Background: Protein-DNA docking is a very challenging problem in structural bioinformatics and has important\r\nimplications in a number of applications, such as structure-based prediction of transcription factor binding sites\r\nand rational drug design. Protein-DNA docking is very computational demanding due to the high cost of energy\r\ncalculation and the statistical nature of conformational sampling algorithms. More importantly, experiments show\r\nthat the docking quality depends on the coverage of the conformational sampling space. It is therefore desirable\r\nto accelerate the computation of the docking algorithm, not only to reduce computing time, but also to improve\r\ndocking quality.\r\nMethods: In an attempt to accelerate the sampling process and to improve the docking performance, we\r\ndeveloped a graphics processing unit (GPU)-based protein-DNA docking algorithm. The algorithm employs a\r\npotential-based energy function to describe the binding affinity of a protein-DNA pair, and integrates Monte-Carlo\r\nsimulation and a simulated annealing method to search through the conformational space. Algorithmic techniques\r\nwere developed to improve the computation efficiency and scalability on GPU-based high performance computing\r\nsystems.\r\nResults: The effectiveness of our approach is tested on a non-redundant set of 75 TF-DNA complexes and a newly\r\ndeveloped TF-DNA docking benchmark. We demonstrated that the GPU-based docking algorithm can significantly\r\naccelerate the simulation process and thereby improving the chance of finding near-native TF-DNA complex\r\nstructures. This study also suggests that further improvement in protein-DNA docking research would require\r\nefforts from two integral aspects: improvement in computation efficiency and energy function design.\r\nConclusions: We present a high performance computing approach for improving the prediction accuracy of\r\nprotein-DNA docking. The GPU-based docking algorithm accelerates the search of the conformational space and\r\nthus increases the chance of finding more near-native structures. To the best of our knowledge, this is the first ad\r\nhoc effort of applying GPU or GPU clusters to the protein-DNA docking problem.
Loading....