Process of Efficiently Parallelizing a Protein Structure Determination Algorithm

Abstract

Computational protein structure determination involves optimization in a problem space much too large to exhaustively search. Existing approaches include optimization algorithms such as gradient descent and simulated annealing, but these typically only find local minima. One novel approach implemented in REDcRAFT is to instead of folding a protein all at the same time, fold it residue by residue. This simulates a protein folding as each residue exits from the generating ribosome. While REDcRAFT exponentially reduces the problem space so it can be explored in polynomial time, it is still extremely computationally demanding. This algorithm does have the advantage that most of the execution time is spent in inherently parallelizable code. However, preliminary results from parallel execution indicate that approximately two-thirds of execution time is dedicated to system overhead. Additionally, by carefully analyzing and timing the structure of the program the major bottlenecks can be identified. After addressing these issues, REDcRAFT becomes a scalable parallel application with nearly two orders of magnitude improvement.