Autostereoscopic Visualization and Geometric Computing for Biological Macromolecules

美国国家科学基金会0429753生物大分子自动立体可视化和几何计算计划中的研究旨在为大型蛋白质组学数据开发可视化和几何推理的工具和技术。提出的研究有三个主要组成部分--基于隐函数的蛋白质表示，蛋白质静电学的快速计算，以及实时自动立体可视化。这项研究将开发一个统一的隐含功能框架，可以紧凑而有效地表示非结合蛋白的性质，包括蛋白质的静电学和可及性。通过紧密地耦合这些性质，基于隐函数的表示可能会在这些问题的解中产生协同有利的交叉。此外，为隐式函数提供自适应的、渐进的和分层的表示将允许它们在多分辨率方案中用于存储、传输和计算控制。蛋白质的静电相互作用在许多生物过程中都是非常重要的。蛋白质及其性质的自动立体可视化对于理解蛋白质的三维结构非常重要。这项研究将利用图形共享软件的最新进展来开发绘制隐式曲面和体的新方法，并采用高效的方法来进行个性化的自动立体显示。

NSF 0429753 Autostereoscopic Visualization and Geometric Computing forBiological MacromoleculesThe planned research seeks to develop tools and techniques forvisualization and geometric reasoning for large proteomics data. Theproposed research has three main components -- implicit-function-basedrepresentations for proteins, faster computation of proteinelectrostatics, and real-time auto-stereoscopic visualization. Theresearch will develop a unified framework of implicit functions thatcan compactly and efficiently represent non-bonded protein propertiesincluding protein electrostatics and accessibility. By tightly couplingthese properties, implicit-function-based representations will likelyresult in synergistically favorable crossovers in the solutions to theseproblems. Further, providing adaptive, progressive, and hierarchicalrepresentations for implicit functions will allow them to be used inmultiresolution schemes for storage, transmission, and computationalsteering. Protein electrostatic interactions are of central importance formany biological processes. Autostereoscopic visualization of proteinsand their properties is important in understanding the 3D structure ofproteins. The proposed research will develop new methods for renderingimplicit surfaces and volumes using the recent advances in graphicshardware as well as adapt efficient methods for autostereoscopic displayswith personalization.