The following is a collection of posts from the Yahoo AFFX board. Most are by a contributor named Tonygt19, who identifies himself as a molecular biologist. I found them very insightful and well-written.
Enjoy - MML
tonygt19
Nov 29 1997
9:16 P.M PST
In my opinion Affymetrix's gene chips represent the biggest jump in genetic detection since PCR but without the laboratory skill which kept PCR from being the commercial success Roche and others were banking on. Does anyone out there think there is another technology poised to commercially supplant Affymetrix's future dominance is this potentially extremely profitable field?
I'm familiar with Hyseq, a company whose name I believe refers to Hybridization sequencing. Hyseqs's application of gene chip technology is more specialized and less commercially promising than Affymetrix's. With Hyseq you can determine sequence variations
between/among samples. This is info important to a researcher like myself but not necessarily important to the relatively lay medical/agricultural community which is the explosive potential market for Affx. Affx's high density chips promise to virtually detect any among thousands of different disease conditions for which it has been "programmed" by virtue of the probe sequences synthesized onto its grid. This will absolutely revolutionize diagnostics. This has not escaped the notice of others in the industry. Check out this company's website for its collaborators list.
It reads like a Who's Who in Pharmaceuticals.
tonygt19
Dec 3 1997
10:01 A.M PST
The detection of infectious disease from any type of sample from which DNA can be isolated is the main promise of a technology like AFFX's. However, the micro-miniaturization achieved by synthesizing DNA probes to a fixed medium via photolithography
currently allows for more than 40,000 different probe patches to be fit on a single chip! This means not only can you test for the presence of a disease organism in a sample but you could simultaneously test for what specific strain it might be, what antibiotic resistance it might have, geographical origin, etc., etc., all in a single test. That would probably use up a few hundred probe patches. With the left over 39,000+ probes you could perform a detailed test for the metabolic response of the infected host cells by probing gene activity levels. To me this is the equivalent to doing 5 years of experiments
in a few minutes. Believe me, this is a major breakthrough. But I wouldn't be as optimistic about this company's future as an investor if it weren't for the fact that they are making all the right moves with sealing business agreements early on with major players like Glaxo, Roche, Pfizer, Merck, etc. AFFX may well turn into the Microsoft of biotechnology. But bear in mind the current value of its stock is still based entirely on promise.
Don't expect short term gains based on earnings.
rgabel_cougar
Dec 3 1997
3:10 P.M PST
Affy's public position is that they intend to focus heavily on research applications and "high-value" diagnostic tests in the near future, as they do not have the manufacturing capacity to turn out millions of TB tests, for example, and that they are not sure what the market will be for these.
By "high-value" tests, they are referring to applications like a genetic screen of an AIDS patient to determine which available drug therapies will be most effective and to which drugs the patient has naturally immunities to. These tests, costing $500 or so, would potentially save AIDS patients tens of thousands of dollars spent on ineffective drugs and manage the disease treatment better than is currently available.
A private company in San Diego named Nanogen, backed by Kleiner Perkins, is developing a slightly different technology and focusing on disease diagnosis tests more so than Affy.
Don't fear, though, if you feel Affy is missing out on a big market opportunity. Their technology is completely adaptable to lower-cost, high-volume diagnostic tests if the market develops and will have manufacturing expertise driven by sales of GeneChips in the research market.
ÿ
tonygt19
Dec 3 1997
4:12 P.M PST
Agree with msg #8. The problem with this technology is that AFFX must chose the targets (i.e. what disease condition is being diagnosed) of its gene chips carefully. Designing good DNA probes/primers can be painstaking work for just a single set. Designing a massive array in which each probe gives an unambigous answer could be a monumental task. Therefore don't expect to see chips for lots of different conditions coming out quickly. Nevertheless, once a design is complete production should eventually be rapid and cheap. This technology is definitely in its infancy but its promise is enormous. Reference your earlier question about "few organisms" if you are referring to a disease condition being in very low titer I don't know what the level of detection is for these chips. That could be a problem unless an amplification is performed first.
ÿ
dthea
Dec 4 1997
6:53 A.M PST
A high sensitivity of the test that is comparable to available PCR techniques is crucial for this application. If amplification (i.e. PCR) were needed for a suitably configured GeneChip to pick up latent or very low titer infection then there would be no real advantage over simple PCR. If the Genechip WERE as sensitive as PCR then it may well supplant this technique. Regardless, as has been pointed out, the attraction of this technology for ID would be to enable a large assortment of potential pathogens to be detected (or ruled out) in one test. One could imagine a "fungal" chip or a "mycobacterial" chip which wouldbe embedded with gene sequences from all known pathogenic strains among these groups. With the recent understanding that ribosomal DNA (now used for molecular speciation) is highly conserved among members of individual species, I would imagine that it would be a ready source of sequences making the process less "painstaking". Am I missing something?
tonygt19
Dec 4 1997 ÿ
Gene chips will not replace PCR but rather should complement the technique. An hybridization-only technique will not have the sensitivity of an amplification of the target DNA/RNA as in PCR or of the signal as in Chiron's Branched DNA technology. But PCR only answers one question at a time and is labor intensive. AFFX technology uses a myriad of carefully selected oligos synthesized in known positions on a massive grid miniaturized onto a thumbnail sized chip. Each oligo patch can ask only one question: is my complement found in this sample, yes or no? AFFX apparently uses fluorescent tags to signal the answer. The power of the chip lies in how many different yes or no questions can be asked and how cleverly they are phrased. A carefully designed question matrix can give you information of complexity limited only by the size of the matrix. Actually coming up with the sequences would not be the hardest part. Designing and manufacturing each mask used in the photolithograhy process which adds each nucleotide would be a bear. Testing the chip and retooling to eliminate ambiguities is probably not a walk in the park either. Then software has to be written to analyze each particular chip.This is probably why AFFX currently has only three chips completed (HIV , p53, and p450 analytical assays). Each targets a high profile area of medical research. The design and production of each chip was probably tantamount to writing and producing a commercial software program. If your favorite organism is not the favorite organism of a lot of other medical researchers with big bucks behind it AFFX will probably not include it on a chip anytime soon. They are however making gene expression assay chips for humans, mice and yeast which
will get a lot of generic use. I'm probably getting way too technical for a financial chat area. If you have more questions of this type my email is tonygt19@concentric.net. Again this technology is in its infancy but everything about it will grow exponentially. In ten years it is likely everyone in an industrial nation will have been screened with a gene chip several times. For a long term investor that bodes well. ÿ
jgre1
Dec 4 1997
2:17 P.M PST
You're hitting the nail on the head when you say that the first commercially available chips target high profile areas of medical research! These particular targets demontrate that there is a built-in "sensitivity" possible in the form of the redundant overlap of the grid oligos. It also allows for input DNA that is in vastly different states of integrity to be monitored relatively equivalently - not a small deal when minimal patient samples may be involved. A great new technology - this company should remain the broad field leader with others picking up niches.
tonygt19
Dec 5 1997
8:23 A.M PST
Unfortunately I'm not privy to all the details of the synthetic DNA probes used in AFFX's chips, but there are two ways to increase shelf life of this kind of info. One is to maintain them in a film of concentrated ammonium hydroxide (the normal way of storing DNA primers) in which no microbes are viable and the other is to use a relatively new chemistry to make PNA which is the same as DNA with a peptide rather than a phosphate backbone. PNA is analogous to plastic, a human invention which is not biodegradable. I'm sure the AFFX people are clever enough to have already thought of this.
ÿ
will_on_the_hill
Dec 8 1997
3:25 P.M PST
I recently heard that there may be problems in the development. I refer to the possible non-specificity of the assay caused by DNA secondary structure (of the sample DNA). Has anyone else heard of this? If so, do you have any opinion on how far the gene chip technology may be compromised-if at all.
tonygt19
Dec 8 1997
8:18 P.M PST
Since the probes are designed to capture known sequences, avoiding sequence areas whose secondary structure would be problematic even under stringent hybridization conditions would be inherent in probe design just as it is in PCR. Check out the papers published on the technique listed on AFFX's website. They are getting very good results so far. Nevertheless, there may be regions of a genome that just aren't well suited to this kind of analysis. This will not negate the value of the technology.
ÿ
Maystack
Dec 9 1997
7:57 A.M PST
Since you seem to understand this technology, what do you think of Hyseq? Could you describe their process vs. AFFX? Thank YOU!
ÿ
tonygt19
Dec 10 1997
5:07 P.M PST
Maystack,
It appears Hyseq uses something akin to an inkjet technology to spray tiny spots of conventionally synthesized DNA probe solution onto a filter support, probably charged nylon. As you probably know inkjet technology has progressed to the point that even cheap printers can put 1440 discrete dots/inch. This array of dots is laid down in a known pattern so many different hybridizations can be done at once. Single nucleotide differences can be detected by using many similar probes.
Affx by contrast synthesizes its probes in place using an ingenius blend of photolithography and nucleotide synthesis chemistry. Affx will eventually be able to progress much further with miniaturization than Hyseq but its chips will be much harder to design. Hyseq's filter arrays would probably be modifiable by the user. Affx's genechip is a black box. Either they make a chip that is useful to you or they don't.
This is all of less interest to a potential investor than what market each company targets. Hyseq targets the relatively lucrative genomics research market. There are big bucks associated with the Human Genome Project and other species genome projects. Its process has its greatest promise in actually determining the sequence of DNA in genomic regions which for which that info is unknown. Affx on the other hand deals with DNA for which the sequence is already known. You can determine slight variations in sequence as in finding mutations but the company's greatest promise is in the detection of the PRESENCE of certain sequences such as disease organisms or gene expression profiles. It will see its greatest application (in the near future) in diagnostics and routine screens as in medical/physical exams.
As an investor you can dispense with trying to make sense of all the technical gobbledygook above. All you need to ask yourself is : How many people do I know who have ever spent money on DNA sequencing equipment? and :How many people do I know who have ever spent money on a medical diagnosis and on medical/physical exams? The answer should guide your investment.
TomInLA
Dec 14 1997
7:56 P.M PST
Recently read an article about a company...Nanogen that has its APEX technology which electronically addresses DNA. I understand the process is faster than the Gene Chip along with other advantages.. Can someone shed some light on this? Thanks.
tonygt19
Dec 17 1997
9:04 P.M PST
Nanogen has an interesting approach to microhybridization technology. It addresses the one weakness in AFFX's approach. Nanogen purports to have achieved electronic control over the hybridization conditions at each individual probe site on its chip via conventional chip circuitry. If this is true then they can control the speed of hybridization and optimize the quality of the signal by finetuning the hybidization for each probe. AFFX's chips (which,investors, are much farther along in production and are already being used by most major pharmaceutical firms) expose all the thousands of probes on the chip to the same hybrization conditions. This is not optimal but it still works because although the optimized conditions for each probe would be somewhat different all complementary short DNA sequences will hybridize to some degree under moderately stringent conditions. The variance in the strength of the individual probe signals can be modulated by the detection and analysis software so that a pleasantly readable and equalized pseudocolor image of the array is displayed on a computer screen.
One poster recommended AFFX buy out HYSEQ to eliminate its competition. Hyseq is not competition for AFFX and their technology would not enhance AFFX's but Nanogen could be and their technology would. But to paraphrase an old military dictum 'Affymetrix was there fustest with the mostest'. And that's how you win the battle and finally the war. Case in point: Microsoft's operating system is second rate, nevertheless....
Affymetrix's stock price is not tethered to earnings because they don't have any yet per se. What they have is cooperative agreements and licensing contracts that will make their technology the pervasive standard worldwide. I know this because I'm negotiating with them now to buy a system for my lab. Their response has been that they're not sure about pricing yet because they haven't actually sold anything without it being part of a collaborative contract. Expect AFFX's stock price to be buffeted about by the market's general impression on poorly understood high tech stocks. Those of us who understand this technology's potential consider it a real bargain.
tonygt19
Dec 31 1997
6:33 A.M PST
BRC_97,
From a technical standpoint I can't imagine AFFX showing any of its potential profitability for the next 2-3 years. Right now all their efforts appear to be targeted toward a global strategy of making their technology the defacto standard for drug efficacy testing. If they succeed, and they appear to be well on their way to doing so, then their chips will become a ubiquitous fixture in medicine and their profitability rise will be meteoric. This is definitely a long term proposition and, as always, a risk. There is always the chance of another breakthrough technology kicking over this anthill. However, this technology is rock solid and will be have a strong niche regardless. The reason I feel this company bears attention now is that its current price is entirely arbitrary, but when it starts to move it will move very, very quickly as its market will already be in place and will be monopolized. It will probably be hard to react in time to get a good price if you're standing flatfooted when that happens.
tonygt19
Jan 30 1998
10:58 P.M PST
Are you so uninformed on this company that you actually think they are selling gene chips at this time and making a profit? (And if you're looking for revenues generated from their licensing agreements, that is meaningless.) Who, other than their research collaborators, do you think has their $150K hybridization, detection and analysis equipment necessary to use gene chips? And
even if someone did, what meaning would the results have? Pharmaceutical companies are only now doing the research which correlates gene activity and specific DNA sequences with disease progression or remission. That is the purpose of all AFFX's announcements on cooperative agreements. They are trying to show that this necessary part of their strategy is being aggressively pursued.
Those of who think this technology will revolutionize diagnostics and make billions doing so are absolutely right. Those of you who think AFFX will be the direct beneficiary of all this are simply betting on the front runner very early in the race.
Why are you arguing about AFFX's current value? It is whatever the investing herd thinks it is. They don't read TN_Jed's impressive technical analyses nor do they subscribe to your nearly religious fervor for this issue. They just want to make money. Since
AFFX isn't going to be making money any time soon - its stock price will remain entirely a function of promise and hype. Since those are very fickle issues I would expect it to be very volatile ride for the next couple of years. And I doubt that any of the peaks or valleys will have anything to do with the true progress of the company. However, there is one significant announcement that is part of the company's strategy that will make AFFX a sure bet. That concerns their attempt to get their chip format declared the worldwide standard for gene chip analysis. If that happens and it probably will, hang on, because regardless of the price you paid so far you will do very, very well.
ÿ
tonygt19
Feb 13 1998
7:20 P.M PST
It would be difficult to predict Affx's financial future even if you assume it will dominate this technology because its impact on medicine will be unprecedented. Financial models may not encompass its potential explosive growth. I get the impression that many people think that this technology will merely make detection of pathogens easier. It goes way beyond that.
The current density of the chips and the eventual finetuning of the coding sequence probes for each gene in the human genome will allow a pharmaceutical company to compare an absolutely precise healthy gene expression profile with that of a disease state profile from a healthy and sick subjects respectively. They then can test the effects of prototype drugs by monitoring sick patients
at time and dosage points with the gene chips. They will look for the best protocol which returns the sick to the healthy state by monitoring the subject at the level of gene expression. This is far closer to what is actually going on in the body than observing classical symptomology although it will have to be correlated with that to satisfy the current crop of physicians. This is what is happening now with Affx's collaborations with pharmaceutical companies. The impact of this step to the financial future of Affx cannot be understated. It means no less than that the next generation of pharmaceuticals' effficacy will be DEFINED by Affx gene chip readouts. That means that the next generation of physicians will need a series gene chip readouts to determine if
the drug they are using is working according to the manufacturer's specs. HMO's will immediately see this as a precise and documented way to avoid costly errors in treatment. This is predicated on a conceptual premise that most people don't really grasp. Disease is not the presence of a pathogen in our body. Disease is largely our metabolic response to its presence. Gene chips will define that both qualitatively and quantitatively. Within 5 years or so I would expect medical schools to initiate some format of gene chip monitoring into every aspect of medical education. If that format is Affx's (and it probably will be) does that give you any insight into Affx's potential earnings?
TN_Jed
Feb 13 1998
9:04 P.M PST
There are two questions: 1) <Affx's potential earnings> and 2) proper discount rate.
ÿtonygt19
Feb 14 1998
8:48 A.M PST
Let me be direct. I don't know the answer to those questions and neither does anybody else. Your demand for quantification, while prudent, is simply one that cannot be satisfied with any reliable accuracy at the current state of this technology. I would recommend you forget about this issue and turn your attention to more mature and financially quantifiable technologies before you lose all your hair and begin mumbling to yourself as you walk down the street.
ÿ
tonygt19
Feb 18 1998
5:34 A.M PST
Before we start sounding like two thirds of the three stooges let me say that in principle I agree with most of what you have posted. I have evaluated this company from a technical and scientific outlook because my lab might be working with them in the near future. And since I'm a molecular geneticist involved with infectious disease detection I felt I was qualified to shed some light on how they do things and what if any potential promise their technology has in the diagnostic market. While I am not qualified to do financial analyses I am familiar with the concept of garbage in garbage out. Any quantitative assumptions one plugs into a financial prediction formula (regardless of how sophisticated the formula is) can only be wild guesses. Therefore, now
matter how much one might desire an accurate forecast the basic information needed to produce it simply doesn't exist yet. For example, you can't query physicians on their desire to use the product because most haven't yet heard of it and many of those who have don't understand it fully. So what you are left with is an evaluation of raw potential of the technology and the probable
dominance of this company. Also, I have usually included in my posts the caveat that AFFX may fail completely to continue in the lead.
TN_Jed
Feb 20 1998
10:20 P.M PST
The other day I had dinner with a guy who works for a Fortune 100 manufacturing company. They have expertise in coatings and thin film technology. They are also pumping tons of cash into manufacturing gene chips. They do not do the biochem. They license that part. They already have customers lined up for their chips, but nothing in production, yet. They hope to produce in 2-3 yrs
at a cost of <$15/chip and make chips by the thousands per day. I hear it takes AFFX a day to make a chip and they will sell for $hundreds.
Lots of companies are getting into this area. They smell a lot of cash.
FYI
tonygt19
Feb 21 1998
5:50 P.M PST
I tried to give you the benefit of the doubt in my last post. But your last post shows a fundamental lack of understanding
of this technology so deep that you would indeed be insane to invest money in it. What you have said is analogous to saying that someone has discovered a way to print words on hundreds of sheets of paper, bind them into a book and sell them for several dollars less than Stephan King does and therefore they will be serious competitors for his market of readers. Don't you understand that the value of gene chips is in proprietary MEANINGFUL CONTENT and not the technique of slapping DNA onto glass substrates? That's why the argument over patents is superfluous. There are dozens of ways to make microprobe arrays and they all work. I can
make them myself in my lab. But I can't apply them to clinical trials of proprietary drugs unless I have an ironclad contract with the pharmaceutical company making the drug. And I can't do that because AFFX has beat me and everyone else to the best contracts. Not only do they have nearly an insurmountable lead in contract agreements with a WHO's WHO list of the leading
pharmaceutical firms on the planet but they are pulling away with each passing day. All the other gene chip companies will have to license AFFX's format eventually or find esoteric niches for themselves in the research market (one far less lucrative than medicine).
The only weakness in AFFX's format is the inability to electronically address each probe patch. If they buy into Nanogen's technology that will allow them to jump quickly to the next generation of gene chips,i.e. semi-dynamic gene chips. If they were to do that nobody could catch them.
tonygt19
Feb 22 1998
8:16 P.M PST
TN_Jed,
Now you are broaching a subject for which there are some answers. Concerning your comment : "I don't think that in the long run the serious money will be made from specialized, proprietary chips for drug research. The big money will be made by mass producing simpler chips for routine blood or genetic analysis."
That's a good point and would be generally true for most technologies. But not here. The high density already available on AFFX chips would allow the entire expressed human genome to be represented as a human expression chip. That one chip would be used for all applications. One subset of genes might be used for breast cancer exams while another would be used for prostate or routine physicals. When you buy these chips from AFFX you tell them what you want to use them for and they would modify the software they sell you so that your equipment would only analyze the appropriate subset of genes. You would then pay them by data point analyzed/chip. Everybody buys the same mass produced chip but the price is a function of the number of data points analyzed. High chip densities will also allow probes for all major pathogens of the species for which the chip is designed to be included. So the chip will work for any application from the most complex to the simplest routine screening exam. This is analogous to a dictionary
which can be used by anyone regardless of the type of writing they are doing. Chip production now is expensive because the technology is not mature. They will be reduced in price dramatically as that happens. The proprietary part of the system is in the decoding and analysis not so much in the hardware.
It is not technologically unfeasible that AFFX could dominate this industry because both the medical and pharmaceutical industries desperately want there to be a single standard format for gene chip analysis. That would simplify their work enormously and save them lots of money. AFFX's biggest challenge will probably be in the form of anti-trust legislation.
ÿ
tonygt19
Feb 24 1998
10:08 A.M PST
It was clear to me you were simply mirroring the development integrated circuits in making your predictions and that is where you were mistaken. This is a very different industry. It is based on discrete data sets,i.e. the genomes of organisms of interest. These are being sequenced at an exponentially increasing rate. The probe density of chips already exceeds the number of
predicted genes in the human genome which is undoubtably the genome on which the greatest production of chips will be based. Some of what you say is will probably be true such as the great number of uses which will extend to agriculture, etc. But I really do believe your timeline is way off (possibly by an order of magnitude). Think in terms of 3 to 5 years for pervasive use,certainly no more than 10.
I only respond to technical issues because that's where my expertise lies so I don't feel qualified to dissect your financial calculations except to point out that there can be no historical data on an unprecedented industry.
As for your "prediction" on when I bought: You can't make predictions on the past unless you're caught in some kind of time loop. I bought in the low 20's with the full understanding that this stock would go nowhere for several years. It's in some IRA's for a couple of young nephews of mine.
I can discuss why marketing, and patents are far less important to this industry than in most but I'm at the end of my lunch break and Uncle Sam is calling. |
