A memorial to Mark Buller, PhD, and our response to the propaganda film "Demon in the Freezer".

Earlier this year my friend and colleague Mark Buller passed away. Mark was a noted virologist and a professor at Saint Louis University. He was struck by a car while riding his bicycle home from the lab, and died from his injuries. Here is Mark's obituary as published by the university.

In 2014 and 2015, Mark and I served as advisors to a WHO scientific working group on synthetic biology and the variola virus (the causative agent of smallpox). In 2016, we wrote the following, previously un-published, response to an "Op-Doc" that appeared in the New York Times. In a forthcoming post I will have more to say about both my experience with the WHO and my thoughts on the recent publication of a synthetic horsepox genome. For now, here is the last version (circa May, 2016) of the response Mark I and wrote to the Op-Doc, published here as my own memorial to Professor Buller.

---

Variola virus is still needed for the development of smallpox medical countermeasures

On May 17, 2016 Errol Morris presented a short movie entitled “Demon in the Freezer” [note: quite different from the book of the same name by Richard Preston] in the Op-Docs section of the on-line New York Times. The piece purported to present both sides of the long-standing argument over what to do with the remaining laboratory stocks of variola virus, the causative agent of smallpox, which no longer circulates in the human population.

Since 1999, the World Health Organization has on numerous occasions postponed the final destruction of the two variola virus research stocks in Russia and the US in order to support public health related research, including the development of smallpox molecular diagnostics, antivirals, and vaccines.  

“Demon in the Freezer” clearly advocates for destroying the virus. The Op-Doc impugns the motivation of scientists carrying out smallpox research by asking: “If given a free hand, what might they unleash?” The narrative even suggests that some in the US government would like to pursue a nefarious policy goal of “mutually assured destruction with germs”. This portion of the movie is interlaced with irrelevant, hyperbolic images of mushroom clouds. The reality is that in 1969 the US unilaterally renounced the production, storage or use biological weapons for any reason whatsoever, including in response to a biologic attack from another country. The same cannot be said for ISIS and Al-Qaeda. In 1975 the US ratified the 1925 Geneva Protocol banning chemical and biological agents in warfare and became party to the Biological Weapons Convention that emphatically prohibits the use of biological weapons in warfare.

“Demon in the Freezer” is constructed with undeniable flair, but in the end it is a benighted 21st century video incarnation of a middling 1930's political propaganda mural. It was painted with only black and white pigments, rather than a meaningful palette of colors, and using a brush so broad that it blurred any useful detail. Ultimately, and to its discredit, the piece sought to create fear and outrage based on unsubstantiated accusations.

Maintaining live smallpox virus is necessary for ongoing development and improvement of medical countermeasures. The first-generation US smallpox vaccine was produced in domesticated animals, while the second-generation smallpox vaccine was manufactured in sterile bioreactors; both have the potential to cause serious side effects in 10-20% of the population. The third generation smallpox vaccine has an improved safety profile, and causes minimal side effects. Fourth generation vaccine candidates, based on newer, lower cost, technology, will be even safer and some are in preclinical testing. There remains a need to develop rapid field diagnostics and an additional antiviral therapy for smallpox.

Continued vigilance is necessary because it is widely assumed that numerous undeclared stocks of variola virus exist around the world in clandestine laboratories. Moreover, unsecured variola virus stocks are encountered occasionally in strain collections left behind by long-retired researchers, as demonstrated in 2014 with the discovery of 1950s vintage variola virus in a cold room at the NIH. The certain existence of unofficial stocks makes destroying the official stocks an exercise in declaring “victory” merely for political purposes rather than a substantive step towards increasing security. Unfortunately, the threat does not end with undeclared or forgotten samples.

In 2015 a WHO Scientific Working Group on Synthetic Biology and Variola Virus and Smallpox determined that a “skilled laboratory technician or undergraduate student with experience of working with viruses” would be able to generate variola virus from the widely available genomic sequence in “as little as three months”. Importantly, this Working Group concluded that “there will always be the potential to recreate variola virus and therefore the risk of smallpox happening again can never be eradicated.” Thus, the goal of a variola virus-free future, however laudable, is unattainable. This is sobering guidance on a topic that requires sober consideration.

We welcome increased discussions of the risk of infectious disease and of public health preparedness. In the US these topics have too long languished among second (or third) tier national security conversations. The 2014 West Africa Ebola outbreak and the current Congressional debate over funding to counter the Zika virus exemplifies the business-as-usual political approach of throwing half a bucket of water on the nearest burning bush while the surrounding countryside goes up in flames. Lethal infectious diseases are serious public health and global security issues and they deserve serious attention.

The variola virus has killed more humans numerically than any other single cause in history. This pathogen was produced by nature, and it would be the height of arrogance, and very foolish indeed, to assume nothing like it will ever again emerge from the bush to threaten human life and human civilization. Maintenance of variola virus stocks is needed for continued improvement of molecular diagnostics, antivirals, and vaccines. Under no circumstances should we unilaterally cripple those efforts in the face of the most deadly infectious disease ever to plague humans. This is an easy mistake to avoid.

Mark Buller, PhD, was a Professor of Molecular Microbiology & Immunology at Saint Louis University School of Medicine, who passed away on February 24, 2017. Rob Carlson, PhD, is a Principal at the engineering and strategy firm Biodesic and a Managing Director of Bioeconomy Capital.

The authors served as scientific and technical advisors to the 2015 WHO Scientific Working Group on Synthetic Biology and Variola Virus.

Guesstimating the Size of the Global Array Synthesis Market

(Updated, Aug 31, for clarity.)

After chats with a variety of interested parties over the last couple of months, I decided it would be useful to try to sort out how much DNA is synthesized annually on arrays, in part to get a better handle on what sort of capacity it represents for DNA data storage. The publicly available numbers, as usual, are terrible, which is why the title of the post contains the word "guesstimating". Here goes.

First, why is this important? As the DNA synthesis industry grows, and the number of applications expands, new markets are emerging that use that DNA in different ways. Not all that DNA is produced using the same method, and the different methods are characterized by different costs, error rates, lengths, throughput, etc. (The Wikipedia entry on Oligonucleotide Synthesis is actually fairly reasonable, if you want to read more. See also Kosuri and Church, "Large-scale de novo DNA synthesis: technologies and applications".) If we are going to understand the state of the technology, and the economy built on that technology, then we need to be careful about measuring what the technology can do and how much it costs. Once we pin down what the world looks like today, we can start trying to make sensible projections, or even predictions, about the future.

While there is just one basic chemistry used to synthesize oligonucleotides, there are two physical formats that give you two very different products. Oligos synthesized on individual columns, which might be packed into 384 (or more) well plates, can be manipulated as individual sequences. You can use those individual sequences for any number of purposes, and if you want just one sequence at a time (for PCR or hybridization probes, gene therapy, etc), this is probably how you make it. You can build genes from column oligos by combining them pairwise, or in larger numbers, until you get the size construct you want (typically of order a thousand bases, or a kilobase [kB], at which point you start manipulating the kB fragments). I am not going to dwell on gene assembly and error correction strategies here; you can Google that.

The other physical format is array synthesis, in which synthesis takes place on a solid surface consisting of up to a million different addressable features, where light or charge is used to control which sequence is grown on which feature. Typically, all the oligos are removed from the array at once, which results in a mixed pool. You might insert this pool into a longer backbone sequence to construct a library of different genes that code for slightly different protein sequences, in order to screen those proteins for the characteristics you want. Or, if you are ambitious, you might use the entire pool of array oligos to directly assemble larger constructs such as genes. Again, see Google, Codon Devices, Gen9, Twist, etc. More relevant to my purpose here, a pool of array-synthesized oligos can be used as an extremely dense information storage medium. To get a sense of when that might be a viable commercial product, we need to have an idea of the throughput of the industry, and how far away from practical implementation we might be. 

Next, to recap, last year I made a stab at estimating the size of the gene synthesis market. Much of the industry revenue data came from a Frost & Sullivan report, commissioned by Genscript for its IPO prospectus. The report put the 2014 market for synthetic genes at only $137 million, from which I concluded that the total number of bases shipped as genes that year was 4.8 billion, or a bit less than a duplex human genome. Based on my conversations with people in the industry, I conclude that most of those genes were assembled from oligos synthesized on columns, with a modest, but growing, fraction from array oligos. (See "On DNA and Transistors", and preceding posts, for commentary on the gene synthesis industry and its future.)

The Frost & Sullivan report also claims that the 2014 market for single-stranded oligonucleotides was $241 million. The Genscript IPO prospectus does not specify whether this $241 million was from both array- and column-synthesized oligos, or not. But because Genscript only makes and uses column synthesis, I suspect it referred only to that synthesis format.  At ~$0.01 per base (give or take), this gives you about 24 billion bases synthesized on columns sold in 2014. You might wind up paying as much as $0.05 to $0.10 per base, depending on your specifications, which if prevalent would pull down the total global production volume. But I will stick with $0.01 per base for now. If you add the total number of bases sold as genes and the bases sold as oligos, you get to just shy of 30 billion bases (leaving aside for the moment the fact that an unknown fraction of the genes came from oligos synthesized on arrays).

So, now, what about array synthesis? If you search the interwebs for information on the market for array synthesis, you get a mess of consulting and marketing research reports that cost between a few hundred and many thousands of dollars. I find this to be an unhelpful corpus of data and analysis, even when I have the report in hand, because most of the reports are terrible at describing sources and methods. However, as there is no other source of data, I will use a rough average of the market sizes from the abstracts of those reports to get started. Many of the reports claim that in 2016 the global market for oligo synthesis was ~$1.3 billion, and that this market will grow to $2.X billion by 2020 or so. Of the $1.3B 2016 revenues, the abstracts assert that approximately half was split evenly between "equipment and reagents". I will note here that this should already make the reader skeptical of the analyses, because who is selling ~$260M worth of synthesis "equipment"? And who is buying it? Seems fishy. But I can see ~$260M in reagents, in the form of various columns, reagents, and purification kit. This trade, after all, is what keeps outfits like Glenn Research and Trilink in business.

Forging ahead through swampy, uncertain data, that leaves us with ~$650M in raw oligos. Should we say this is inclusive or exclusive of the $241M figure from Frost & Sullivan? I am going to split the difference and call it $500M, since we are already well into hand waving territory by now, anyway. How many bases does this $500M buy?

Array oligos are a lot cheaper than column oligos. Kosuri and Church write that "oligos produced from microarrays are 2–4 orders of magnitude cheaper than column-based oligos, with costs ranging from $0.00001–0.001 per nucleotide, depending on length, scale and platform." Here we stumble a bit, because cost is not the same thing as price. As a consumer, or as someone interested in understanding how actually acquiring a product affects project development, I care about price. Without knowing a lot more about how this cost range is related to price, and the distribution of prices paid to acquire array oligos, it is hard to know what to do with the "cost" range. The simple average cost would be $0.001 per base, but I also happen to know that you can get oligos en masse for less than that. But I do not know what the true average price is. For the sake of expediency, I will call it $0.0001 per base for this exercise.

Combining the revenue estimate and the price gives us about 5E12 bases per year. From there, assuming roughly 100-mer oligos, you get to 5E10 difference sequences. And adding in the number of features per array (between 100,000 and 1M), you get as many as 500,000 arrays run per year, or about 1370 per day. (It is not obvious that you should think of this as 1370 instruments running globally, and after seeing the Agilent oligo synthesis operation a few years ago, I suggest that you not do that.) If the true average price is closer to $0.00001 per base, then you can bump up the preceding numbers by an order of magnitude. But, to be conservative, I won't do that here. Also note that the ~30 billion bases synthesized on columns annually are not even a rounding error on the 5E12 synthesized on arrays.

Aside: None of these calculations delve into the mass (or the number of copies) per synthesized sequence. In principle, of course, you only need one perfect copy of each sequence, whether synthesized on columns or arrays, to use DNA in any just about application (except where you need to drive the equilibrium or reaction kinetics). Column synthesis gives you many more copies (i.e., more mass per sequence) than array synthesis. In principle — ignoring the efficiency of the chemical reactions — you could dial down the feature size on arrays until you were synthesizing just one copy per sequence. But then it would become exceedingly important to keep track of that one copy through successive fluidic operations, which sounds like a quite difficult prospect. So whatever the final form factor, an instrument needs to produce sufficient copies per sequence to be useful, but not so many that resources are wasted on unnecessary redundancy/degeneracy.

Just for shits and giggles, and because array synthesis could be important for assembling the hypothetical synthetic human genome, this all works out to be enough DNA to assemble 833 human duplex genomes per year, or 3 per day, in the absence of any other competing uses, of which there are obviously many. Also if you don't screw up and waste some of the DNA, which is inevitable. Finally, at a density of ~1 bit/base, this is enough to annually store 5 TB of data, or the equivalent of one very beefy laptop hard drive.

And so, if you have access to the entire global supply of single stranded oligonucleotides, and you have an encoding/decoding and sequencing strategy that can handle significant variations in length and high error rates at scale, you can store enough HD movies and TV to capture most of the new, good stuff that HollyBollyWood churns out every year. Unless, of course, you also need to accommodate the tastes and habits of a tween daughter, in which case your storage budget is blown for now and evermore no matter how much capacity you have at hand. Not to mention your wallet. Hey, put down the screen and practice the clarinet already. Or clean up your room! Or go to the dojo! Yeesh! Kids these days! So many exclamations!

Where was I?

Now that we have some rough numbers in hand, we can try to say something about the future. Based on my experience working on the Microsoft/UW DNA data storage project, I have become convinced that this technology is coming, and it will be based on massive increases in the supply of synthetic DNA. To compete with an existing tape drive (see the last few 'graphs of this post), able to read and write ~2 Gbits a second, a putative DNA drive would need to be able to read and write ~2 GBases per second, or ~183 Pbits/day, or the equivalent of ~10,000 human genomes a day — per instrument/device. Based on the guesstimate above, which gave a global throughput of just 3 human genomes per day, we are waaaay below that goal.

To be sure, there is probably some demand for a DNA storage technology that can work at lower throughputs: long term cold storage, government archives, film archives, etc. I suspect, however, that the many advantages of DNA data storage will attract an increasing share of the broader archival market once the basic technology is demonstrated on the market. I also suspect that developing the necessary instrumentation will require moving away from the existing chemistry to something new and different, perhaps enzymatically controlled synthesis, perhaps even with the aid of the still hypothetical DNA "synthase", which I first wrote about 17 years ago.

In any event, based on the limited numbers available today, it seems likely that the current oligo array industry has a long way to go before it can supply meaningful amounts of DNA for storage. It will be interesting to see how this all evolves.

Planning for Toy Story and Synthetic Biology: It's All About Competition (Updated)

Here are updated cost and productivity curves for DNA sequencing and synthesis.  Reading and writing DNA is becoming ever cheaper and easier.  The Economist and others call these "Carlson Curves", a name I am ambivalent about but have come to accept if only for the good advertising.  I've been meaning to post updates for a few weeks; the appearance today of an opinion piece at Wired about Moore's Law serves as a catalyst to launch them into the world.  In particular, two points need some attention, the  notions that Moore's Law 1) is unplanned and unpredictable, and 2) somehow represents the maximum pace of technological innovation.

DNA Sequencing Productivity is Skyrocketing

First up: the productivity curve.  Readers new to these metrics might want to have a look at my first paper on the subject, "The Pace and Proliferation of Biological Technologies" (PDF) from 2003, which describes why I chose to compare the productivity enabled by commercially available sequencing and synthesis instruments to Moore's Law.  (Briefly, Moore's Law is a proxy for productivity; more transistors putatively means more stuff gets done.)  You have to choose some sort of metric when making comparisons across such widely different technologies, and, however much I hunt around for something better, productivity always emerges at the top.

It's been a few years since I updated this chart.  The primary reason for the delay is that, with the profusion of different sequencing platforms, it became somewhat difficult to compare productivity [bases/person/day] across platforms.  Fortunately, a number of papers have come out recently that either directly make that calculation or provide enough information for me to make an estimate.  (I will publish a full bibliography in a paper later this year.  For now, this blog post serves as the primary citation for the figure below.)

Visual inspection reveals a number of interesting things.  First, the DNA synthesis productivity line stops in about 2008 because there have been no new instruments released publicly since then.  New synthesis and assembly technologies are under development by at least two firms, which have announced they will run centralized foundries and not sell instruments.  More on this later.

Second, it is clear that DNA sequencing platforms are improving very rapidly, now much faster than Moore's Law.  This is interesting in itself, but I point it out here because of the post today at Wired by Pixar co-founder Alvy Ray Smith, "How Pixar Used Moore's Law to Predict the Future".  Smith suggests that "Moore's Law reflects the top rate at which humans can innovate. If we could proceed faster, we would," and that "Hardly anyone can see across even the next crank of the Moore's Law clock."

Moore's Law is a Business Model and is All About Planning -- Theirs and Yours

As I have written previously, early on at Intel it was recognized that Moore's Law is a business model (see the Pace and Proliferation paper, my book, and in a previous post, "The Origin of Moore's Law").  Moore's Law was always about economics and planning in a multi-billion dollar industry.  When I started writing about all this in 2000, a new chip fab cost about $1 billion.  Now, according to The Economist, Intel estimates a new chip fab costs about $10 billion.  (There is probably another Law to be named here, something about exponential increases in cost of semiconductor processing as an inverse function of feature size.  Update: This turns out to be Rock's Law.)  Nobody spends $10 billion without a great deal of planning, and in particular nobody borrows that much from banks or other financial institutions without demonstrating a long-term plan to pay off the loan.   Moreover, Intel has had to coordinate the manufacturing and delivery of very expensive, very complex semiconductor processing instruments made by other companies.  Thus Intel's planning cycle explicitly extends many years into the future; the company sees not just the next crank of the Moore's Law clock, but several cranks.  New technology has certainly been required to achieve these planning goals, but that is just part of the research, development, and design process for Intel.  What is clear from comments by Carver Mead and others is that even if the path was unclear at times, the industry was confident that they could to get to the next crank of the clock.

Moore's Law served a second purpose for Intel, and one that is less well recognized but arguably more important; Moore's Law was a pace selected to enable Intel to win.  That is why Andy Grove ran around Intel pushing for financial scale (see "The Origin of Moore's Law").  I have more historical work to do here, but it is pretty clear that Intel successfully organized an entire industry to move at a pace only it could survive.  And only Intel did survive.  Yes, there are competitors in specialty chips and in memory or GPUs, but as far as high volume, general CPUs go, Intel is the last man standing.  Finally, and alas I don't have a source anywhere for this other than hearsay, Intel could have in fact gone faster than Moore's Law.  Here is the hearsay: Gordon Moore told Danny Hillis who told me that Intel could have gone faster.  (If anybody has a better source for that particular point, give me a yell on Twitter.)  The inescapable conclusion from all this is that the management of Intel made a very careful calculation.  They evaluated product roll-outs to consumers, the rate of new product adoption, the rate of semiconductor processing improvements, and the financial requirements for building the next chip fab line, and then set a pace that nobody else could match but that left Intel plenty of headroom for future products.  It was all about planning.

The reason I bother to point all this out is that Pixar was able to use Moore's Law to "predict the future" precisely because Intel meticulously planned that future.  (Calling Alan Kay: "The best way to predict the future is to invent it.")  Which brings us back to biology.  Whereas Moore's Law is all about Intel and photolithography, the reason that productivity in DNA sequencing is going through the roof is competition among not just companies but among technologies.  And we only just getting started.  As Smith writes in his Wired piece, Moore's Law tells you that "Everything good about computers gets an order of magnitude better every five years."  Which is great: it enabled other industries and companies to plan in the same way Pixar did.  But Moore's Law doesn't tell you anything about any other technology, because Moore's Law was about building a monopoly atop an extremely narrow technology base.  In contrast, there are many different DNA sequencing technologies emerging because many different entrepreneurs and companies are inventing the future.

The first consequence of all this competition and invention is that it makes my job of predicting the future very difficult.  This emphasizes the difference between Moore's Law and Carlson Curves (it still feels so weird to write my own name like that): whereas Intel and the semiconductor industry were meeting planning goals, I am simply keeping track of data.  There is no real industry-wide planning in DNA synthesis or sequencing, other than a race to get to the "$1000 genome" before the next guy.  (Yes, there is a vague road-mappy thing promoted by the NIH that accompanied some of its grant programs, but there is little if any coordination because there is intense competition.)

Biological Technologies are Hard to Predict in Part Because They Are Cheaper than Chips

Compared to other industries, the barrier to entry in biological technologies is pretty low.  Unlike chip fabs, there is nothing in biology that costs $10 billion commercially, nor even $1 billion.  (I have come to mostly disbelieve pharma industry claims that developing drugs is actually that expensive, but that is another story for another time.)  The Boeing 787 reportedly cost $32 billion to develop as of 2011, and that is on top of a century of multi-billion dollar aviation projects that had to come before the 787.

There are two kinds of costs that are important to distinguish here.  The first is the cost of developing and commercializing a particular product.  Based on the money reportedly raised and spent by Life, Illumina, Ion Torrent (before acquisition), Pacific Biosciences, Complete Genomics (before acquisition), and others, it looks like developing and marketing second-generation sequencing technology can cost upwards of about $100 million.  Even more money gets spent, and lost, in operations before anybody is in the black.  My intuition says that the development costs are probably falling as sequencing starts to rely more on other technology bases, for example semiconductor processing and sensor technology, but I don't know of any real data.  I would also guess that nanopore sequencing, should it actually become a commercial product this year, will have cost less to develop than other technologies, but, again, that is my intuition based on my time in clean rooms and at the wet bench.  I don't think there is great information yet here, so I will suspend discussion for the time being.

The second kind of cost to keep in mind is the use of new technologies to get something done.  Which brings in the cost curve.  Again, the forthcoming paper will contain appropriate references.

The cost per base of DNA sequencing has clearly plummeted lately.  I don't think there is much to be made of the apparent slow-down in the last couple of years.  The NIH version of this plot has more fine grained data, and it also directly compares the cost of sequencing with the cost per megabyte for memory, another form of Moore's Law.  Both my productivity plot above and the NIH plot show that sequencing has at times improved much faster than Moore's Law, and generally no slower.

If you ponder the various wiggles, it may be true that the fall in sequencing cost is returning to a slower pace after a period in which new technologies dramatically changed the market.  Time will tell.  (The wiggles certainly make prediction difficult.)  One feature of the rapid fall in sequencing costs is that it makes the slow-down in synthesis look smaller; see this earlier post for different scale plots and a discussion of the evaporating maximum profit margin for long, double-stranded synthetic DNA (the difference between the orange and yellow lines above).

Whereas competition among companies and technologies is driving down sequencing costs, the lack of competition among synthesis companies has contributed to a stagnation in price decreases.  I've covered this in previous posts (and in this Nature Biotech article), but it boils down to the fact that synthetic DNA has become a commodity produced using relatively old technology.

Where Are We Headed?

Now, after concluding that the structure of the industry makes it hard to prognosticate, I must of course prognosticate.  In DNA sequencing, all hell is breaking loose, and that is great for the user.  Whether instrument developers thrive is another matter entirely.  As usual with start-ups and disruptive technologies, surviving first contact with the market is all about execution.  I'll have an additional post soon on how DNA sequencing performance has changed over the years, and what the launch of nanopore sequencing might mean.

DNA synthesis may also see some change soon.  The industry as it exists today is based on chemistry that is several decades old.  The common implementation of that chemistry has heretofore set a floor on the cost of short and long synthetic DNA, and in particular the cost of synthetic genes.  However, at least two companies are claiming to have technology that facilitates busting through that cost floor by enabling the use of smaller amounts of poorer quality, and thus less expensive, synthetic DNA to build synthetic genes and chromosomes.

Gen9 is already on the market with synthetic genes selling for something like $.07 per base.  I am not aware of published cost estimates for production, other than the CEO claiming it will soon drop by orders of magnitude.  Cambrian Genomics has a related technology and its CEO suggests costs will immediately fall by 5 orders of magnitude.  Of course, neither company is likely to drop prices so far at the beginning, but rather will set prices to undercut existing companies and grab market share.  Assuming Gen9 and Cambrian don't collude on pricing, and assuming the technologies work as they expect, the existence of competition should lead to substantially lower prices on genes and chromosomes within the year.  We will have to see how things actually work in the market.  Finally, Synthetic Genomics has announced it will collaborate with IDT to sell synthetic genes, but as far as I am aware nothing new is actually shipping yet, nor have they announced pricing.

So, supposedly we are soon going to have lots more, lots cheaper DNA.  But you have to ask yourself who is going to use all this DNA, and for what.  The important business point here is that both Gen9 and Cambrian Genomics are working on the hypothesis that demand will increase markedly (by orders of magnitude) as the price falls.  Yet nobody can design a synthetic genetic circuit with more than a handful of components at the moment, which is something of a bottleneck on demand.  Another option is that customers will do less up-front predictive design and instead do more screening of variants.  This is how Amyris works -- despite their other difficulties, Amyris does have a truly impressive metabolic screening operation -- and there are several start-ups planning to provide similar (or even improved) high-throughput screening services for libraries of metabolic pathways.  I infer this is the strategy at Synthetic Genomics as well.  This all may work out well for both customers and DNA synthesis providers.  Again, I think people are working on an implicit hypothesis of radically increased demand, and it would be better to make the hypothesis explicit in part to identify the risk of getting it wrong.  As Naveen Jain says, successful entrepreneurs are good at eliminating risk, and I worry a bit that the new DNA synthesis companies are not paying enough attention on this point.

There are relatively simple scaling calculations that will determine the health of the industry.  Intel knew that it could grow financially in the context of exponentially falling transistor costs by shipping exponentially more transistors every quarter -- that is the business model of Moore's Law.  Customers and developers could plan product capabilities, just as Pixar did, knowing that Moore's Law was likely to hold for years to come.  But that was in the context of an effective pricing monopoly.  The question for synthetic gene companies is whether the market will grow fast enough to provide adequate revenues when prices fall due to competition.  To keep revenues up, they will then have to ship lots of bases, probably orders of magnitudes more bases.  If prices don't fall, then something screwy is happening.  If prices do fall, they are likely to fall quickly as companies battle for market share.  It seems like another inevitable race to the bottom.  Probably good for the consumer; probably bad for the producer.

(Updated)  Ultimately, for a new wave of DNA synthesis companies to be successful, they have to provide the customer something of value.  I suspect there will be plenty of academic customers for cheaper genes.  However, I am not so sure about commercial uptake.  Here's why: DNA is always going to be a small cost of developing a product, and it isn't obvious making that small cost even cheaper helps your average corporate lab.

In general, the R part of R&D only accounts for 1-10% of the cost of the final product.  The vast majority of development costs are in polishing up the product into something customers will actually buy.  If those costs are in the neighborhood of $50-100 million, the reducing the cost of synthetic DNA from $50,000 to $500 is nice, but the corporate scientist-customer is more worried about knocking a factor of two, or an order of magnitude, off the $50 million.  This means that in order to make a big impact (and presumably to increase demand adequately) radically cheaper DNA must be coupled to innovations that reduce the rest of the product development costs.  As suggested above, forward design of complex circuits is not going to be adequate innovation any time soon.  The way out here may be high-throughpu t screening operations that enable testing many variant pathways simultaneously.  But note that this is not just another hypothesis about how the immediate future of engineering biology will change, but another unacknowledged hypothesis.  It might turn out to be wrong.

The upshot, just as I wrote in 2003, is that the market dynamics of biological technologies will  remain difficult to predict precisely because of the diversity of technology and the difficulty of the tasks at hand.  We can plan on prices going down; how much, I wouldn't want to predict.