynthesis is an example of fundamental electron chemistry and is an efficient energy transfer system — processes which are fundamental in many areas of scientific research. The "Monte Carlo" in the title refers to simulations in which data are obtained by simulating a statistical model in which all parameters are numerically specified.

The project has been delayed somewhat by code development, porting, and other changes. After the project was under way, it was determined that the calculation could be made much more efficient by expressing the wave functions in a less dense representation and by using a Slater basis. In order to accomplish the latter, NERSC has acquired and installed the ADF software package for first-principles electronic structure calculations on NERSC's IBM SP, and the Lester group is now using this software.

The INCITE1 team has improved their QMC code by implementing an alternative algorithm, and as a result, it is more capable of scaling up to large systems involving hundreds of electrons. A sparse representation of the wave function was accomplished via a grid acceleration technique. The improvements yield a 16-fold reduction in wallclock time for one of the systems of interest (the spheroidene molecule shown in the figure). In terms of scale and computational complexity, this test system is very similar to the photosynthetic target systems. Calculations on a 44-electron test system (hexatriene) were completed to validate the new INCITE1 code. Large-scale test runs will be followed by full production runs.

Highest occupied molecular orbital (HOMO) of the singlet state
of spheroidene molecule. This molecule is responsible for
photoprotection in photosynthetic reaction centers. This is the
largest biological molecule ever treated using the accurate
quantum Monte Carlo method.

Coding of grid acceleration has been finished, which greatly improves the code's ability to scale up to large molecular systems such as the photosynthetic target systems outlined in the original proposal. Debugging of Schmidt-Moskowitz correlation functions and wave function optimization continues. The team has found the region of code where the disagreement occurs and is currently analyzing it. Debugging of the Umrigar-Nightingale-Runge algorithm for random walks is near completion. Optimization of single-particle moves and determinant updates appears to be necessary and is under way. NERSC consultant David Skinner has worked extensively with the INCITE1 team to optimize their codes; working notes about the code optimization are available here.

See also the Berkeley Lab View article, "Berkeley Lab Scientists Gain New INCITE on Photosynthesis" and the visualizations of the project's simulations.