CAMBRIDGE, Mass.--When Nat Goodman talks about how the life sciences industry should support academic efforts in computational biology, he's not just promoting altruism. "There is an urgent need to make sure that academia is working productively in pursuit of important new computational problems," Goodman told BioInform. "It's of strategic importance that academics be nurtured to work on new generations of problems."
As the director of Compaq Computer's new bioinformatics solutions center here, Goodman is equipped to get the word out. Here, in part-two of BioInform's exclusive interview, Goodman, formerly a senior scientist with the Jackson Laboratory, shares his ideas for nurturing the bioinformatics field.
BioInform: The shortage of computational tools for genomics research that you described sounds like a much bigger problem than even Compaq will be able to make a dent in.
Goodman: We have to play a role in focusing industry's attention around this larger problem. It's going to require commitments from all of the key players. Pharmaceutical companies are going to have to do more to invest in the academics who are developing these new methods. That's the real critical element here. Most of the new statistical and algorithmic work, say to analyze single-nucleotide polymorphisms or gene expression, is being done in academia.
At the same time, there's been this steady brain drain out of academia. It's important that we work not only to reverse that, but to make sure the academics have the support and the opportunities they need to work on these new kinds of data that are becoming available.
What's necessary to make that happen is not rocket science. Academics work on interesting new problems. They're motivated by the intellectual challenge of it all, they also need the money to do it, and most importantly, they need access to the data. And that's one of the biggest problems to date.
In sequence analysis there has been a long tradition of making data available, so academics have had access to lots of sequence data for the last 10 years or so. In most other areas--gene expression, proteomics--they can't afford the chips, and companies won't give them the data. There are companies in the DNA-array field that restrict their customers from publishing raw data that comes off their instruments, which means academics can't see it. So, an academic can't do a collaboration with a company and say, You give me the data and I'll analyze it, because that's prohibited by an intellectual property agreement.
One has to appreciate that in order for academia to work you have to present the academic community with hard problems and access to lots of different kinds of data, and freedom to pursue the work wherever it goes. If you're a company, you have to be willing to let them draw negative conclusions about your methods or your products.
BioInform: Does Compaq have specific plans for contributing to academic research in bioinformatics?
Goodman: We do contribute a fair bit already--we make computers available to many of the major academic software developers and people who are developing new methods--but we have to do a better job. Part of my role is to make sure all of the right people doing the important work are being supported. Another part is to provide forums for that knowledge to be communicated to industry and for industry to communicate back to the academics.
BioInform: Is support for the bioinformatics software industry another part of the equation?
Goodman: Finding the right vehicle for software companies to succeed is just as much a challenge that we have to face up to. I think it's fair to say there are no terribly successful bioinformatics companies that have moved to the front of the pack and gained dominance. It's been difficult for them to get investment. But even beyond that, it is necessary, along with nurturing academia, to help them understand how they can get past the current barriers and become more successful.
The academics have to create the knowledge, the algorithmic methods, and then there has to be an industry that can turn that into robust products that customers can use. Finding a way for that industry to succeed is going to be a key element.
BioInform: You don't believe pharmaceutical and biotech companies will just decide to develop their own computational tools and solutions?
Goodman: That's certainly what is going on today, but if we look down the list of all the different kinds of software that are needed across all the different kinds of data, it's too daunting. Companies cannot afford to do this themselves. I estimate that if a company wanted to do that they would need to put in place something like a 200-person bioinformatics software organization. That would be the entry-point for being self-sufficient.
Look at all the different kinds of data and software they have to deal with. It's databases, algorithms, visualization, and putting the components together to form more complete systems. And, then you have to do that for genomes, gene sequences, gene expression, protein expression, genetic maps of the SNP data, physical maps, protein-protein interactions, and pathways. And, of course, there are model organisms that you have to bring in too. It's really a lot of work.
So, there has to be a software industry and it has to become successful, but there are a lot of impediments and it's not obvious how they are going to succeed.
BioInform: When the SNP Consortium was formed by several major players in the pharmaceutical industry, frustration over what they were being charged by database companies seemed to be a driver. Is there a chance that those big companies would similarly fund academia to do software development, thereby eliminating the need for a bioinformatics software industry?
Goodman: One possible outcome could be a consortium approach. The SNP Consortium is an exciting development. We need to wait and see how well it can compete with the private sector. It's an important new direction that the industry might go in, but even if it proves successful on the lab side, I don't know whether it would be successful in software.
On the laboratory side, the SNP Consortium was able to fund three top genome centers that could, by themselves, create the resource they set out to create. This won't work on the software side, because you have to go out and get the software from the dozens of academics who are developing it.
Still, I suspect a consortium could be the answer to a lot of these problems. It could be the way out.