No. of Recommendations: 0
Linda: Thanks for your excellent job in researching Affymetrix!

May I suggest an additional point to look for next time? You are a doctor, so you are used to looking at the development of parameters in your patients. Do the same with the quarterly figures of stocks! They are easily found in the SEC 10-Q and 10-K files, and I am confident you will find hard quantification instructive.

For all those not familiar with the biotech intricacies on this board, I also recommend reading Hedden´s excellent Affymetrix RM analysis:

and Hedden´s extremely useful categorization of biotech companies:

There needs to be mass acceptance of the technology. It is heavily dependent
on R&D spending of the pharmas and research institutions. Before acceptance
of its use in diagnostic testing becomes widespread, I believe there will have to
be major progress in the treatment of genetic diseases. For example, TH
mentions its possible use in genetic testing for Alzheimer's disease, Huntington's
disease and inherited tendency to breast cancer. Unless there are good
treatments for these diseases, the use may not be widely accepted by patients.

I think this is too narrow a description of potential uses for this technology.


Affymetrix has TECHNOLOGICAL advantages over the competition that make a host of applications possible.
The first advantage is the high density of probes on a chip. A silicon chip of a quarter square inch can harbour up to 320,000 different probes on tiny fields measuring 20 by 25 micrometers. This means you can quantify lots of genes at the same time. Spotting techniques are not able to array probes at that density. Miniaturization by photolithography, a technique adapted from chip makers, allows these high densities and should reduce production costs over time.

The second advantage is the use of short probes (so-called "oligonucleotides", single-stranded DNA pieces of maybe 25 bases) on the chip versus the long DNA probes (500+ bases) on the chips of competitors. With a long probe, you can only quantify a matching molecule in the sample assayed. With short probes, you can do that as well, but in addition you are able to determine whether there is a perfect match or, conversely, a single base pair mismatch-- in other words, a point mutation. In short, you can not only tell which genes are in your sample, but also whether they are o. k. or defective.

A third feature is true for chips from all producers: you can quantify the genes themselves (DNA) or a similar molecule that is only produced when the respective gene is active (messenger-RNA). By measuring RNA, you determine which of all these genes are switched on.

Apart from technological advantages, Affymetrix also has "intellectual property" advantages by being the first to closely cooperate with pharmaceutical companies and academic research institutions in the design of chips. Constructing gene chips is more than putting tiny pieces of DNA on an array-huge intellectual input is necessary to make the whole thing work in a way that you can get useful information out of it.

Affymetrix cites these fields of APPLICATIONS:

1. gene expression profiling
2. polymorphism analysis
3. disease management

(4. the technology would also work just to sequence unknown pieces of DNA, but this application is apparently covered by patents of Hyseq-the company whose president is a patent lawyer)

I will try to explain this mumbo-jumbo:

1. Gene expression profiling is mainly required by big pharma in high-throughput screening for new drug candidates. Right now, this is the main source of revenue for Affymetrix. Pharmaceutical companies frequently need to know whether drug candidate molecules turn certain genes on or off. The need for this technology is obvious from the fact that practically all big pharmaceutical companies have signed comprehensive agreements with Affymetrix ("EasyAccess-silver and -gold") that provide for access to the technology and delivery of large quantities of GeneChips at reduced prices. By signing up all of big pharma, Affymetrix is creating a new industry standard ("GATC") that will influence the gathering and handling of complex genetic data for years to come.

2. Polymorphism analysis. Only identical twins are completely alike. The rest of us, although having by and large the same genes, differ from one another by small differences in the DNA sequence of our genes. This can be useful in forensic applications (as in the search for criminals), but also in determining a predisposition to certain diseases. For example, small mutations in genes called BRCA1 and BRCA2 can increase the likelyhood to develop breast cancer, mutations in the APC gene to develop colon cancer. Many people would prefer to know such predispositions to have themselves monitored more closely. Other small genetic differences (polymorphisms) can result in different metabolization rates of drugs such as beta blockers or antidepressants. If you knew how the person will react, you could adjust dosage beforehand. New genetic risk factors are being discovered all the time. By assaying for them, in many cases people would be able to reduce these risks by adjusting their lifestyles. It is like measuring your cholesterol level and then adjusting your diet.

3. Disease management.

Example 1. A sore throat can be caused by a host of bacteria or viruses. Antibiotics help against some bacteria, but not against viruses. Currently, the doctor takes a swab and sends it to a lab. There, the bacteria are grown, the type of bacteria is determined and they are tested for possible resistance to a panel of antibiotics. If the causative agent were a virus, it would be even harder to identify. All that takes too much time and too much work. GeneChips will soon be able to provide a solution: what do we, bacteria and viruses have in common? We all contain characteristic DNA and/or RNA! With the help of a single chip, we will be able to find out which virus or bacterium caused the disease and to which antibiotics the bacteria are resistant (because resistance, too, is due to the presence of specific genes in the bacteria).
In the same vein, just a few days ago, Affymetrix signed an agreement with one of the world´s largest water utility companies, Lyonnaise des Eaux of France, to develop a GeneChip solution to test for microbial contaminations of drinking water. Click any Affymetrix news to view the announcement.

Example 2. Currently, if two people have the same malignant tumor, they are administered the same protocol of chemotherapy. Frequently, it works in one person, but in the other, it does not. Why?-- Carcinogenesis is thought of as a step-by-step acquisition of 5 to 10 unrelated mutations in different genes in the same cell. The exact pattern of genes and mutations involved varies from case to case. In other words, the "same" malignant tumors in two persons are genetically different, and therefore may respond differently to chemotherapy. With a needle biopsy of tumor tissue, potential tumor-causing mutations can be diagnosed via a GeneChip. Affymetrix is already selling a GeneChip designed to search for mutations in a gene frequently involved in malignant tumors, creatively termed p53. The status of this gene in a malignant tumor may soon influence the choice of therapeutic procedure. This is the first step in tailoring chemotherapy to the genetic pattern of a malignant tumor.


Taken together, there is huge and, by the results of new research, exponentially growing demand for high-efficiency genetic diagnostics. In this field, Affymetrix has such a lead that I have trouble imagining how anybody else could successfully compete.

Current impediments are high costs and the time and effort necessary to develop useful chips. From a cost standpoint, right now there are cheaper solutions than Affymetrix. But over time, I believe those will run into the wall of their technological limitations. Probe density combined with mutation detection capabilities will ultimately carry the day.

Affymetrix´technology is not perfect however, and there is a small possibility that someone else develops a technology that is so much more efficient that it enables the new competitor to make up for the years of development Affymetrix is currently ahead.

Readout is currently by fluorescence. For reasons I won´t go into here, a form of electronic readout would be more desirable. Nanogen and one or two other companies are apparently working on such solutions. A logical development, if technologically feasible, would be for Affymetrix and one of these companies to eventually pull together.

Incyte has a huge proprietary DNA sequence data base. In time, Affymetrix and Incyte will have to stop wasting their money on patent litigation. Affymetrix will need access to Incyte´s database, and Incyte will need Affymetrix´license money.


Affymetrix is not yet profitable and just starting to have sales. Therefore, in my view valuation is currently completely arbitrary.

Yet, a look at sales figures suggests that Affymetrix is in a transition from a concept stock to a "real" stock:

Quarterly revenue (product+contract+grant):
Q1/1997: $2.8 million, Q2/1997: $4.7 million, Q3/1997: $5.0 million, Q4/1997: $7.2 million, Q1/1998: $9.8 million, Q2/1998: $11.1 million, Q3/1998: $15.0 million, Q4/1998: 16.2 million. Grant revenue was fairly constant around $2 million over this period, product sales and contract revenues were rising nicely.

On the other hand, cost of products sold went up almost as fast as product sales and still was at 68% in Q4/1998. Affymetrix clearly has to work on its manufacturing efficiency. A second manufacturing facility is being built near Sacramento, CA and should start production versus the end of 1999. In addition to alleviating current backlogs, this might also bring a chance to eventually improve gross margins.

Selling, General and Administrative also went up steeply, not least because of intensifying patent litigation.

Due to these negatives, Loss attributable to common shareholders remained more or less unchanged:
Q1/1997: $4.9 million, Q2/1997: $5.2 million, Q3/1997: $5.3 million, Q4/1997: $7.1 million, Q1/1998: $5.9 million, Q2/1998: $6.6 million, Q3/1998: $5.7 million, Q4/1998: 7.2 million. In summary, profitability is not just around the corner. It remains to be seen whether it can be reached in 2000.

Cash reserves are $80 million, quarterly burn rate is a little over $10 million. They better become profitable soon, otherwise they will again have to sell preferred stock to uncle Glaxo!



Print the post  


What was Your Dumbest Investment?
Share it with us -- and learn from others' stories of flubs.
When Life Gives You Lemons
We all have had hardships and made poor decisions. The important thing is how we respond and grow. Read the story of a Fool who started from nothing, and looks to gain everything.
Community Home
Speak Your Mind, Start Your Blog, Rate Your Stocks

Community Team Fools - who are those TMF's?
Contact Us
Contact Customer Service and other Fool departments here.
Work for Fools?
Winner of the Washingtonian great places to work, and Glassdoor #1 Company to Work For 2015! Have access to all of TMF's online and email products for FREE, and be paid for your contributions to TMF! Click the link and start your Fool career.