Aggregation of ABeta peptides is linked to Alzheimer's disease, the most common dementing disorder. The exact cause of ABeta cytotoxicity remains uncertain, but it may be related to the interactions of ABeta peptides with cellular membranes leading to their disruption. Currently, molecular mechanisms of ABeta-membrane interactions are largely unknown. All-atom molecular dynamics simulations with explicitly treated solvent can uncover these mechanisms with the level of detail not available through other methods. However, due to extreme demands of such simulations, existing computational platforms are not adequate. To address these limitations, we propose a novel adaptation of replica exchange molecular dynamics (REMD) simulator to a massively parallel grid computing architecture, the Frontier? Enterprise Computing Platform developed by Parabon Computation. Because Frontier can harness the idle computing power of tens of thousands of computers worldwide and deliver it as a web-based service, we expect that the Frontier-enabled version of REMD, called dREMD, will allow researchers to capture the full potential of REMD in a timely and affordable manner from any standard web browser.
The specific aims of the project include the development of dREMD software based on existing in-house REMD codes developed by the academic Principal Investigator. The dREMD simulator will be thoroughly tested by comparing its functions with the results of conventional REMD performed on Linux clusters and experimental data. We will then apply dREMD to probe the thermodynamics of ABeta-lipid bilayer interactions. By performing free energy computations, we will study the molecular mechanisms of ABeta monomer and oligomer binding and insertion into cellular bilayers. Our objective will be to determine the ability of ABeta oligomers to perturb or disrupt bilayers. Accomplishment of these specific aims will position Parabon for a Phase II STTR project that will expand the usability features of dREMD and apply it to the study of inhibitors of ABeta-induced disruption of lipid bilayers. The business objectives are to bring to market two offerings: (1) dREMD atop the Parabon Frontier Computation Grid as a pay-per-use service offered to pharmaceutical and biotechnology companies performing structural biology and ligand binding research;(2) dREMD packaged with Frontier Enterprise for in-house delivery of the same capabilities. The biomedical significance of the project is twofold. First, the project is expected to transform the practice of biomolecular simulations by making available to the research community a new and powerful computational tool - dREMD. Being general-purpose, affordable, and easy to use, dREMD will expand the scope of biomolecular simulations to new challenging problems. Second, the project is expected to shed new light on the mechanisms of ABeta peptide cytotoxicity implicated in Alzheimer's disease.
This project will develop a Software as a Service (SaaS) application called dREMD, a novel adaptation of the replica exchange molecular dynamics algorithm ported to Parabon Frontier, a massively parallel grid computing architecture developed by Parabon Computation. To illustrate the feasibility of applying dREMD to challenging biomedical problems, it will be used to study interactions of ABeta peptides with lipid bilayers in explicit solvent. dREMD is a general-purpose, affordable, and easy to use computational tool expected to significantly expand the scope of biomolecular simulations.
|Lockhart, Christopher; O'Connor, James; Armentrout, Steven et al. (2015) Greedy replica exchange algorithm for heterogeneous computing grids. J Mol Model 21:243|