Michael Capellas, the recently appointed CEO at Compaq, implores his troops to build “really cool stuff.” It looks like genomics stuff is included.
GT: Compaq has played a major role in facilitating the sequencing of the human genome, with notable installations at Celera, the Whitehead Institute, and the Sanger Centre. What opportunities do you see for Compaq in the “post-genomic” market?
Capellas: I think it’s a little premature to declare that the “post-genomic” era has begun. While the draft of the human genome has been published, it is still a draft and needs to be brought to completion. Compaq is contributing to this effort with our 100 CPU BioCluster, which we have made available to researchers. And work is just beginning on the many other plant and animal genomes that will be sequenced during the coming years.
We are already involved with researchers working in functional genomics and proteomics, which you could describe as a “post-genomic” market. In addition we are collaborating with scientists working in traditional computational chemistry, especially in the area of large, complex molecules and the manipulation and visualization of these molecules.
GT: What do you predict will be the dollar value of Compaq’s business in the biosciences, genomics, and bioinformatics space in the next ten years?
Capellas: It’s difficult to predict with any accuracy what the market will look like in ten years’ time. It is difficult enough to look out even five years. Current forecasts indicate that the market for computational biology ¯ bioinformatics, genomics, proteomics ¯ will be about $10 billion by 2005-2006. Compaq intends to continue to be the leader in this field, so we would expect to have a healthy share of that potential revenue.
GT: What sort of hardware advances do you believe need to be made in order to address the complex computational challenges presented by genome and protein data analysis?
Capellas: We believe that the best way to provide the high performance computing needed in the genomics and proteomics market is to build flexible and scalable clusters from our standard AlphaServer systems. We provide clusters from as few as two nodes to clusters as large as 256 nodes, with each node having multiple CPUs.
We are already researching the ultra-high-performance supercomputers that will be required to push the frontiers of biological research, including detailed protein analyses and cell systems simulation. Our partnerships with the Pittsburgh Supercomputing Center and the Department of Energy to build 12 teraflop and 30 teraflop systems, respectively, demonstrate the kind of performance we believe will be required in the post-genomics world.
GT: What can you tell us about Compaq’s activities in bioinformatics software?
Capellas: We have decided not to supply our own versions of application code. We prefer to work with the leading commercial software houses and the academic code providers to provide better and better applications.
Researchers at Compaq’s Cambridge Research Labs (CRL) have been collaborating with the MIT Genome Center and other Human Genome Project institutions. CRL researchers play a key role in the development of algorithms for gene recognition, genomic annotation by semantic tags, gap closing in genomic assemblies, comparative genomic analysis, large-scale duplication analysis, and ancient gene family identification and classification.
For the future, we believe that there will be major advances in database technology applied to this arena. We also expect to see faster and more capable applications being written over the next few years.