Personalized cancer treatments have been a regular topic in New World Investor for a while now. The most familiar treatment process that we have discussed is probably Dendreon’s (DNDN) prostate cancer vaccine Provenge. But if you recall, this is not the only company revolutionizing the care that cancer patients receive. Back in October 2006, I introduced you to Affymetrix (AFFX) and gene assays.
While Dendreon uses cells from a cancer patient’s tumor to create a vaccine, Affymetrix’ biotech and pharmaceutical company customers create a treatment that is based on the genetic structure of the tumor. Basically, a genetic test, using DNA chips to scan messenger RNA (ribonucleic acid) from the thousands of other genes in a tumor, indicates how active a particular gene is within a cell. This allows pharmaceutical companies to develop very selective drugs, and allows doctors to determine which drugs will be most effective in treating a particular cancer based on the person’s genetic makeup.
DNA chips from Affymetrix and others were revolutionizing biological research even before the human genome was first sequenced in 2001. Then it turned out that sequencing those three billion letters created an ocean of complex data, and the only way to get a handle on it was using DNA chips. Having proven their value in research, these chips are now moving into the far larger markets for clinical diagnosis, and from there they will move into doctors’ offices — an even bigger market — to dramatically improve the practice of medicine.
We are going to be riding this MegaShift for a long time, and I want you to have a solid understanding of this technology and industry, especially since I am making a new gene chip recommendation today. So, let’s start with the process of creating DNA chips.
Glass, Not Silicon
DNA chips, called microarrays, are built using a semiconductor-like process, but on glass substrates instead of silicon. Semiconductor production uses photolithography to build layers of a circuit, with certain areas masked to prevent the light from hitting and hardening photosensitive material that creates circuit lines. In 1991, Dr. Stephen Fodor of the Affymax Research Institute published a paper in Science showing that photolithography could also manage light to activate biological compounds. Here’s how that process works:
From “The Economist,” November 30, 2006
As you can see in the diagram above, the surface of the glass-based array is covered with the basic material to build proteins or DNA. The material only sticks in areas that are exposed to light, and the unneeded material is then washed away. Then another layer of different material (pieces of RNA) is flooded onto the chip, with another photolithographic mask controlling which areas will stick to the layer beneath. At each point, a short vertical strand of protein or DNA that has a unique structure compared to its neighbors gets built up.
The reason this works is that DNA has a unique molecular structure, the familiar double helix or two chains that wind around each other and also connect at certain points, looking something like a ladder that has been twisted. Each chain is constructed from four bases called A (adenine), C (cytosine), G (guanine) and T (thymine). These can occur in any order, but the connections between the two chains are always this: A pairs with T and C pairs with G.
That’s why the single strands of DNA built on a chip can be used to probe the genetic structure of material that passes over the chip. When the gene chip is done, the material to be analyzed flows across its surface and the DNA in the material will bind to the complementary gene fragment on the chip. Points on the chip that react can be detected under fluorescent light, and the computer keeps track of the exact structure of the DNA or protein strand that was made at each point.
The first gene chips carried short molecules called peptides, but the plan from the beginning was to increase complexity and speed to the point that the chips could be used for DNA analysis. And now they’re used for the whole genome analysis.
Affymetrix Appears
Dr. Fodor started Affymetrix in 1993 as a spin-off from the lab, funded with a $3.5 million grant from the National Institutes of Health. After a series of patent fights and licensing deals in the 1990s, AFFX emerged as the leading company in making gene chips. Their first product was a chip to detect mutations in the HIV virus, but they had similar programs in DNA sequencing, gene detection and pathogen analysis, simply because they had no idea what would pay off first.
The sequencing of the human genome turned out to be the driver for an explosion in sales of gene expression arrays in the late 1990s. Affymetrix could not keep up with the demand, even though the equipment cost almost $200,000 for a single analysis station, and the chips cost several thousand dollars apiece. It is a razor-and-blades business, and Affymetrix managed to get the highest market share of workstations (razors) and arrays (gene chips, or blades). The total market is now up to about $700 million.
Some of Affymetrix’s competitors include: Illumina, Agilent and General Electric Healthcare. Its biggest competitor, Illumina, has focused on Single Nucleotide Polymorphisms (SNPs), which are the small genetic changes that account for the differences between individuals, even though we all have the human genome. However, Affymetrix sued Illumina for patent infringement and won, and I would not be surprised to see a 15% royalty rate agreed to in the near future. In the meantime, AFFX has introduced its newest generation of SNP chips and seems to have caught up to Illumina’s technology. SNP chips are useful both for understanding and predicting predispositions to disease and responses to different therapies. SNP chips accounted for only 9% of the gene chip market in 2004, when Affymetrix filed its lawsuit against Illumina, but are projected to grow to 45% of the market by 2009.
Just this week, Nature published a paper based on the work of 256 researchers trying to link specific diseases to specific genes. With the financial support of the Wellcome Trust, they focused on seven diseases: bipolar disorder, coronary-artery disease, Crohn’s disease, hypertension, rheumatoid arthritis, type I diabetes (juvenile) and type II diabetes (adult-onset). They studied 2,000 people with each disease and compared them with 3,000 symptom-free people, using gene chips to look for associations between a disease and one or more of 500,000 SNPs.
They found 24 places in the genome that were so strongly associated with a disease that there was only one chance in two million that the link is accidental. Another 58 associations had less than one chance in 100,000 of being wrong. They had especially strong results in Crohn’s disease with nine near-certainties and type I diabetes with seven near-certainties. Many of the candidates corresponded to already-known genes, but others didn’t, so the next phase of the project will be to search the areas of the genome that these new SNPs point to for likely genetic suspects. Once the gene is discovered, the protein it encodes will become a new target for pharmaceutical research — using gene chips, of course.
The Future of Gene Chips
Just as in the semiconductor industry, gene chip prices fall year by year. In the mid-1990s, Affymetrix had to throw away 90% of the chips that they made because they failed quality control. Today, they throw away about 3%. They’ve also gotten better at packing more and more DNA strands on a single chip, to the point that their human genome set is only two chips, and it will soon be one. As costs come down, the chips can move out of the research lab into diagnostics, and eventually into doctor’s offices.
At the various Money Shows, I used to tell the story of the New York lawyer who had an inoperable brain tumor in the mid-1990s. He did his research and then paid $100,000 out of his own pocket to have the tumor tissue analyzed on Affymetrix chips. It was a metastasized breast cancer tumor. A small fraction of men get breast cancer, and it typically metastisizes to the brain. While there is little to be done for brain cancer besides surgery, there is an effective drug for breast cancer: Herceptin. He went on Herceptin and then went into remission. His $100,000 analysis in the mid-1990s would cost less than $10,000 today, and it will cost less than $1,000 in a few years. That’s why the unit volume of Affymetrix arrays will continue to grow at very high rates.
The first microarray diagnostic test was approved in 2004 in both the U.S. and Europe. Roche and Affymetrix collaborated on the AmpliChip to identify over 30 genetic variations in two genes that affect how quickly a person metabolizes many prescription drugs. This is the first of many steps towards truly personalized medicine, where drugs are given to the 25% to 50% of patients who will benefit most from it, not wasted on the 45% to 70% who won’t respond, and carefully kept away from the 5% likely to have serious side effects.
Roche has several other AmpliChip diagnostics in development, all based on Affymetrix microarrays and many for the treatment of cancer. One of them can detect 20 different subtypes of leukemia, while another finds mutations in the crucial p53 gene that may determine a cancer patient’s prognosis.
Of course, one of the biggest applications of gene chips is in drug discovery. Merck, for example, does about 40,000 gene chip experiments a year, and about 20% of their drugs now in clinical trials were developed with microarrays. MIT and Harvard are collaborating on a data base of gene expression profiles for all approved drugs in the U.S. They expect to discover new applications for existing drugs and also to identify new methods of action.
After revolutionizing biological research, this MegaShift is just beginning to create genomics-based diagnostics and therapeutics, and it has a long way to run. I’m not willing to buy Illumina until we know what their royalty rate to Affymetrix will be, as it could take a substantial chunk of their gross profit. But I have found another company, like Affymetrix, that is using gene chips to create the diagnostic tests of the future.
Buy Sequenom
Sequenom (SQNM) has its own genetic analysis platform called MassARRAY, but in addition to selling general-purpose equipment to research labs and biotech companies, as Affymetrix and Illumina do, it develops specific tests to run on its machines and also does contract research. The MassARRAY system is a high-performance nucleic acid analysis platform for DNA analysis, including SNP genotyping and discovery, and gene expression analysis. The company collaborates with Quiagen in Germany to develop prenatal diagnostics, including a more accurate test of placental fluid for Down syndrome to replace amniocenteses. A second collaboration with Lenetix Medical Screening Labs has developed a non-invasive Rh negative test for mother/baby incompatibility, which will be launched in July.
Sequenom is also developing tests for cancer focused on circulating genetic material from tumors. Their other major internal program is run by a UK-based subsidiary that is developing and using genetic tests to select animals to breed for higher-quality offspring.
March first-quarter revenues grew 43% from the prior year to $9.9 million, so they seem to be getting traction in these markets. Both sales of MassARRAY systems and contract research services are growing. However, they are managing their R&D and marketing expenses higher, so the net loss of $3.7 million was about the same as last year. They had $38.7 million in cash after an April private placement of $18.4 million, and they have no debt.
Sequenom is expecting to do $37 million to $39 million this year, up 37% from 2006 to the high end of the range. However, that would mean no sequential growth from the March quarter, in spite of the Rh negative test that’s about to launch, so I suspect they are low balling by a little. They also said that they will report a $23 million to $25 million loss, with a cash burn of $16 million to $18 million. That would be higher than the $12.2 million they burned through in 2006, with almost the whole difference accounted for by increased marketing and R&D.
While I think they will do over $40 million this year, I think their cash burn is on target, so they will lose about 40 cents a share. They can cut that loss in half in 2008 and turn profitable in 2009. I want you buy SQNM before the Rh test launches in July. Buy a partial position under $4.50, and leave room to double up if it goes under $4. My target for this time next year is $8.
Content on Demand
Motorola (MOT), which I recommended in a Flash Alert on Monday, is starting to drift up with the semiconductor group. If you recall, Motorola has great exposure to the cell phone industry, which is growing this year by about 15%. Cell phones are a huge part of the Content on Demand MegaShift, and that’s why I recommended that you buy the January 2009 $17.50 LEAP calls (VMAAW) under $4 for a $10.50 target in 18 months. While MOT’s common stock has moved up this week, the LEAPs are still trading under my buy limit and you still have a chance to get on board.
Silicon Image (SIMG) should benefit in the second half of the year from the pre-holiday build of consumer electronics, because now hundreds of products include the DVI or HDMI interface in the form of chips bought from SIMG or intellectual property licensed from the company. About 80% of SIMG revenues come from product sales and 20% from licensing fees. In addition, it is now clear that PC manufacturers are not going to choose the lower cost DisplayPort standard to save money, because HDMI is required to connect to all the consumer devices, whether mobile or sitting in the living room. Intel essentially blessed the HDMI standard at the recent, huge Computex show in Taiwan when they introduced their G35 Express Chipset with HDMI, using a Silicon Image chip.
The consensus earnings estimates for this year have drifted up from 35 cents a share 60 days ago to 38 cents, and I think the company will do over 40 cents this year and over 60 cents next year. Put this together with the strong balance sheet — $224 million in cash ($2.55 a share) and no debt — and SIMG is a timely buy all the way up to $13 for my $20 target.
Zhone Technologies (ZHNE) has been a disappointment so far, but the world is going their way. As Internet-delivered video takes off, especially high-definition video delivered by telephone and cable companies, the demand for bandwidth is exploding. That includes the famous last mile, with the connection to the home or business being a potential bottleneck. Zhone specializes in solving that problem for both phone and cable companies and has a reputation for delivering rock-solid equipment that is easy to install and backed by knockout customer service. If the current management can’t deliver in that environment, we’ll have to sell the stock next year. But in anticipation of better times, I have no problem buying ZHNE under $2 for my $5 target — on their own or in an acquisition engineered by the active venture capitalists still on the board of directors.
New Energy Technology MegaShift
Energy prices have soared over the past couple of weeks. They started up after a powerful cyclone hit Oman, which shut down a major terminal that ships 650,000 barrels of oil a day. It continued today, pushing oil over $67 a barrel, when the Department of Energy announced both a drop in refinery utilization last week due to maintenance downtime and an increase in gasoline inventories for the summer driving season.
I have been on the road a bit more than usual lately, and it sure looks like a lot more RVs and general traffic than the last couple of years. It could be the weak dollar has convinced Americans to vacation at home this year, but high gas prices are not affecting traffic.
Just a few months ago, the official OPEC line was that they are “comfortable” with oil around $50 a barrel. Now they are saying they are “comfortable” with oil between $60 and $65. That’s because they don’t think there’s any low-cost oil left to be found, and they are not worried about the alternatives being available in sufficient volume to make a difference in the next 10 years. They are right on both counts. All of our New Energy Technology MegaShift recommendations could show incredible growth and make us rich, without putting a dent in oil prices for five to ten years.
In the “oil shock” in 1981, after long lines at the gas pumps, Exxon spent $40 billion to find new oil. Now, with oil prices about double those “oil shock” levels, Exxon spent about half as much on exploration in 2006 as they did in 1981, while paying much higher prices for labor, drilling rigs and everything else.
BP, the old British Petroleum, has been pumping oil from Alaska’s Prudhoe Bay for almost 30 years. Their original estimate for the life of the field was 25 years. To keep it going, they spend billions of dollars maintaining old equipment and pipelines. Recently, they discovered a leak that forced them to shut down a fourth of their production — 100,000 barrels a day — and the news caused oil prices to jump. But the entire field is at end of its life and in the oilfield equivalent of intensive care. It will be gone soon.
The same story is playing out at older oil fields around the world. Worldwide oil production is supposed to hit 87 million barrels per day by the end of 2007, but it is going to be tough. This could be the year of peak oil production. With the low level of new discoveries and the high cost of finding oil, it will take just a few more of the old monster fields going into depletion mode to start annual production shrinking.
What does this mean for us? With the hurricane season officially underway, even a return to normal from last summer’s extraordinarily quiet period would be an additional shock to the system. A higher-than-normal year, as is being forecast by the same folks who blew the forecast last year, would be a disaster for oil prices — but great news for our energy technology stocks.
Energy Conversion Devices (ENER) announced a showcase installation of a 1.1 megawatt UniSolar system for an almond and pistachio grower and processor, Paramount Farms. That’s enough power to run 300 homes, and Paramount said that the system will pay for itself in six years. It’s a wake-up call for other businesses that want to “go green” in a way that makes economic sense. ENER is a strong buy at current levels and up to my $35 limit for my $55 target.
Royal Dutch Shell (RDS.A) is well above my $75 target, and those who bought it are understandably anxious to lock in their profits. But RDS.A can hit $80 to $85 in a heartbeat if a big Gulf of Mexico hurricane takes out a chunk of refining capacity. So I want you to continue to hold RDS.A for a higher exit price, even though I am not going to raise my target price. I’ll let you know when the time is right to sell.
WiMAX MegaShift
MobilePro (MOBL) drew a number of questions after an Associated Press story about how cities across the country that installed their own Wi-Fi systems are getting less interest than forecast, resulting in millions of wasted taxpayer dollars.
Folks, this is good news for MOBL, not bad, and they certainly could use some. These systems are installed by the city themselves, rather than franchise a company like MOBL to do the job. Typically, they refuse to sell advertising and have no sensible revenue model. Then, under pressure from the naysayers, they never deploy the whole system and it limps along, much like — surprise! — municipal transport.
When MOBL won the contract for Sacramento, they were very excited. Then the politicians bowed to the “inclusive” argument and said that the system had to be free, with limits on advertising. MOBL walked away.
Subscriber Ron noted that the article said the technology is wrong, overpromised and underdelivered, and he asked. “Isn’t this technology too big of a gamble right now?”
Ron, a mesh Wi-Fi network like the one MOBL is running in Tempe, Arizona, is not difficult to install. This is not bleeding-edge stuff. And while city councils have a long history of over promising and under delivering, MOBL signs contracts for systems that will be economically viable and then delivers them. They have to. My hope is that articles like this take the municipally-owned, not-for-profit crowd out of the picture, and open up most of the remaining 700 systems that municipalities say they want to build to MOBL and its competitors.
MOBL remains a hold until they sort out their capital structure.
Print This Post
