The research objective of this EArly-Concept Grants for Exploratory Research (EAGER) award is to create visualization refinements to an existing methodology that will support virtual protein manipulation as an experimental framework for formulating design theory. The investigators will integrate insights from known constraints on mechanisms to create more efficient protein folding algorithms so as to enable nano-machine design. In this high-risk, high-pay-off project, the process will be visualized on high performance computing platforms that have been designed for protein folding. Spline models of knots - inclusive of their polygon control structures - will be visualized. The user will be able to quickly conduct extensive 'what-if' visual experiments to understand the 3D design ramifications of differing protein manipulations. The visualization capabilities will prompt new intuition and novel conjectures by focusing attention on the design properties arising from the inherent crossings and self-intersections. The designer will be able to efficiently explore a much richer design space than can be achieved by existing algorithmic methods. Hence, visualization, engineering design, and high performance computing form the foundations for this experiment to improve protein folding algorithms.
If successful and the performance improvements that have already been shown within small scale prototypes can be transferred to realistically complex models, then this work would have significant impact across numerous communities, with substantial impact in industry. These algorithms are currently the subject of extensive study throughout the bio-molecular, medical and pharmaceutical communities. There will also be a broader impact in educating students, as the students involved will have the opportunity to work with realistically complex and voluminous data sets from the industrial partner, IBM. Students will also be exposed to the complexity of the state of the art high performance computing platforms, such as IBM is providing.
The research of this EArly-Concept Grants for Exploratory Research (EAGER) analyzed `snap shots' of protein folding, with integration to high performance computing simulations of these molecules. This process is know as dynamic scientific visualization and was conducted in collaboration with IBM Research, which graciously supplied access to their high performance computing platforms for relevant experiments. A key insight from mechanical engineering was to reduce the angular freedom within the folding for significant algorithmic efficiency. This approach permits the biochemical end user to quickly conduct extensive 'what-if' visual experiments to understand the 3D design ramifications of differing protein manipulations. The ability to quickly `see' these configurations facilitates scientific insight that can easily be obscured by purely numerical analysis of the voluminous data generate by these simulations. An undergraduate student was supported through a supplement under Research Experiences for Undergraduates. The student produced code that is publicly accessible which permits 3D manipulation of data. This is being considered for further integration with these molecular simulations as well as for editing during the creating of computer animations in the film industry -- an unexpected spin-off. The options with the motion picture industry are being investigated for technology transfer with Innoventive Software, LLC, San Diego, CA under related NSF funding.
