Sun Shone Brightly on BioITWorld Conference & Expo

Sun Shone Brightly on BioITWorld Conference & Expo
By Steve Meloan
November 2002

The BioITWorld Conference & Expo (presented by IDG World Expo, in San Diego, CA, November 12-14, 2002), gathered leading visionaries from the pharmaceutical, biotechnology, academic, and government markets. They explored ways to better store, analyze, and interpret the avalanche of information being generated by the life sciences industry.

Among the speakers at the conference was Howard Asher, Director, Global Life Sciences, Sun Microsystems, Inc. Asher delivered the November 14th keynote address at the conference.

As the life sciences market leader for UNIX servers, as well as a leader in the realm of high-end graphics visualization technologies, it's fitting that Sun Microsystems assumed a high profile at the BioITWorld gathering. Asher was joined at the conference by such world-renowned bioinformatics heavy hitters as Dr. J. Craig Venter, President, The Center for the Advancement of Genomics. Venter founded Celera Genomics, which shared the spotlight in 2000 for completing the first "working draft" of the human genome.

Many New Challenges

The value of the global life sciences industry already stands at $600 billion. By many estimates, the convergence of IT and the life sciences is poised to virtually revolutionize our health care system -- indeed, our very society. And in this post-genomic era (since the completion of the Human Genome Project), the industry is predicted to grow by even greater leaps and bounds. But those in the know also recognize a gathering dark cloud on the horizon.

The first challenge to be met is simply one of effectively storing and managing the sheer volume of data and information being generated by the burgeoning life sciences industry. "Sun has seldom seen such data requirements in other vertical markets," says Howard Asher. "We are dealing with terabyte, and even petabyte, files."

By some estimates, the volume of data being generated is doubling every six months. "I was visiting a very large multinational drug company yesterday," reports Asher, "and we asked the IT staff about their data requirements. The head of their staff looked at his watch and joked, 'Well, around noon, it was probably 30-40 petabytes. But since it's now 3 PM, it could be higher.'"

Need for New Ways, Better Ways

But the needs of the life sciences and IT industries go well beyond simple storage capacity. The flood of information coming out of high-throughput sequencing efforts -- as well as companion studies in proteomics (the study of protein structure and function) -- includes a plethora of data formats and file types. Not only that, hardware systems are dispersed geographically (located in both the private and public sectors), and there are a variety of operating systems and programming languages -- located on everything from desktop PCs to supercomputers.

This post-genomic information includes numerical data, biochemical-pathway data, genomic data, proteomic data, imaging data, and more. And these wildly divergent data types are not discrete informational domains; in order to truly turn data into knowledge, it becomes necessary to logically interconnect these diverse informational realms. Such interconnection requires new and innovative data architectures, knowledge management facilities, search and analysis applications, and visualization tools. In short, just as important as initially amassing the data, is the task of representing, storing, interconnecting, processing, and sharing the data in a meaningful way.

The Business End

Meanwhile, there are equally daunting challenges in the administrative realms of the life sciences industry. "It now takes 17 years and $800 million to develop a pharmaceutical drug, from discovery to market approval," reports Asher. "Compare that to 8 years and $6 million thirty years ago. At the rate we're going, it could soon reach 20 years and $1 billion. At that point, we're effectively out of the drug development business."

The pharmaceutical drug development process is extremely complex, and highly regulated. "In the development of a drug," explains Asher, "we go from applied research, to development, to third party reimbursers, to the regulators, to the hospitals, to the doctors, and then, at last, to the patient. And in the process, 30-40 million documents are typically generated."

Clearly, there is a great need to streamline and automate the bureaucratic and administrative side of the drug development process -- in terms of more effective processing of records, correspondence, submissions, reports, clinical reviews, and more.

Java Technology: Meeting the Challenge

To bring the promise of the post-genomic era to full fruition, researchers at diverse facilities and locations need to be able to seamlessly share data and knowledge. This requires open architectures that facilitate both cross-platform applications, and cross-platform data. And it is here that Sun technologies and Java technology are almost perfectly positioned.

Sun technology offers scalable, secure, high-performance hardware solutions.

PatternHunter, a genomic search engine, provides a Java technology-powered 2D visualization tool to better enable genetic sequence comparisons between chromosomes, and even between entire genomes.

"It's critical to the life sciences industry that they deal with best of breed systems," says Asher. "And Sun technology is already part of the standard family in the back-office install base of the industry."

Meanwhile, Java technology provides secure, scalable, network-aware, cross-platform software solutions. The Java platform also offers tools and technologies to seamlessly access legacy systems and diverse databases, as well as a rich tool set for processing XML -- which is increasingly becoming the de facto data representation technology for the life sciences.

And in an era when the medical privacy issues being raised by post-genomic discoveries are increasingly in the spotlight, the security of life sciences computational systems is of utmost importance. But here, too, Sun has its bases covered. "Sun knows how to spell security," says Asher, "and has known how to since the company was founded. Security is built into our systems from the ground up."

Dr. Christoph Sensen, Professor of Biochemistry and Molecular Biology at the University of Calgary, standing inside the Sun technology-powered CAVE.

Finally, Sun is an ongoing leader in high-end graphics visualization technologies. The Sun technology-powered CAVE (CAVE Automatic Virtual Environment) is an immersive, graphics-driven system that is increasingly being used in bioinformatics research. Located at the University of Calgary, the system utilizes a Sun Fire 6800 server, four Expert3D graphics accelerators, and Sun StorEdge disk arrays. The CAVE -- utilizing stereo projection displays supplied by Fakespace Systems, Inc.-- provides a 270-degree view (across three screens), with a fourth screen situated under the viewer's feet. Inside this 10-x-10 foot space, researchers can view high-resolution, immersive 3D renderings of molecular, genetic, and cellular images. Such visualization technologies allow researchers to see spatial relationships that would be impossible to discern with conventional imaging.

In short, Sun possesses the technology, the architecture, the knowledge base, and the expertise necessary to lead the increasing convergence between the life sciences and IT. "At Sun, we have people who understand all three legs of the stool," says Asher, "the regulatory leg, the IT leg, and the life sciences community leg. And what is going to make a standing stool, is effectively integrating all three of those realms."

Return on Investment

Unlike the returns promised by many industries, which may be only monetary in nature, the convergence of IT and the life sciences offers the promise of an entirely new paradigm of health care for the human race.

With the knowledge of the post-genomic era, researchers can now begin to conduct "in-silico" computational simulations of living systems -- beginning at the cellular level, on up to the organ level, and eventually to the level of entire organisms. Such simulations offer the promise of predictive models, where the effects of a potential new drug can be tested in-silico prior to any animal or human testing. "We can eventually get out of the business of animal testing, as well as Phase I and II clinical trials altogether," says Asher. In effect, a new drug could be tested (by predictive simulation) upon the entire human race. "With in-silico drug discovery and testing, all we'd then have to do is a confirmational Phase III clinical trial," adds Asher. "So we can drastically reduce both the time and the expense of developing new drugs."

Here are the recent findings (September 9, 2002)-- from the Tufts Center for the Study of Drug Development -- on potential cost savings that would result from boosting new drug R&D efficiency:

• Preclinical screens that, for example, boost clinical success rates from the current 1 in 5
• to 1 in 3, would lower the average cost of bringing a new drug to market (now pegged at $802 million by the Tufts Center) to about $560 million.
• A one-year reduction in the preclinical period would cut $18.4 million from total capitalized cost per approved new drug.
• Cutting one year from Phase III clinical studies would save an average of $71.4 million from the total cost of a new drug.
• A 33% reduction in development and regulatory review time (four years) would decrease average capitalized cost per approved new drug by $167 million (21%).

And we may soon be entering an era of "personalized medicine." It has long been known that responses to pharmaceutical drugs are variable and sometimes unpredictable. (Adverse drug reactions led to more than 2 million hospitalizations, and 100,000 deaths in one year studied.) And it also known that such variability in drug response is in good part genetically determined. Current chemotherapy regimens for diseases such as cancer end up being a somewhat trial-and-error process. But although cancer cells may look nearly identical between two patients, they are actually quite different at the genetic level. Through microarray analysis, which can discern the specific genes turned on in a given patient's disease process, it will eventually become possible to develop personalized drug treatment regimens.

The Future

Sun Microsystems continues to lead the way in the convergence between IT and the life sciences. In 2000, Sun established its Informatics Advisory Council (IAC) to help address life sciences computing challenges and to gain the perspectives of industry leaders.

And in 2001, as a result of discussions within the IAC, Sun helped to form the Interoperable Informatics Infrastructure Consortium (I3C). I3C is a community-driven effort to facilitate data exchange, data management, and knowledge management across the life sciences community, promoting protocols to ensure interoperability in an open, consistent, and robust fashion. More than 40 organizations have contributed to the I3C initiative, including Avaki, BIO, Blackstone, EBI, IBM, INCOGEN, LabBook, Millennium Pharmaceuticals, National Cancer Institute, Oracle, Sun Microsystems, Inc., Whitehead Institute, TimeLogic, and TurboGenomics.

Web Services

The next step is the development of web services. It is generally agreed that an integrated life sciences architecture should be "web-centric." Web services are self-describing, modular, encapsulated functions that can discover and engage other web services to complete complex tasks over the Internet.

Web services workflow using UDDI

Web services communicate using SOAP (Simple Object Access Protocol), which encodes messages in XML. SOAP messages can be sent over HTTP, so web services can be hosted on servers that are already in place. Such web services are defined using WSDL (Web Services Description Language), which is an XML-formatted language used to describe the service's capabilities. Using UDDI (Universal Description, Discover, and Integration) registries, life sciences applications can find needed web services (such as genetic sequence analysis), and web services can even find other web services to perform given computational tasks.

The I3C has now developed a Technical Architecture Working Draft toward a web services architecture for the life sciences. And this architecture maps neatly to the elements of the Sun ONE (Sun Open Net Environment) architecture. Sun ONE is Sun Microsystems' standards-based vision, architecture, platform, and expertise for building and deploying web-based applications and next-generation web services.

Meanwhile, with Sun's active involvement in the Liberty Alliance (a coalition of major corporations dedicated to establishing an open standard for federated network identity) the growing security needs of the life sciences industry will also be handily met. As the company that "makes the Net work," Sun Microsystems is perfectly positioned for the challenges and opportunities of the web-centric, post-genomic revolution. "We offer high-performance computing solutions and technologies that could not be at a sweeter spot for the life sciences industry," says Asher.

About the Author

Steve Meloan, a frequent contributor to java.sun.com, is a writer and former software developer. His work has appeared in Wired, Rolling Stone, BUZZ, San Francisco Examiner, ZDTV's "The Site," and American Cybercast's "The Pyramid."

Back to top