It wasn’t long after announcing the completion of the sequenced human genome that Craig Venter, Celera Genomics’ maverick president and chief scientific officer, again generated splashy headlines with an announcement that his company intended to become the star of the burgeoning proteomics market.
And recently Venter dared to take one major leap further by revealing plans to develop a cancer vaccine.
“We are really trying to do something about cancer,” Venter told BioInform.
The first step toward that goal will take place in September when its sister company, Applied Biosystems (until recently PE Biosystems), delivers its Tandem Time of Flight machines to Celera. These machines
cansequence 3,000 samples an hour, 10 times more than the machines currently on the market.
At a later stage Applied Biosystems plans to develop yet another machine that will be able to sequence 30,000 samples an hour. With this level of power, Venter said Celera would be able to generate the information necessary to understand proteins’ links to disease.
“Applied Biosystems thinks that in a couple of years, it will be able to do 30,000 sequences per machine an hour,” said Venter, adding that Celera wants to buy enough machines to measure as many as 1 million protein sequences a day.
“That’s pretty radical,” said Venter.
A New Proteomics Center
Using a portion of the $1 billion it raised earlier this year, Celera has already started building a proteomics facility at its Rockville, Md., headquarters, an operation that will be the equivalent of what it developed for DNA sequencing.
Accelerators from Paracel will likely be integrated with mass spectrometry machines to speed the proteomics calculations.
Venter said that Gene Myers, Celera’s vice president of informatics research, and his staff, who were credited with developing the algorithms needed for sequencing the human genome, have also begun devoting some time to devising proteomics algorithms. New algorithms will be needed to associate the results from mass spec sequencing with Celera’s human genome database to understand the links to disease. This will require “tremendous computational power,” Venter said.
Although Celera will not have first mover advantage in the proteomics and drug discovery sectors, Venter said that he is not afraid of the competition because few companies are set up for such a high-throughput proteomics effort.
Celera will also have the benefit of getting a sneak preview of Applied Biosystems’ technology due to an early access partnership between the two companies.
Stephen Martin, director of Applied’s Proteomics Research Center at the company’s PerSeptive Biosystems division in Framingham, Mass., said that the deal for the proteomics sequencer is similar to the one for the ABI 3700 genome sequencing machine.
Celera and Applied Biosystems will also join forces on the bioinformatics front. Noting that bioinformatics is even more important in proteomics than it has been for genomics, Martin said at least a third of the people in the proteomics initiative would be bioinformaticians.
Although mass spec and sample preparation are critical, “being able to take that data and give it some context is at least a third of that total process,” he said.
New software tools will also be needed, including basic database and data transfer tools and software algorithms that allow users to compare data from proteins with that of DNA, SNPs, and clinical data.
Martin said Celera’s wet lab employees would be working at the Framingham center, which was founded in March. Some of the PerSeptive Biosystems people will work at Celera’s Maryland complex on the informatics.
Looking for Antibodies
Once enough training has been done, Celera will sequence serum samples from people with different kinds of cancer to look for new surrogate markers. Such markers will enable new tests analogous to the prostate-specific antigen (PSA) test for prostate cancer to be created. Venter noted that there are no good markers for colon and breast cancer, for example.
Celera plans to make antibodies for proteins present in cancer tumors as part of its effort to understand which are the most important ones to change to fight cancer. Venter said such antibodies would help Celera develop new cancer vaccines. Venter added that Celera does not have any plans to develop small molecule therapeutics or protein therapeutics like insulin.
Even if Celera moves methodically to ensure its future success, some industry players said that the company will have to move fast to catch up with competitors in drug and diagnostics development.
Kenneth Conway, president of Millennium Predictive Medicine, a subsidiary of Millennium Pharmaceuticals, said Celera would likely need to find partners in order to catch up.
“They’re going to have to partner, build a broader technology platform, and it’s going to be challenging,” said Conway. “If it were easy to do, drug companies and diagnostic companies would not be coming to companies like Millennium to access our platform.”
Millennium, which for the past seven years has been building disease-specific biological expertise and trying to understand the underlying causes of different diseases, has developed platforms that include transcriptional profiling, proteomics, mouse modeling, and bioinformatics technologies.
Beyond discovering personalized vaccines and new markers, companies need to get tissue samples and prove in clinical trials that the concepts actually stand up, Conway added.
Venter said Celera has not yet decided whether it would do clinical trials by itself or through a partnership. The company does, however, want to leverage its information business so that the vaccines it takes into clinical trials will have a good chance of succeeding. Venter said he wants to avoid “biotech roulette,” investing tremendous resources into a single small molecule drug, which may only have a one in 20 chance of successs.
Ruedi Aebersold, a co-founder of the Institute for Systems Biology in Seattle, was optimistic about Celera’s ability to apply what it knows about generating, managing, and analyzing large amounts of data from genomics to proteomics.
“They will have to of course re-tool and tackle different quest-ions. But many of the elements are the same,” said Aebersold.
The main difference is that genomic sequencing has a defined endpoint — the linear sequence of all the nucleotides in a genome. Proteomics, meanwhile, is a dynamic measurement where the composition of the proteins in a cell or tissue changes all the time, said Aebersold.
Aebersold is collaborating with Applied Biosystems to develop improved instruments that will have higher sensitivity and throughput to enable further automation of proteomics.
Cancer vaccines are an “obvious application” for pro-teomics, Aebersold said, since that information could help scientists design a vaccine directed against one or several proteins that are expressed only in the cancer cell. In this way, a person’s white blood cells could be trained to attack cancerous cells.
Gene Logic Shifts Away From Services to Focus on More Lucrative Content, CEO Says
A while back, BioInform spoke to Gene Logic, a developer of genomic expression databases, about the company’s decision to build its Data Logic bioinform-atics business. At the time, then CEO Michael Brennan, now Gene Logic’s chairman, said he expected the Data Logic division to grow into a $15 million annual business.
Now, less than two years after that announcement, Gene Logic of Gaithersburg, Md., has a new CEO and a new plan that includes putting bioinformatics on the back burner. Instead, the company will focus primarily on selling subscriptions to its GeneExpress database, a move intended to help the company to increase revenues and narrow losses (see Money & Finances section).
To date, Gene Logic has analyzed about 50 percent of the human genome based on thousands of healthy and diseased tissue samples that the company of 230 employees has collected.
BioInform recently caught up with the new CEO, Mark Gessler, 38, to discuss the move toward content as well as what Gene Logic plans to do with the $270 million it raised earlier this year.
BioInform: Why did you decide to become a content provider?
Gessler: The Data Logic division was set up initially to provide database solutions for specific customers where we felt we could use our OPM (Object Protocol Model) database technology to create databases that would include each customer’s own data.
That’s still part of what we’re doing with GeneExpress, except it’s just kind of integrated into our overall total product offering. So, in fact, it’s not a specific service that we’re looking to sell independently. Now, we are trying to blend best-of-breed technologies both on the IT side and on the laboratory platform side.
We scaled up with the Affymetrix gene chip technology, and we took the IT infrastructure that we had and the tools that we have developed over the years to look at gene expression information and to manage gene expression information, and we started deploying those to build this large-scale reference database.
BioInform: Is selling content more profitable?
Gessler: Content businesses by their nature can be quite profitable and the reason is that, once you’ve covered your fixed costs for building the database, each successive time you sell the data the margins are incredibly large and you don’t have to provide additional services. So, the operating leverage for that kind of business is much more favorable than a service business.
BioInform: What are the technological advantages that you have?
Gessler: There are three areas of core competencies – on the front end, it’s the acquisition of the bio-repository to obtain samples from a worldwide network of clinical centers of excellence.
The second core competence is the large production scale that we’re running which uses Affymetrix gene chip technology as well as our own technology in the READS (differential display technology) area. This gives us broader coverage of some of the genes that aren’t presently on the chips.
We also focus and concentrate on what I call the heavy lifting. These are incredibly large data sets that we’re building. We are putting the data in a format that allows you to readily obtain the information that you want. That is to obtain both the clinical information that we have on all the samples as well as the genetic information that we have on the samples that have gone through production.
BioInform: What about proprietary technologies?
Gessler: The technology has been based on the work Victor Markowitz (Gene Logic’s chief information officer) has done on OPM and we have built on that. We’re using Corba Java interfaces and a variety of different technologies including the OPM proprietary piece to pool together the GeneExpress product for our client.
BioInform: What can researchers do with this information?
Gessler: The database allows researchers to make comparisons. Take, for instance, cancer: Researchers can look at the changes in gene expression as cancer moves from stage one to stage two to stage three.
BioInform: How much does a subscription cost?
Gessler: Pricing ranges between $3 million and $6 million for the GeneExpress product on a subscription basis. So, it’s an annual fee and the subscriptions are typically three years.
BioInform: How much of the genome have you completed?
Gessler: Now we’re approaching about 50 percent of the genome. We’re scanning about 50 percent of the genes with the 60,000 chip sets that we receive under a supply agreement from Affymetrix. So our data is on all the major disease areas. We’re covering cancer, cardiovascular disease, neuro-psychiatric disease, inflammation, and metabolic diseases, such as diabetes.
We’re doing a broad spectrum of diseased tissues and then measuring the gene expression in those tissues, again, for about 50 percent of the genome presently. And our goal and objective is really to move that up to as close to 100 percent of the genome as we possibly can. We’re shooting for right now about 30,000 human samples being run over the 100,000 genes to basically assemble what we believe will be the definitive database of gene expression.
BioInform: Who are your target customers?
Gessler: Big pharma and biotech clients as well as diagnostic companies, really anybody who is involved with product development at any stage.
They may be at the early research stage, where they may be looking at compounds and trying to prioritize compounds. That’s where the ToxExpress (predictive toxicology) module that we’ve developed is targeted, so that people can start to manage the attrition of these compounds. They’re looking to weed out the bad actors early on. And the earlier the better because then you are spending less money.
The next stage is trying to sort out which patients should be involved in clinical trials and that’s some of the work that we see coming our way.
And, really, the long-term vision of this company is taking this data all the way to the patient and to the physicians to treat them. And in that way trying to understand what is going on with a particular patient on a genetic level and comparing that against the vast resource of information that we have in this reference database which stands as a backdrop to contrast what might be best for that particular patient.
BioInform: Do you fear competitors such as Incyte?
Gessler: Incyte has built a marvelous business on the sequence side and we feel that we’ve cut our teeth and we’ve built the best gene expression database. And they have announced that they are going to try and compete in this area but we’re just going to have to see how that sorts itself out. We’ve got a substantial headstart.
BioInform: What differentiates Gene Logic from companies such as DoubleTwist and Compugen’s LabOnWeb?
Gessler: What’s fundamentally different here and what is difficult to replicate is the fact that we’re pulling together over the next couple of years here 30,000 human patient samples from all the major organs of the body under virtually all the major disease circumstances.
That information is not available out there nor is it likely to be available out there in the public domain. Unlike sequence information, this particular area of RNA expression offers a much more difficult task. It’s much more difficult to assemble all the clinical samples, all the clinical records and then to run them over a highly quality-controlled environment in order to obtain the data.
BioInform: What are your revenue projections?
Gessler: I guess the Street estimates are somewhere in the $30 million revenue range this year and we’re seeking to obviously meet those expectations. I believe that if we’re able to put that in place the hockey stick effect is going to start occurring. We hope to achieve revenue numbers consistent with the Street next year, which are in the $50 to $60 million dollar range.
BioInform: The stock has tumbled about 80 percent after hitting a 52-week intraday high of $152.50 in March. What do you make of the market’s reaction?
Gessler: It’s an incredibly volatile area. The amount of movement in stocks like Gene Logic and stocks that sit in this space are enormous over time. And I think that there’s going to be some settling out and hopefully a decrease in volatility over the year, but frankly at this point it’s hard to imagine that.
I think investors understand what we’re doing. We subscribe to the fact that it’s a monumental achievement, but the next thing is sorting out what’s going on with the genome. I think we’ve situated ourselves to be the best-positioned company to play in that space and I think there’s going to be recognition of that.
BioInform: How does your recently formed partnership with Life Technologies reflect your current and future strategy?
Gessler: This is around a major full length cloning and sequencing and patenting effort that we have put in place. There we’re taking for the first time the Gene Express database that we have built to sort out which are the most important genes and to go after full-length sequences and full-length clones. Then we sequence those and file patents on them.
And, we are making these clones available to our customers through a click-through option in the database called CloneExpress, where they can obtain full-length clones. And we’re making the IP available to our subscribers. So it’s really Gene Logic’s foray into a very aggressive patenting stra-tegy around all the new things we are going to hopefully unearth with the GeneExpress database.
BioInform: What about acquisitions?
Gessler: We’re opportunistic. We raised a quarter of a billion dollars in the first quarter and we didn’t assemble that cash to sit on it. We’re looking at organic growth and growth that can come through acquisitions of key technologies or key companies that might play a role in our space.