No Patents on Genes

The Supreme Court ruled unanimously yesterday in Association of Molecular Pathology v. Myriad Genetics that a gene, such as BRCA1 or BRCA2, does not qualify for a patent. The fact that Myriad isolated the DNA is not enough to distinguish it from its in situ counterpart as the information it contains is the same. However, cDNA, a version of the gene that has been stripped of non-coding sequences is subject to patent.

gene dollarWith this ruling the price of most of Myriad’s tests will fall as competition enters the market (the BRACAnalysis tests are actually a number of different tests, as I read the technical specifications, only some of these depend on cDNA. The markets appeared to have been initially confused about this.).  Even more importantly, the Myriad patents were broad and they prevented researchers from freely studying the BRCA1 and BRCA2 genes, from improving the tests or from developing additional applications. The giants demanded payment (video) from those who would stand on their shoulders. I think the restrictions retarded progress–as have similar restrictions–to an extent that made the patents difficult to justify.

Although I am broadly in agreement with the ruling, it’s also clear that the limited flexibility of patent law–you get a 20-year patent or nothing–and the fact that patent law is not based on patent theory (pdf) greatly hampers the ability to tailor patent law optimally. The ruling, for example, says that a firm can’t patent a gene that it discovers but it can patent the cDNA that it develops. It’s the discovery, however, that’s expensive. The development of cDNA is today a trivial step. Thus, you can patent the trivial step but not the giant leap.

You might think that the law draws a bright line between discoveries which cannot be patented and inventions which can but that’s not correct. Discoveries can be patented and the ruling goes out of its way to push back against the view that they can’t. The ruling correctly notes that a “considerable danger” is that patents on basic ideas and tools would “inhibit future innovation”. Yet the law makes no mention of these considerations and the court provides no guidance on implementation.

Coherent or not, the recent patent cases do indicate that the SC is no longer acceding to the United States Court of Appeals for the Federal Circuit–they are reestablishing control and pushing back in the right direction on the Tabarrok curve.


To be fair, SCOTUS hasn't acceded to the Federal Circuit in a long time, if ever. There are unanimous reversals once every couple of years on similarly big issues.

Agreed. Also the Fed Circuit has patent specialists who see a lot of patent cases. The Supreme Court has Scalia who complained in this opinion that he didn't understand the science!

To be fair, the law is based on a distinction between things in nature (discoveries are not patentable) and things created by humans (patentable). Where so much modern science is engineered, it's hard to distinguish what is discovered and what is engineered.

It seems like discoveries are a place for "citation," one of the roles filled by the incentives and rewards of the academia system.

At least he was honest. The actual opinion is riddled with errors.

I have never understood the concept of letting courts decide science.

Courts don't decide science, they rule on laws applicable to scientific matters. If not them, who?

Okay. I think there are cases where the courts drift into trying to define science. I'd have less laws to reduce the need for the court to wade close to that line. For example, there are ideas towards markets for patent protection and copyright "insurance."

That post lists two errors, one of which is basically a typo, and one of which is not an error, the author was just unaware of processed pseudogenes. I was actually surprised by the level of scientific understanding displayed in the opinion. Though I am definitely a fan of the humility of Scalia's opinion.

Re: "Coherent or not, the recent patent cases do indicate that the SC is no longer acceding to the United States Court of Appeals for the Federal Circuit–they are reestablishing control and pushing back in the right direction on the Tabarrok curve."-

You'd better get a trademark on that "Tabarrok Curve" (TM Bill 2013) thing pretty soon, or you might have to pay me.

Noone can really steal the "Tabarrok Curve."

Actually, I made up The Tabarrok Curve myself.

See? It doesn't work.

First use and confusingly similar to Alex's. I can see a dispute coming.

I first read that as "TM bill " and wondered what "trademark bill in 2013" Bill was talking about. :-)

At least for now, there remain levels of stupidity beyond which even our government dare not tread despite business lobbying.

The shortest description of this ruling is that you can patent photos of the sky, but not the sky itself. And yet I'm pretty sure you can't patent photos of the sky.

Well, you most assuredly can obtain a copyright a photograph of the sky. But that doesn't prevent other people from taking *other* photos of the sky; it just means they cannot use yours without permission.

People tend to overestimate the breadth of authority that copyrights. trademarks, and patents give to their holders. All three protect not generic ideas, but specific implementations of a general idea (at least in theory). It's a big step to go from "this particular photograph of the sky" to "the idea of photographing the sky."

you can 'patent' photos of the sky,

The word you were looking for was copyright.

This distinction comes up almost every time. I wonder how much of this problem is misapplication of the various IP modalities. Every generation the courts and policymakers have to learn why things are they way they were.

Really? Because the SC used the word patent. Maybe the word *they* were looking for was copyright? Or maybe you didn't read the decision?

And any particular cDNA is not a new process or invention, it's an almost perfect analog of a photograph of a naturally occurring object--nothing new is added, but a formula relates the "product" to the original. (like the formula governing the 2d projection of a 3d object)

It's like copyrighting a photo of a natural object of which, for whatever reason, only one photograph is possible. Under copyright law, one would imagine that in such a situation you could prevent people from copying your photo, but not from taking their own photo which (because of whatever mysterious quality of the object being photographed) turns out identical to yours.

Re: "Thus, you can patent the trivial step but not the giant leap."
To be clear, I believe the Court said that the "trivial step," cDNA synthesis, was "patent eligible," not necessarily patentable. An applicant would still need to demonstrate that its patent eligible subject matter satisfied the additional requirements for novelty and non-obviousness to obtain a patent.

"An applicant would still need to demonstrate that its patent eligible subject matter satisfied the additional requirements for novelty and non-obviousness to obtain a patent"

Aren't we the optimist!?!

Possibly guilty on that front. I do think it's helpful to distinguish between the different ways a given patent can fail to meet the objectives of good patent policy. The requirement for non-obviousness, of course, tends to ensure the advance is substantial enough to warrant protection, while the requirement that the subject matter fall within certain classes of subject matter tends to limit patent rights to fields in which the physics of the situation imposes some meaningful limits. This statutory subject matter requirement keeps us, for example, from issuing patents on truly novel, non-obvious short stories or musical compositions, fields in which economics and physics impose few constraints beyond what you can conceive, and fields in which a sufficiently abstracted patent claim could preempt an enormous amount of meaningful subject matter with little up-front effort. In contrast, a patent claim to something tangible, like a stapler, might also be broad and abstract, but the subject matter encompassed is only valuable to the extent the economics and physics of staplers yield a useful device. To my mind, the problem with patents on DNA, and software more generally, is that even for truly non-obvious and novel advances, the invention is ultimately about manipulating information, more like a short story, musical composition, or other subject matter we've typically excluded, and patents on these fields become questionable for the same reasons.

"It’s the discovery, however, that’s expensive. The development of cDNA is today a trivial step. Thus, you can patent the trivial step but not the giant leap."

However, I've seen nothing in the ruling limiting the ability to patent methods or machines to practice the 'giant leap'. If you can develop a method to make the giant leap easier and less costly, you can patent that and use that patent to derive value from how you share it. That makes sense - you have an opportunity to invest in the actual hard part and have an opportunity to derive value from it.

Yes. Exactly. In particular, "discovery" in the sense of discovering something that exists in nature isn't patentable: inventing a method for discovery is. The problem with "the hard part" here is everybody's working on new, cheaper, faster ways of doing it, so no one cares about the old hard way and the patent on that old hard way isn't of much value. The trick for getting from the whole gene to the cDNA, even if it's a cheap trick, is patentable as long no one else thought it up. Maybe cheap tricks are more valuable as patents than hard tricks. Hard tricks will quickly be outmoded by cheaper tricks, whereas a cheap trick may remain the only game in town for a long time.

Also, genes are getting cheaper and cheaper to find, although it looks as though there are fewer and fewer genes out there that have a big effect. We've already found all the major genetic diseases. Everything else is tiny increments in probability of getting a disease; where tiny is way smaller than blind random luck and environment.

My general rule, though, is that if Clarence Thomas says something I think I agree with, I need to think long and hard about that something to make sure I haven't made a horrible mistake somewhere. It's a really useful rule; even here, there's food for thought.

Even a blind squirrel finds a nut once in a while.

You need to update your patent & policy paper with Boldrin and Levine.

And for once, the evil rent seeking monopolists lose!


It's unfortunate that they are not burned alive instead of being merely denied their unjust profits, but it's a start.

Here is another question: genome sequencing is pretty trivial now compared to when the human genome project started. And yet the value of the genome is (let's say) trivial compared to the value in the near future. So, it seems pretty inefficient to incentivize exhorbitant spending when the price of sequencing is dropping quickly to create squatting opportunities for future windfalls when what is actually needed is getting from here to the place where sequencing is trivial. ?

Patent law is by its very nature difficult to implement logistically. Complicated theories of how patent law should work will often fail real life tests. While it would be great to simply tailor patent law perfectly if this isn't an option they you have to choose the best real world option. From all that I've seen about the state of patent law lately I get the general impression its more innovation suppressing then generating these days. So I tend to side against patent right in general.

"The key element in the economic theory is that pioneer firms have large, hard to recoup, sunk costs. Yet patents are not awarded on the basis of a firm's sunk costs."

I have a competing economic theory of the value of patents. It allows inventors of new, valuable ideas to coordinate with companies much larger than themselves with the means of production and economies of scale. Without good intellectual property protection, larger companies would just run roughshot over every young innovative company.

Too many observations about intellectual property ignore the extra value that start-up companies get on their exit price due to protections they have from the acquiring company simply copying their innovation (including patents, copyrights, trade secrets, etc.)

My competing theory is that the value of IP is not based only on the recoup of sunk R&D costs, but also on the value of eliminating the need to create excess production / commercialization capacity by start-ups where it already exists, which would occur more often if innovators have weak protections preventing established firms from stealing their ideas.

Except today patents are a way for large companies to squash start ups by owning huge databases of vague patents that could apply to anything and then threatening them with prolonged legal battles that they can't afford. It's basically an extortion racket, with lawyers acting as the hired goons.

Part of me wonders if the space on the Tarrabok curve above the red dot is ever possible to achieve, or if it is completely unaccessible due to other commercial trade-offs, legal trade-offs, regulatory costs and/or monitoring costs.

The only supporting data I have for the above is that the U.S. is still the most innovative country in the world by most measures.

The Supreme Court's decision reflected bad understanding of science.

Generally, what's to stop a company to setup shop outside the U.S. in a country where patent laws are friendlier? Would it have to abide by a decision by the SCOTUS?

What do you expect? Judges are scientists. As such it devolves into who can best influence a layman's opinion. This is why patent law that rests on high levels of government competence in evaluating their value is a doomed enterprise.

When patents start getting complicated they become less about invention and more about who has the most legal and political muscle.

You can patent the trivial step but not the giant leap. That's why we have the NSF. To cover the costs of making the giant leaps in basic research. Besides, there are other strategies apart from patents for reaping the financial rewards from making discoveries about nature.

The Supreme Court is even more clueless than Alex when it comes to patent law. And that's saying a lot! And Alex is just flat wrong on the law when he says Myriad (the patent owners) "prevented researchers from freely studying the BRCA1 and BRCA2 genes, from improving the tests or from developing additional applications."

Anybody can freely study a patented invention as long as they do not make, sell, or use the thing. Others can also improve the tests and develop additional applications as long as, again, you avoid use of the patent (absent a license). Improvements are developed all the time, which is one reason for cross-licensing of patents. The same for applications. You can get a patent on a new use of a known product.

"Anybody can freely study a patented invention as long as they do not make, sell, or use the thing"

Can you rephrase that?

Tabarrok is correct. The SC gutted the research exemption:

"Under the patent statute, one who uses a patented invention without authorization from the patent holder is liable for patent infringement.1 However, an accused infringer can escape a finding of infringement by invoking the common-law experimental use exception,2 which permits a de minimis use of a patented invention where the use was motivated by an experimental purpose.3 Recent decisions from the Federal Circuit addressing experimental use have demonstrated that the exception is extremely narrow and unavailable in most practical circumstances."


Reading over the linked paper on the mismatch between patent theory and patent law, it occurred to me that an interesting solution might be in sight. Have you heard of IPXI? They are a new financial exchange that seek to turn licence rights to patented discoveries into tradable commodities rather than relying on one-on-one negotiation with the patent holder.

Now, it seems to me that if IPXI or some similar IP exchange were successful in creating a deep and transparent market in commoditised licence rights, then the question of quantifying the potential benefit of a patented discovery is, for all intents and purposes, solved. All we would have to do is redefine the terms of patents from a time limit to a system where the patent expires when the patent-holder has earned a pre-determined amount of money from the licence rights sold on such an exchange (based on the costs of discovery and perhaps some reasonable profit on top) and then we let the market determine the value of the discovery's licence rights and therefore the rate of return to the patent-holder and the ultimate duration of the patent. if you want to know more about what these guys are up to.

Comments for this post are closed