Tangible Augmented Interfaces for Structural Molecular Biology
IEEE Computer Graphics and Applications
13 W ith the prevalence of structural and genomic data, molecular biology has become a human-guided, computer-assisted endeavor. The computer assists the essential human function in two ways: For exploring scientific data, it searches for and tests scientific hypotheses. For collaborating between two or more scientists, it shares knowledge and expertise. As databases grow, as structure and process models become more complex, and as software methods become more diverse, access and manipulation
... digital information is increasingly a critical issue for research and collaboration in molecular biology. We have developed an augmented reality (AR) system that lets virtual 3D representations of molecular structures and properties be overlaid on autofabricated models of the molecules (for more information, see the "Molecular Biology Trends and Tools" sidebar). While using our tangible interaction environment, users can intuitively manipulate molecular models and interactions, easily change the representation shown, and access information about molecular properties. Design of physical models We use our Python Molecular Viewer (PMV) 1 both to create virtual objects and design tangible models, great-ly simplifying the integration of the models with the virtual environment. PMV is a modular software framework for designing and specifying a wide range of molecular models, including molecular surfaces, extruded volumes, backbone ribbons, and atomic balland-stick representations. It allows the design of models at different levels of abstraction for different needs. The software uses representations that focus on molecular shape when large systems and interactions are presented and incorporates atomic details when needed to look at functions at the atomic level. We built PMV within the interpreted language Python, which serves as a glue layer to interconnect different software components at a high level. PMV includes our generic 3D visualization component (DejaVu), which provides a high level interface to the OpenGL library and its geometry viewing application. To this, we added components that provide all manner of molecular modeling and visualization functionality. We have added output formats (.stl and .vrml) for the molecular models that serve as input for the solid printers. Figure 1 (next page) shows a variety of physical models produced in PMV and printed on either a Zcorp 406 color solid printer or a Stratasys Prodigy ABS plastic printer. Currently, researchers in structural molecular biology primarily use computer-generated 3D models. Collaboration, both remote and local, is aided by shared viewing of these interactive visual representations of molecular data. Yet, recent advances in the field of human-computer interfaces have not yet impacted molecular biologists-most work in biomolecular structure and genomics is performed in front of a workstation display using a mouse and keyboard as input devices. Early structure research relied heavily on physical models: Linus Pauling used his newly invented space-filling models to depict the molecular structures that he solved by crystallography and to predict the basic folding units of protein structures. James Watson and Francis Crick used brass-wire molecular models to help them devise an atomic model of the DNA double helix, which reconciled decades of genetic data. These researchers "thought with their hands" by using physical analogs to produce important scientific results. Current research in molecular biology now focuses on larger assemblies and on more complex interactions, for which the traditional physical molecular models are inadequate. The evolving technology of computer autofabrication (3D printing) now makes it possible to produce physical models for complex molecular assemblies. Computer autofabrication technology-sometimes called solid or 3D printing-has evolved over the past decade from a rapid prototyping tool for product design and manufacture to a class of more broadly applied output devices used in many contexts where physical representations are helpful. All of these technologies use a layer-by-layer buildup of the physical part with some method of support for overhangs in the vertical build direction. The great advantage of these methods is that nearly any shape can be built-limited only by the imagination and the structural integrity of the building material.