Visualization in the Life Sciences in the Post-Genomic Era
The data sets generated by these techniques have grown steadily in size and number. This explosion of data generating and processing capability made the Human Genome Project feasible. And now that many key genomic databases are complete, the automated sequencing process has become a given, and current state-of-the-art research efforts are focusing on finding value from this data. We are now in the post-genomic phase of this research.
The emphasis in life sciences research now rests on managing, understanding and utilizing the massive quantities of genomic data. Through disciplines such as bioinformatics and molecular modeling, the structure, function and disease associated with specific genes can be identified. A key to developing approaches efficiently to treat disease rests on the ability to model proteins and small chemical molecules (putative drugs). These in silico interactions between the models of proteins and small molecules are used to predict protein-to-protein or protein-to-chemical interactions in the body. Such calculated interactions help researchers efficiently select agents to evaluate on the basis of their high probability of therapeutic success in vivo. This is the ultimate objective of post-genomic research.
Visualization technology: the next major research tool
The large molecular databases now available were built up through the power of high-performance computing (HPC). We have now moved into a different phase of this research that requires a different set of tools. Because the shapes of protein and chemical molecules determine their functions, we are intensely interested in their 3D structures — how they fold into their final configurations and how they interact with each other. To understand these processes well enough to make useful decisions in an acceptable timeframe, we have to visualize molecular structures and interactions, and, for this, more than HPC is needed. Once visualized, molecular models move from the abstract to the readily comprehensible concrete. It is now clear that the ability to see and manipulate three-dimensional molecular models clarifies functions and speeds insight. That is why visualization technology has become a key to the post-genomic phase of life and chemical sciences research.Visualization supercomputers can display very large 3D molecular data sets in interactive environments. Protein and chemical molecules can be visualized, moved and rotated in stereoscopic environments like real-world objects, with little or no time lag. In the past, researchers looked at separate images of molecules and attempted to visualize intuitively the ways in which they could interact. They can now bring three-dimensional proteins and small molecules together on the screen to simulate interactions.
The remarkable ROI of Immersive Visualization Systems
Interactive visualization of molecular models on desktop workstations, while enhancing productivity, does nothing to alleviate some of the basic inefficiencies of the drug design process. The traditional research scenario consists of two or three scientists crowded around a workstation, manipulating models in an attempt to reach consensus. This process leads inevitably to the development of ad hoc groups bound by a common opinion, to contention or competition between groups, and ultimately to a splintered and disparate approach to key research issues. This process resists efficiency, raises costs and eats up valuable development time.Visualization technology is resolving these long-standing problems. Research organizations are using visualization supercomputers to display stereoscopic 3D graphics images in room-size, theater-style visualization centers. Stereoscopic visualization simulates a total immersive experience in which 3-D images appear to leave the screen, allowing the audience to interact with the objects as though they are "real." In this environment, colleagues and interdisciplinary teams can meet, manipulate images and discuss results collaboratively. A typical environment presents images on a large, rear-projection screen, with the audience viewing and interacting with the data through stereoscopic glasses.
The return on investment from visualization technology expresses itself in several obvious ways that result from an improved process. One is decision quality. The interdisciplinary team, manipulating molecular models in real time, generates synergy and insight that cannot be present in other environments. Another saving is in the time required of each team member to contribute to the decision, because teams come together in the visualization center with the expectation that they will reach a decision before they leave. This leads to reduced time to discovery and to market.
The recent development of new technologies is taking this return on investment to a much higher level. The whole phenomenon of consolidation-mergers, acquisitions, and globalization has produced enormous corporations with very large research staffs and facilities, spread across campuses, regions and the globe. Research teams include members who are multiple time zones apart. Bringing them together has become a major cost of research; travel fares, accommodation, lost time and fatigue will soon make the price of routine travel prohibitive.
A new technology known as Visualization Area Networking (VAN) enables scientists at geographically separated visualization centers to view and manipulate the same graphics simultaneously. VAN, which provides a framework for distance collaboration networking, is a business process-focused tool that combines graphical supercomputing technology with high-bandwidth networks to facilitate productive meetings. Scientists can now share visualizations of very large data sets with other users anywhere, anytime, on any device, in real time. Every key specialist, regardless of location, can be brought into the decision-making process.
Big rewards for early adopters
Although visualization techniques are rapidly being adopted by leading researchers in bioinformatics, genomics and computational chemistry, these technologies have already been widely adapted into the product development processes in other industries. Collaborative visualization is already an indispensable part of automotive design, aerospace design and petroleum exploration processes worldwide. In these fields, improved decision-making, time to market advantages and overall ROI were the ultimate reason behind its adoption.The adoption of visualization technology followed differing paths for each of these applications. For the oil and gas business, scientists were the first to realize the potential of visualization in understanding their huge data sets, and they embraced the technology in parallel with their adoption of supercomputing techniques. In automotive design, engineers were the first to realize the enormous productivity potential of the technology and were quick to adopt it. In the pharmaceutical industry, however, researchers who are looking for ways to enhance the drug-discovery process and work more efficiently must first establish the ROI of visualization techniques as applied to the life sciences. They need to understand how their existing technologies can become more productive and efficient through visualization.
Protein-folding studies at the Delaware Biotechnology Institute
For the early adopters, the rewards have been great. Just one among many examples comes from the academic community.Dr. Yong Duan of the University of Delaware's Chemistry and Biochemistry department uses visualization tools at the Delaware Biotechnology Institute (DBI) to study peptide-folding processes in amino acids. Duan's folding simulations, which depict events that take just a few nanoseconds in the real world, may require a computer run of five days or more on a single-processor computer. Duan pauses regularly during the simulation to capture the current status of the configuration of the folding peptide. He then combines these captured positions into a movie that can be displayed in immersive stereo on a visualization supercomputing system. Starting from a straight chain conformation, Duan simulates all-atom models model in continuum solvent. Much of the rapid initiation of the folding process takes place within one nanosecond, while unfolding of intermediate states, which constitutes the rate-limiting step, takes about 10 nanoseconds, followed by the formation of helix-turn-helix and the completion of the folding. It is striking and highly informative to watch the 16-amino-acid straight-chain molecule fold, contort, and gradually assume a helix formation.
Other visualization techniques being developed at DBI (www.dbi.udel.edu) include remote immersive collaboration, immersive viewing of in vivo biological images, and a new approach to the visual data mining of biosequence data. These and other innovative uses of visualization technology in the life sciences were discussed at "Delivering Technology Leadership for Life Sciences," an executive summit for biotechnology professionals hosted by DBI and SGI on October 3, 2024 at the University of Delaware. The meeting addressed trends and challenges in high-performance computing and visualization for life and chemical sciences as it relates to drug discovery and bioterrorism research.
A technology whose time has come
We have refined and automated the prodigious number-crunching processes that accomplished the goal of the Human Genome Project. We have moved into a phase of life sciences research that focuses on gene function, which inevitably means the study of three-dimensional molecular structure. Interactive visualization is the most effective tool for this work. It reduces time to insight and accelerates decision-making with realistic displays that can be manipulated by collaborative research teams in real time.Summary
We are in the early phases of adoption of this technology, but its acceptance in the life and chemical sciences is moving very quickly. We have observed the adoption curve for industries where visualization has become an embedded part of the process, a mission-critical technology that is required for successful competition. The pattern is clear: investigation, demonstrations, establishment of ROI benefits, acceptance. It is our expectation that visualization technology will soon become an integral part of the life sciences laboratory infrastructure.About the author
Dan Stevens, Ph.D., is Market Manager, Life Sciences, at Silicon Graphics, Incorporated.More information is available from:
