AT A GLANCE: Received PhD in computer science from Yale University in 1975. Prior to founding TurboGenomics, held position of vice president at Scientific Computing Associates, where he had served as principal investigator on several government- funded research projects in which SCA developed parallel computational chemistry software and visual programming environments for parallel computation.
QWhere will bioinformatics be in 2 years? 5 years?
ARight now bioinformatics is just entering what some have called the “post-genomic” era. I share Nat Goodman’s view that the days of one-by-one analysis are about over and that most bioinformatics analyses will be applied to massive numbers of sequences, proteins, images, or whatever. Clearly the short-term focus will be proteins, but then we see it moving toward clinical applications.
QWhat are the biggest challenges the bioinformatics sector faces?
AObviously, new analytic methods are required to understand proteins and related pathways and cellular functions. Another challenging area is image analysis, where we’re just scratching the surface. Overarching all of this are major challenges in coping with the massive explosion of data: DNA, proteins, images, personal genomes, etc. This isn’t just about database integration, though that’s important; it’s also about computational speed and automation. Many computations just aren’t feasible with current software, because either it’s too slow, or it requires a human in the loop.
QWho do you see as your customers?
AGenaissance Pharmaceuticals was our first major customer. We see biotech and pharmaceutical companies of all sizes as the primary customers for our software, and we are now working with a number them. The university and research communities are also important for us, not only as customers, but also as sources of new technology.
QWhat non-existing technology is number one on your customers’ wish list?
AHigh performance solutions. Everyone wants an easy-to-use enterprise platform that delivers integration and high performance, yet is open and flexible enough to adapt readily to new bioinformatics applications. It’s a difficult challenge that we’re addressing with our TurboBench product.
QDo you see yourself more as a software provider or as a consultant?
AWe’re definitely a software provider, but we do provide consulting services as needed. We particularly like to enter into win-win collaborations where delivering high-value solutions to our clients pays off for us in new software technology and know-how.
QWhat products do you have in the pipeline?
AWe’ve recently released TurboBlast, our accelerated Blast implementation. It exploits parallel computing technology on clusters of workstations, PCs, or Macs. TurboBlast partitions Blast jobs into manageable tasks, runs the tasks on multiple processors using the unmodified Blast programs from NCBI, and then delivers a unified output exactly matching what a single Blast processor would have produced. You get the same correct answer — just very much sooner.
One direction we’re pursuing is development of additional parallel solutions — we call them TurboTools. These will use the same run-time infrastructure as TurboBlast, but Blast will be replaced by other programs. Because of our technology and experience, we can easily accomplish this sort of replacement. Therefore, we expect to have a number of TurboTools soon.
In addition, we are also developing TurboBench, our comprehensive, automated, high-performance enterprise bioinformatics platform. TurboBench will meet the bioinformatics challenges outlined earlier. It will integrate large volumes of complex life sciences data and dramatically speed up all types of bioinformatics processing. All this is made possible by our proprietary parallel and distributed supercomputing technology.
Using TurboBench, scientists can easily exploit diverse data sources and multiple analysis tools, without worrying about such details as program or data location, data interoperability, and communication. TurboBench includes a number of easy-to-use user interfaces. To complete the picture, we’ve built TurboBench as an open platform so that users (life scientists — not expert programmers) can extend it by adding new bioinformatics functionality as they need it.