Advanced Electrostatic Computation in Molecular Dynamics

描述（由申请人提供）：计算硬件和分子建模技术的进步已经彻底改变了我们模拟静电相互作用的能力，静电相互作用在生物分子的稳定性，结构，折叠和功能中起着重要作用。目前，基于泊松-玻尔兹曼方程（PBE）的连续体静电描述提供了建模保真度和计算成本的最佳组合。虽然非常大的生物分子组件，如核糖体，含有数十万个原子，已成功地解决了使用PBE求解器，遇到的主要困难时，这些求解器被纳入分子动力学（MD）和能量最小化（EM）代码。在分子表面预测的静电场的准确性受损的MD模拟不稳定，并阻碍EM计算的收敛。所提出的努力将提供新的计算工具，以准确和有效地预测所需的分子表面附近的静电性能。以前的技术，特别是一种新的自适应笛卡尔网格结构，可以最大限度地提高大规模生物分子的计算性能，将被修改以实现表面的高精度。由此产生的计算工具应该证明是非常宝贵的：（i）MD和EM计算，通过有效地提供高保真度，PBE质量的静电能量梯度和（ii）研究人员有兴趣在相关的静电表面性质（例如，电势和感应电荷）到生物分子功能。

DESCRIPTION (provided by applicant): Advances in computational hardware and molecular modeling techniques have revolutionized our ability to simulate electrostatic interactions which play a fundamental role in the stability, structure, folding and function of bio-molecules. Currently, the continuum electrostatic description based upon the Poisson- Boltzmann Equation (PBE) offers the best combination of modeling fidelity and computational cost. While very large bio-molecular assemblies such as ribosomes, containing hundreds of thousands of atoms, have been successfully solved using PBE solvers, major difficulties are encountered when these solvers are incorporated into molecular dynamics (MD) and energy minimization (EM) codes. Impaired accuracy of predicted electrostatic field at molecular surfaces destabilizes MD simulations and hinders convergence in EM computations. The proposed effort will provide new computational tools to accurately and efficiently predict the required electrostatic properties near the molecular surface. Previous technology, notably a novel adaptive Cartesian grid structure that maximizes computational performance for large-scale bio-molecules, will be modified to achieve high accuracy at the surface. The resulting computational tool should prove invaluable to: (i) MD and EM calculations by efficiently providing high fidelity, PBE-quality electrostatic energy gradients and (ii) researchers interested in relating electrostatic surface properties (e.g., potential and induced charge) to bio-molecular function.