CAMBRIDGE, Mass.--Since 1990, Ken Fasman has practiced bioinformatics in the public sector. He headed sequencing informatics at the Whitehead Institute/MIT Cen ter for Genome Research here for the past 18 months, and for the preceding seven years, he ran informatics at the Genome Data base at the Johns Hopkins Univer sity.
Earlier this month Fasman, 39, started a job as director of informatics at the new Astra Bioinfor matics Center here, an expanded spinoff of a bioinformatics center established three years ago in Lund, Sweden, by the Stockholm-based pharmaceutical firm Astra.
Fasman recently spoke with BioInform about his new role at Astra, his decision to leave Whitehead, and his bioinformatics career.
BioInform: What's the mission of the Astra Bioinformatics Center?
Fasman: The Astra Research Center Boston (ARCB) was established about three and a half years ago with a very focused mission: to develop new therapeutics to eradicate Helicobacter pylori, the bacterium implicated in peptic ulcer diseases. Astra bought access to the H. pylori sequence from Genome Therapeutics several years ago to allow ARCB to exploit that sequence as one means of developing therapies against it. It's been quite successful.
This group has been a model for how genomics and bioinformatics will affect drug discovery and development throughout the rest of Astra. Based partly on its success and the desire to tap into the US and Canadian academic sectors, Astra is building a large R&D campus on the outskirts of Boston, in Waltham, Mass. We're breaking ground in the middle of May and will start occupying buildings on that site two years from now. That's where my center will be. Within five years there could be as many as 400 Astra research employees at that site. It's going to be a major effort.
We may end up renaming the center the Astra Research Informatics Center. That's an effort to indicate explicitly that this is more than just bioinformatics, more than just molecular biology and genetics. Within the mandate of the center will fall chemoinformatics, structural biology, and chemistry, and, over time, essentially all research computing within the R&D units of Astra.
Astra, like other pharmas, already has a substantial effort in chemical information systems and chemoinformatics in place. Integration of this new program and the existing ones will not happen by simple fiat, but rather by the concerted collaboration of everyone involved.
BioInform: Is this job your first foray into private industry?
Fasman: I had my own consulting company from 1988 through 1991. That's how I got involved with the Genome Database in the first place: my first year of involvement was as a consultant.
BioInform: You must have been one of the first bioinformatics consultants.
Fasman: I guess so. That wasn't even the term I used back then. We called it "research computing."
BioInform: What about this jump to industry? You have had some experience in private industry, but was it a hard switch to make?
Fasman: Not really. I've joked with so many of my colleagues in this field that we're all going to end up working in the pharmaceutical companies at some point or another. Of course I certainly have a few friends who are hoping never to leave their current positions, but this is where so much of the activity is going on now, where so many of the resources are focused.
Switching back and forth from academia to industry strikes me as being less and less of a problem. I know a number of people who have moved in both directions. It used to be, when I was in college and graduate school, that once you left academia for industry that was the end of your academic career. But I know a growing number of people who have come back to academics from industry and done quite well.
Part of what's going on now is just this tremendous pressure in terms of the shortage of experienced people in this field. The demand in industry is so great that there's a tremendous pull in that direction, and it's going to be a problem in terms of training the next generation of people in bioinformatics.
BioInform: In your corporate capacity do you foresee doing any work to promote bioinformatics training in academic institutions?
Fasman: Absolutely. The center is going to be looking for every opportunity to encourage bioinformatics on the academic side. Partly for our own selfish reasons, to ensure that we have a steady flow of people to draw from, but also because so much innovation goes on in the academic sector, where people are less concerned about near-term deliverables. So it's very important to me personally and to the success of the Astra informatics center that there be strong and growing academic programs.
BioInform: What motivated you to leave Whitehead to go to Astra?
Fasman: The challenge of working on informatics on such a large scale, the ability to address the integration of both chemical and biological information systems. It was a chance to do something on a much larger scale than I was able to do at the Whitehead.
BioInform: What's your educational background?
Fasman: I have a BSE in electrical engineering and computer science from Princeton, and a PhD in biomedical engineering from Johns Hopkins School of Medicine. My PhD training was in neurophysiology. I have some background in molecular biology, genetics, and chemistry, but as we expand we'll be bringing in real experts in each area, many of whom will come from existing chemical informatics groups and projects within Astra.
BioInform: How many people are you planning to bring aboard?
Fasman: The initial group will be 8-10 but that will probably grow pretty quickly as we expand into each of those disciplines. The notion here is to deal with the problem in reasonable pieces. So our initial focus, no surprise, will be genome informatics--getting that uniformly implemented throughout the company and making sure that researchers in all of our different research units have access to the same quality of genome informatics analysis. As that project starts to reach maturity we will expand into the structural biology, chemistry, and chemoinformatics. We're trying to tackle these things one front at a time.
BioInform: Are there any models in industry for bringing bioinformatics and chemoinformatics into one department like that?
Fasman: Most of the companies that I'm aware of, when they create a bioinformatics department, are really talking about what I'm calling here genome informatics: informatics in support of molecular biology and genetics. The chemoinformatics is done somewhere else. In fact, in most pharmaceutical companies that group has been around much longer and probably for most of its history hasn't been referred to as informatics at all.
The word informatics has spread like a weed, so that what maybe was called computational chemistry in the past or chemical information systems is now chemoinformatics. Everything is something-informatics: bioinformatics, chemoinformatics, I'm sure there'll be 10 other informatics within a year's time.
BioInform: Do you foresee a trend of pharmaceutical companies bringing those together in one department?
Fasman: There's certainly a demand for it. For example, companies have large compound databases where they track all of the different small molecules that they test and to which they have proprietary access. They have other databases where they store their genomic information about potential targets. But bringing those datasets together, having tools that can help you look for large-scale trends and building a data warehouse that combines the chemical information and the biological information, ideally requires a group that has some kind of mandate within the company across those disciplines.
This is an issue that a number of companies are wrestling with. We are hopefully well positioned to deal with it by virtue of the company recognizing that it is an integrated effort from the very beginning.
BioInform: Did the founders of Astra's original bioinformatics center in Sweden start out thinking that way?
Fasman: Yes, but they were having difficulty recruiting foreign bioinformatics experts to move to Sweden. One of the reasons for moving this effort to the US is to take advantage of a larger pool of potential recruits.
BioInform: How about hardware and software purchases? Are you in the process of building that infrastructure?
Fasman: Absolutely, but as you can appreciate, I've only been on the job for eight days. It's pretty early in the process. I can say that I am looking very hard at purchasing as much as possible.
I know some bioinformatics groups have decided that they've got to build their own everything. We're just starting to see enough maturity among the products from the bioinformatics companies--Pangea, Incyte, Base4. These products are far enough along now that one can actually start to think about building a serious infrastructure on top of them, instead of duplicating those efforts. I would not say that they've all reached full maturity, but we're going to be looking to do as much as possible with off-the-shelf solutions.
BioInform: Are there ready-made answers for a group like yours that wants to put chemoinformatics and bioinformatics in one?
Fasman: No, there's no single product that pulls it all together. There will definitely be a lot of integration work on our part. Some of the vendors are already recognizing that this is a need and beginning to branch out in this area, but I wouldn't say that any of them is very far along. Some companies, Oxford Molecular is one that comes to mind, have been running parallel efforts in chemical and biological information systems. So they're obviously positioning themselves to deal with this issue.
BioInform: How about on the hardware side? How do you allocate your resources, for example between workstations and PCs? Have you found any hardware vendor that you feel is particularly eager to work with you or who is especially well versed in the specific challenges of bioinformatics?
Fasman: Our standard platform is Microsoft NT workstations on the desktop with X server software installed, and larger, multiprocessor Unix systems for web, database, and computer servers. Many of the Unix vendors are vying for attention in the bioinformatics and pharmaceutical sectors. We don't favor a particular one.
BioInform: What about the bioinformatics-specific hardware from companies like Paracel or Compugen or TimeLogic? Do you see a place for that type of equipment in your lab?
Fasman: We are still evaluating the requirements within Astra for hardware accelerators for sequence alignment and searching, but I believe they could be justified for some of our efforts.
Coming June 8 in BioInform: Part 2 of our exclusive interview with Ken Fasman in which he discusses his top priorities at Astra and the company's bioinformatics future.