Solvation modeling for next-gen biomolecule simulations

项目摘要/摘要 计算生物物理学和药物发现需要更快、更好、在某些情况下 蛋白质溶剂化和蛋白质-蛋白质相互作用的完全重新制定的物理模型： 用于设计可以夹在大的蛋白质-蛋白质界面中的大环化合物； 模拟生化途径；用于计算多抗体的运动、结合和识别； 用于配制针对折叠和聚集不稳定性的治疗性蛋白质解决方案；以及 缓解蛋白质聚集的疾病。 实现快速、准确和可扩展的蛋白质建模，这些蛋白质是大的、复杂的或 聚集体，以及在水中的，需要一个团队能够从四个主要非 重叠研究社区：原子蛋白质MD、蛋白质-蛋白质对接、蛋白质-胶体 液态理论和水统计力学。大型企业需要将这些方法结合起来 关于生物相关时间和空间的快速和准确的计算机建模的进展 规模，并有适当的统计机制。在这里，我们的团队是6个已经成对的PI 高度协作(42篇联合论文)，每一篇都带来了前沿能力(Simmerling，一个关键 显影剂或琥珀和GBNECK；CLUSPRO的开发商科扎科夫，顶级蛋白质-蛋白质相互作用 卡普里的Web服务器；Coutsias，数学几何学家，其BRIKARD提供了经过验证的加速 受限搜索减少100倍；Hribar-Lee，他的Wertheim理论成功地预测了Simple 蛋白质聚集；Fennell，快速准确的水模型SEA的开发者；以及Dill，开发者 水和已被证明成功的MD加速器MELD的统计力学模型 在履约协助方案中)。 我们的5年目标包括：(A)超越僵化的蛋白质-蛋白质对接，包括 构象灵活性、原子细节、对大型系统的可伸缩性和亲和力。(B)预测 蛋白质和抗体聚集热点以及对盐和辅料的依赖。(C) 用新一代隐式溶剂开发AmberSB力场，更快、更准确 表面积计算，在CASP、SAMPL和CAPRI事件中进行盲测。(D)发展中 溶液平衡和水动力学的“超快”分析水模型，如 在表面和通过毛孔的扩散、粘度和传输。团队管理计划是 建议在协调领导下优化协作研究，并为持续的 交流、参与和集体智力的发展。

Project Summary / Abstract Computational biophysics and drug discovery need much faster, better, and in some cases completely reformulated physical modeling of protein solvation and of protein-protein interactions: for designing macrocyclic compounds that can sandwich into large protein-protein interfaces; for modeling biochemical pathways; for computing multi-antibody motions, binding and recognition; for formulating therapeutic protein solutions against folding and aggregation instabilities; and to mitigate against diseases of protein aggregation. Achieving fast, accurate and scalable modeling of proteins that are large or in complexes or aggregates, and that are in water, requires a team that can innovate from four largely non- overlapping research communities: atomistic protein MD, protein-protein docking, protein-colloid liquid-state theory, and water statistical mechanics. Combining these approaches is needed for big advances toward fast and accurate computer modeling on biologically relevant time and space scales, with proper statistical mechanics. Here, our team is 6 PIs that have already been pairwise highly collaborative (42 joint papers), and that each bring forefront capabilities (Simmerling, a key developer or AMBER and GBNECK; Kozakov, developer of CLUSPRO, top protein-protein interaction webserver in CAPRI; Coutsias, mathematical geometer whose BRIKARD gives proven acceleration of constrained search by 100x; Hribar-Lee, whose Wertheim Theory successfully predicts simple protein aggregation; Fennell, developer of SEA, a fast accurate water model; and Dill, developer of statistical mechanical models of water and of MELD, an MD accelerator that has proven successful in CASP). Our 5-year Aims include: (A) Going beyond rigid protein-protein docking, to include conformational flexibility, atomic detail, scalability to large systems, and affinities. (B) Predicting protein and antibody aggregation hot-spots and dependencies on salts and excipients. (C) Developing AmberSB force fields with next generation implicit solvent, and faster, more accurate surface-area calculations, with blind testing in CASP, SAMPL and CAPRI events. (D) Developing ‘super-fast’ analytical water models for solution equilibria, and for water dynamics, such as diffusion, viscosities and transport at surfaces and through pores. A Team Management Plan is proposed to optimize collaborative research with concerted leadership, and to provide for ongoing communication, engagement and the development of collective intelligence.