VCs, analysts, and CFOs discuss the risks and rewards of the genome technology biz
By Meredith Salisbury and Adrienne Burke
Last year, "genomics startup" was synonymous with "IPO," and IPO was code for big money. But as the market shakes out, investors are proceeding with caution. The financial climate has emphasized the dichotomy between technology vendors and discovery partners — and that very divergence could mean the difference between black and red for genomics companies.
For people in the know, the struggle is tools versus targets: which will reap the rewards? GT asked venture capitalists, analysts, and CFOs from both sides of the debate to explain their position.
To round out the picture, we also looked at the major money managers who have invested in companies in the Genome Technology Index (page 30) and, for those startups still seeking funding, we provide a directory of venture capitalists (page 38).
THE VC VIEW
Noubar Afeyan is president and CEO of NewcoGen Group, a firm he founded in 1999 to support ventures in information technology and life sciences. He is also managing partner of Applied Genomic Technology Capital Fund, a venture capital firm with $150 million to invest in the genomics industry. Previously, he was senior vice president and chief business officer of Applera. Afeyan also founded PerSeptive Biosystems, which merged with PE in 1998.
Which do you consider the better investment opportunity, targets or tools?
NA: We as a fund and I as an individual believe in either model. They both will make money. We are spreading our bets across the various areas — to me they’re all variations on the same theme.
Is one strategy riskier?
NA: The advantages and disadvantages are these: When you are developing a tool you’re going to want to satisfy a broad market. The costs of development and validating and supporting it are substantial whether it’s gene-expression machines or mass spectrometers or enterprise software solutions.
In the target validation business model you are able to continue to surf wave-after-wave of technologies as they’re coming out. You don’t have to be completely dependent on the one technology that you’ve got being the best.
In the tools business you have to continue to invest in coming up with wave-after-wave of technologies. The risk — which is not something that makes us not want to invest but makes us very cautious — is that you’re stuck between two sources of opposing pressure. You’ve got people doing the same thing as you, trying to commoditize your technology. And [you have] changing customer needs.
[But] business development in the drug discovery partnership model is quite unpredictable. If you have a product for which there are customers and budget approvals, it’s more straightforward to launch that and take it to the market and support it than it is to build infrastructure that is largely predicated on doing three or four big partnerships.
We tend to shy away from the discovery partnership business model because of that.
You’ve just talked us out of investing in either business model.
NA: Investing and to some extent entrepreneurship is a business that requires cautious optimism. Pure optimism produces the Internet bubble and could produce a genomics bubble if people are not careful.
Why be optimistic? There is an industry that spends 15 percent of sales on R&D, a large proportion of which is going to mine this newly explored genomic information [to find out] how does biology happen from a genetic basis. Companies that allow that will be on the critical path for the eventual transformation of this industry. And if the payoff is big enough downstream, people will pay for it upstream.
You named Millennium as an example of a success in the target discovery partnership service model. Who would you name in the technology vendor category?
NA: Applied Biosystems is absolutely the largest success. PerSeptive was a burgeoning success. I consider Affymetrix a success although [they] have to continue to innovate and improve in order to fight the commoditization war.
I’m not sure who is an absolute winner in the software-selling arena of this business. The issue I have with bioinformatics tools is that it’s not at all clear to me what the question is people want to answer with these tools.
The market will choose which things survive and which don’t. The lead users and thinkers in this area will begin to converge on a method for how informatics can be done at a very large scale, then eventually people will develop an operating system for informatics in the biology world.
Dennis Purcell is senior managing partner in charge of the recently formed Perseus Soros BioPharmaceutical Fund, which has made private equity investments in tools companies GeneFormatics and Structural Bioinformatics and drug companies Axcan, Collagenix, Novazine, and Valentis. Purcell also sits on the Biotechnology Industry Organization’s bioethics committee.
Drug discovery or tools, which is the better business model for a genomics company?
DP: The traditional wisdom has been that [drugs] were the way to go because you had pretty high gross margins, long patent protections, and generally a ready market for the product. If you get a product approved, more likely than not you’re going to have a successful company; whereas on the tool side (though there are less regulatory hurdles and a quicker route to market), we’re going to find out over the next year or two whether there’s a market for all these companies.
What do you consider to be examples of failures or successes?
DP: Historically the larger market-cap companies — Amgen, Genentech, Genzyme, Biogen — were really product-oriented. A lot of companies that started in tools, like Millennium or Human Genome Sciences, have taken their early technologies and advanced them into bringing products for their own account to market. So a lot of tools companies now act as hybrids — they’re trying to sell their products to big pharma but at the same time they’re trying to figure out a way that they can develop products for their own accounts.
Why start a fund like yours now?
DP: In the pharmaceutical industry there are a lot of mergers, and on Wall Street a lot of the firms that advanced this industry have been acquired. Biotech goes through cycles; we just went through a very hot, good cycle in 2000. But now we’re going to see more value than in the past — stock prices are coming back down to where they should be. A number of stocks are trading at very attractive valuations because even though the stocks are going down, the science is still moving forward nicely.
As an investor, what do you look for in a genomics company?
DP: We look for people who are able to sell the technology or do transactions with the pharma industry — not just tech-for-hire transactions but ones where real money changes hands and a big pharma company has a commitment to a biotech company. There’s also a premium on good management. We see whether they have a good understanding of what the risks are in terms of moving the company forward and whether the business plans make sense and are realistic.
What’s the five-year outlook for vendors of genomics technologies?
DP: We’re going to see fewer but stronger companies with more complete product offerings that truly do play a role in lowering the cost, lowering the time, or improving the accuracy of discovering new drugs. We’ll find a number of technologies that are useful but probably aren’t sufficient to have standalone companies; they’re going to become part of larger entities in the future.
We had a very odd situation last year. Two-thirds of all the new IPOs were tool companies, so as a percentage of the total we have more tool companies than we’ve ever had before, and that’s something new to the industry. Over the next couple of years we’re going to see a renewed focus on the product companies simply because so many of them are coming to late-stage clinical trials. There’ll be so many new biotech products in the market that investors will revert back to focusing on the product side of the equation.
THE ANALYSTS’ ANGLE
Michael King is an analyst with Robertson Stephens who has specialized in the biotechnology field since 1985. Of the stocks he predicts will outperform the market this year, only two are genomics companies: Human Genome Sciences and Affymetrix.
Which is the better business, drugs or tools?
MK: The market has voted on that, and the answer is drugs. For proof I would point to the valuations of Human Genome Sciences and Millennium. That doesn’t mean, however, that tools are a bad place to be. I think Affymetrix is clearly a case where if you’ve got the right set of tools (or ABI for that matter), that could be a very good business as well. But if you had to choose one, I’d say drugs. Drugs can become billion-dollar products; there are very few tools that can become billion-dollar tools.
You’ve already mentioned Human Genome Sciences and Millennium as success stories. Any other examples?
MK: CuraGen and Myriad might be examples of success because they own the rights to a number of their targets. I would also say companies like Medarex and Abgenix — I think it’s legitimate to think of those guys as therapeutics-based genomics companies because they have a methodology for getting very rapidly from a target right to the drug.
What about success stories for tools companies?
MK: One of the stocks I picked for the year is Affymetrix because it’s driven by revenue growth and they have real profits. There are good values out there, like Sequenom and Orchid — the stocks are cheap, but their performance will be somewhat hampered by the process that investors are going through this year, which is sorting out winners from losers. Orchid and Sequenom are going to be winners, but the process is going to take most of the rest of this year.
What do you look for in these companies?
MK: I want to make sure there’s good technology there, and then see how that drives the business model. That’s why I think that HGS has such a good business model, because they’ve got strong technology and they’re keeping a lot of the rights to the drugs for themselves. Because of their proprietary position they have an ability to receive royalties on the small molecule uses of their targets.
A business plan can be great in theory, but if there’s no underlying technology then a business plan is out the window. A good technology can have fits and starts, but if the core of it is strong then it can overcome whatever shortcomings there are in the business model.
What’s your five-year outlook for genomics companies?
MK: The HGSs and the Millenniums are going to be around for a long time. For a lot of the other ones, there isn’t going to be a lot of business left over. There’s a two- to three-year window in which companies will be able to sell a lot of products, and then at some point we’ll reach saturation. Then the market grows like a mature market.
There’s certainly a first-mover advantage. So if you’re not there now with a product that’s available now that works and that people are using either to genotype patients in clinical trials or do target discovery, I think you’re going to be left behind.
Scott Greenstone covers genomics and proteomics technologies for Thomas Weisel Partners where he is vice president for equity research. He has been following the analytical instruments industry for seven years.
What do you make of the trend of genomics tools and data companies moving into drug discovery?
SG: These companies realize that they make some good money delivering data and tools to help other companies find IP and develop drugs, but as they become experienced and gain a domain expertise they realize they can capture some of that IP themselves.
This is why it’s not easy to say which is the better business strategy. It plays to different risk-return characteristics of the investor, the shareholder base. Some companies are trying to capture both. Not all of them are ultimately going to be drug discovery companies. They’ll end up just having this IP.
Where do you see your technology companies going in five years?
SG: We’re at the forefront of a long growth cycle. I say that because of the vast sums of information that have yet to be identified and still need to be. Will companies continue to buy tools at the pace they have been? Clearly not. Right now they’re starving for gene chips and mass specs and a lot of screening instruments. There just aren’t enough of these tools out there. Once you get them out there you’ll start getting into steady-state growth and replacement cycles. What will drive the industry then will be improvements in the technology.
Is this what’s already happening with high-throughput sequencing instruments?
SG: We’ve probably hit that cycle for sequencers where it just doesn’t make sense to build these big high-throughput factories any more. That’s not to say that sequencing is done. There’s still a ton to do, but you don’t need to build up these massive scale systems. Applied Biosystems will still sell the 3700, but they won’t sell a lot of them. In 1999, they sold 1,100 at $300,000 a piece. That’s by far the biggest product introduction ever in the instrument industry.
The real driver that some people forget is that ABI might not sell the instruments anymore but they’ll sell the reagents, and the reagents are a higher margin business. It’s the razor/razor-blade model. You can’t do sequencing without reagents. The $300,000 sequencer running at full capacity will probably take $100,000 to $150,000 of reagents a year.
And that’s just for one sequencer. If you’re running 30 or 50 machines…
SG: You’re spending a lot of money on reagents!
The interesting thing is that from the time the 3700 and MegaBACE were introduced in the fall of 1998, the solution to the sequencing problem became clear and easy. That isn’t going to be the case for the next stages of genomics and proteomics. There are so many more proteins — much larger and much more diverse molecules, and much harder to deal with. Mass spec, NMR, and x-ray crytallography are the hardware technologies used to look at those, so right now they’re in a big up cycle. Ultimately that cycle will turn, but it will last longer and will take a longer time. It took about two years before the sequencing market fell out of bed. The mass spec market is going to be up for many years.
THE CFO OUTLOOK
Stephen Parker is CFO of Oxford GlycoSciences, which raised $215 million last year in an equity offering. He has also worked as a strategic consultant to the pharmaceutical industry and most recently worked as a healthcare industry investment banker in London.
How did OGS settle on a target-discovery over a tools-vendor strategy?
SP: OGS has had a strategy from the time of its IPO in 1998 to pursue a platform into products. Recent changes have been more a question of the implementation of that strategy rather than a change of strategy per se.
Once it became clear that the genome was going to be published substantially ahead of original estimates, we set up what we termed “the landgrab strategy” last spring. The aim of that was to identify 4,000 disease-associated proteins and their genes and file applications for patents on those to form the basis of our proprietary R&D activity.
What was the financial logic behind that decision?
SP: We wanted to have some space in which we could operate and develop drugs — for the most part to proof of principle — and out-license many of them. And the sorts of patents we are filing — the antigen, its gene, and the disease in which it is over or underexpressed — are the sorts of patents that the courts uphold as rigorous.
Unless you have a proprietary target in the new postgenomic world, you are struggling. By having targets that are proprietary and upholdable in the courts, anyone who wants to discover drugs is going to be paying us a royalty associated with that target.
Do you expect more of your revenues to come from what you do with those targets internally or from what you do with pharma partners?
SP: Having found the targets we are faced with a simple choice: We can either out-license them to big pharma and get very little or add value to them, discover the drug or the antibody, and take it through proof of principle. Clearly the economics of the discussion at that point are different.
That underpinned the rationale for our fundraising of last year, which [gave] us enough of a war chest that we could develop substantial numbers in our own discovery and development programs.
Did you ever consider just selling tools directly to drug companies?
SP: I don’t think we ever considered selling the tools themselves, although in the proteomics space we had to build the tools. We have a very good engineering department that has built a lot of the robotics stations and 2D electrophoresis systems. One reason genomics took off faster than proteomics did was that you could go out and buy a gene sequencer off the shelf. You’re starting to see product coming through now in the proteomics space.
Will you also apply your technologies for custom discovery?
SP: We remain committed to collaborations with pharma and we do that for two principle reasons. At one stage, of course, we did it for cash, as everybody does.
We choose to continue to collaborate because in many cases the tissue samples we receive have been collected with informed consent within clinical trials, so we have direct access to clinical samples, including tracking the progress of disease.
Also, we can earn a great deal of experience in handling those diseases and the medical approaches toward those diseases. In those instances we most frequently are searching for either targets or biomarkers for diseases.
Nurit Benjamini was promoted last August to CFO of bioinformatics-firm Compugen, which raised $36 million in private placement funding and another $53 million in its August 11 IPO. Compugen is also embarking on its own discovery research.
Which do you believe is the better business model and better investment opportunity for a genomics company: applying proprietary tools in-house for target discovery or selling the technology?
NB: It is not a question of either/or. There’s still doubt as to what is the optimal business model. Now the market values the drugs companies — those have higher valuations than tools companies. It’s like with hardware and software; until you have hardware that works there isn’t any value for the software, but as soon as the hardware works the software is very valuable because it can enhance the information or the technology that you have. In the current situation the drugs companies have a higher valuation but in the future it might change. Or in the future a combination of both will be of higher valuation than just the big drug company.
What are examples of successes or failures in the industry to back up your view?
NB: It’s difficult to speak about failures. I have two good examples for success: one is Human Genome Sciences, and the other one is Millennium. Human Genome Sciences, together with a consortium of five companies, is developing drugs based on the Human Genome Sciences proprietary database, which is one of the largest that exists. Millenniun started by building a drug discovery platform and worked with external companies — one famous one is Bayer — and now they’re developing drugs based on that platform.
As a financial officer, what do you think is important in a genomics company business plan?
NB: There’s the question of whether we can sell the technology we develop, because if we’ll take it and utilize it just in-house, it might be that a few years from now we’ll have a much more valuable product pipeline than if we sell the technology on the market. But we need to know that we have a killer application, that our model is stable, and that we have a pipeline of products so that our revenue can grow. Companies like us have to invest a lot in research and development and also marketing and sales until we really become profitable, so my concern is also having enough money in the bank to build our business model.
What’s the five-year outlook for genomics companies?
NB: The winning companies will be those that provide integrated solutions that create a clear biological discovery value. We believe that niche companies will find it very difficult to succeed. You need to be not a niche company and you need to offer a diversity solution to tackle all the things that drug companies will have to face you with three to five years from now.
I can tell you this for sure: A bioinformatics company that doesn’t have any molecular biology lab will find it very difficult to develop tools for a drug company or for other biotechnology companies.
Genomics companies will have to have killer applications that the drug company will not be able to work without — but they will also have to either collaborate or go more in the direction of biological discoveries.