Why the University of California Is Appealing the CRISPR Patent Decision | Hacker News Reader

Why the University of California Is Appealing the CRISPR Patent Decision

theatlantic.com

133 pointsazuajef65 comments

Someone summed up the decision as such: UC Berkeley patented tennis balls. Harvard/MIT patented yellow tennis balls. Harvard/MIT won.

I realize this example is tongue in cheek and imperfect. I also dont pretend to understand the biology, but has anyone produced a good layperson's view of how Harvard/MIT won this?

Disclaimer: I'm a graduate student at UC Berkeley.

use of CRISPR in humans is what Harvard/MIT want to protect. According to the judges, UC Berkeley didn't make that clear.

Even being a biologist who understands CRISPR well and having trained at Harvard/MIT (disclosure to provide context about biases), it's not clear what Harvard/MIT will gain with winning this fight. Because the incentive is incredibly strong, I feel like everyone will find a way to work around whatever patents Harvard/MIT collect. I see UC Berkeley's side in these patents because their discovery proceeded Harvard/MIT and Harvard/MIT (Eric Lander) is just sucking up airtime like Trump does to come out on top. See this: http://www.michaeleisen.org/blog/?p=1825

To stretch your analogy: "Tennis balls colored for enhanced human recognition." So they're patenting a method to 'improve' CRISPER, in this case 'adapting' it for use in humans. To technically use the Harvard / MIT patent you may still need the UCB one.

Why this happened from my experience patenting through UCB and dealing with Harvard is that UCB does not have the best patent department when it comes to helping inventors draft patents that are defensible in terms of pushing for sufficiently general language, or including defensive dependent claims (like for use humans) for precluding people patenting specific extensions as Harvard/MIT did in this case. (They just put my patent memo in the appropriate legal format with a few minor modifications.) In contrast, Harvard/MIT as organizations are far more serious about intellectual property and have the legal departments to reflect it.

Berkeley's IP office is awful.

By comparison, Larry Page patented Page Rank. He assigned it to Stanford, Stanford prosecuted it and licensed it back to him, basically for nothing. There's a lot of trust there between student and institution. Furthermore, Stanford owns and defends the patent and no one fucks with Stanford.

https://www.google.com/patents/US6285999

I say this as a Berkeley grad. University IP departments should be more like Stanford and less like Berkeley.

> I say this as a Berkeley grad. University IP departments should be more like Stanford and less like Berkeley.

I disagree. And I think that anything funded with public grant money should be barred from being patented. If you want to patent it, you should have to fund it privately and take all the risk. If there is even a whiff of public money, you shouldn't get to patent it.

Look at how much patents retarded the 3D printing industry, for example. Those advances were very much funded by public money, and those patents basically made 3D printing economically infeasible until they expired. We could have had this revolution 30 years ago--that's a huge loss to us as a society.

The foundations of the computer industry rest upon the fact that many things escaped being patented because they came from Berkeley. While most people think of Unix, there was something possibly even more important--the circuit simulator SPICE.

The FORTRAN code for Spice2(g6?) became the foundation of the entire VLSI ecosystem that gave us the cheap chips that we all take for granted. It spawned an entire industry.

And Berkeley never wanted that to happen again. So they locked down the rewrite into C (Spice3(e2?)) and that code never went anywhere. And even 30 years on, it cost the NGSpice guys an enormous amount of duplicated work.

> I think that anything funded with public grant money should be barred from being patented.

Grants, green cards, and IP rights are the three pillars of research funding in the US. Any proposal to lop off one pillar needs to come with a plan to replace it or an argument why decimating the US research apparatus is the lesser evil.

> Grants, green cards, and IP rights are the three pillars of research funding in the US

IP rights are definitely not a pillar of research funding. The majority of universities lose money on their tech transfer offices[1].

[1] http://www.nytimes.com/2013/11/21/education/patenting-their-...

This. But the majority of these tech transfer offices are more like Berkeley's, jealously guarding their prerogatives rather than promoting students and faculty interests.

Stanford's model is that alumni who do well are generous. Also, alumni startups are like all startups, strapped. So unlike Berkeley, which wants their cut upfront, Stanford is patient.

So keeping with Berkeley examples, should the Regents not have rights (copyright not patent) for BSD? That too became the foundation of the entire ecosystem. But the Regents defended those rights (which they give away for free) from the UNIX System Laboratories lawsuit. You can't defend what you don't have.

Full disclosure: my copy of Lions' Commentary was entered as evidence in that lawsuit.

Maybe, but were there other cases where Berkeley helped the ecosystem instead of hurt it? How about elsewhere other than Berkeley?

I can point to a handful of technologies where the patents and copyrights got in the way without even thinking hard. I'm having a lot of trouble coming up with technology where the patents/copyrights helped spread it. Unix may be the only one I can come up with.

So you think that patents and copyrights get in the way of the ecosystem. But what do you think motivates developers to invent in the first place? Without that motivation where would the ecosystem be?

Famous case is that when Japan opened up to the world, they sent ambassadors out to learn what was going on on the outside. One said, “we have looked about us to see what nations are the greatest, so that we can be like them. We said ‘what is it that makes the United States such a great nation?’ and we investigated and found that it was patents, and we will have patents.”

http://www.iphalloffame.com/korekiyo_takahashi/

> But what do you think motivates developers to invent in the first place? Without that motivation where would the ecosystem be?

Developers develop because they feel a need to develop. No more, no less. Painting, music, writing, etc. are similar.

Creators create because they must create.

I'd say open source invalidates the assumption that there would be no advancement without patents.

All the while the evidence has piled up that patents actually don't promote advancement anymore: https://www.forbes.com/asap/2002/0624/044_print.html

> http://www.iphalloffame.com/korekiyo_takahashi/

Interesting, but I would point out that he was making those statements at a point in time where the US was literally ripping off anything and everything in the UK that was patented (aka: what China has currently been doing to the US).

That time period is, in fact, a prime example of advancement without strong patent protections.

The original Ethernet patent was an example of this. The license fees were low - $1000 [0] and it became a de facto standard. Token Ring was a competing network technology with considerably higher license fees [1]. Patent owners for subsequent Ethernet standards started out with similar terms but then started shaking down licensees [2].

The Bell Labs transistor patent may be another example. The initial license fees were $25k [3].

[0] https://books.google.com/books?id=qTkEAAAAMBAJ&lpg=PA52&ots=...

[1] https://books.google.com/books?id=OFCXnqlSFKwC&lpg=PA709&dq=...

[2] https://www.ftc.gov/news-events/press-releases/2008/01/ftc-c...

[3] (pdf) http://repository.jmls.edu/cgi/viewcontent.cgi?article=1175&...

I agree with your principle ('publicly funded work shouldn't be patented'), but that's not the legal framework we have, and it hasn't been that way since Bayh-Dole was passed in 1980 [1].

Until that act is reversed, it's foolish for UC to not behave more like Stanford.

1. https://en.wikipedia.org/wiki/Bayh–Dole_Act

Note that prior to the Bayh-Dole Act, publicly-funded IP was assigned to the government, where 95% of it lay unused.

It seems like it would be a good idea to have some kind of sub-patent, time-limited exclusivity granularity. For purposes of saying "I recognize that you are providing value to me as the patent holder by doing something with the patent that may not merit its own patent, but should nonetheless be encouraged (for my own gain and yours)".

Like a sub-piggyback patent that automatically includes licensing of the parent.

Given current IP law, I think the more relevant question is how public Universities should handle their patents (e.g. for public use, for a nominal license fee, or profit) and not if they should patent.

A non-existant or poorly executed patent allows for anyone to follow up, as MIT/Harvard in this case, and effectively patent minor extensions to the idea and still lock it up. Even if a public university wants their inventions to be free to the public, they need to properly patent it if it is going to stay available for public use. Eg. BSD style licenses for patents.

> I disagree. And I think that anything funded with public grant money should be barred from being patented. If you want to patent it, you should have to fund it privately and take all the risk. If there is even a whiff of public money, you shouldn't get to patent it.

What's the economic logic behind this? In these situations, the government is just another early stage investor. If it demands terms that are too onerous, the good projects will go somewhere else, leaving the government with the ones that don't have better options. Or, you'll basically make government funded R&D all into dead ends, because nobody is going to bother to spend the money taking it from lab projects into commercialization.

> If it demands terms that are too onerous, the good projects will go somewhere else, leaving the government with the ones that don't have better options. Or, you'll basically make government funded R&D all into dead ends, because nobody is going to bother to spend the money taking it from lab projects into commercialization.

Yet, somehow, we had almost a century of research and funding that included seminal things like radio, construction design and analysis, gigantic tools, tubes, transistors, etc. long before Bayh-Dole.

To be fair, I don't WANT government funding going into something that will make somebody money tomorrow. That's not what government research is supposed to be for, and it already has distorted "research" institutions into "funding" institutions where professors keep their best ideas close to the chest to go build a startup with. Besides, lots of people are willing to throw money at what looks immediately promising incrementally.

Nobody remembers them.

What everybody remembers are those rare times when funding flowed into things without any apparent use that magically transformed the entire technological landscape.

To be fair, CRISPR-based technology looks to be a much, much bigger deal than Page Rank. It's not really much of a stretch to analogize this to the invention of the planar transistor, or the internal combustion engine.

Deep breath. ICE and the transistor have been around for decades. Centuries in fact, in the case of the former. It's not even clear that pagerank is being used now; I can almost guarantee you it won't be in 50 years. Given the pace of biotech discovery I suspect CRISPR won't either.

PCR -- an essential step for all sorts of DNA analysis -- was worth $2B to Roche

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1523369/

https://www.theatlantic.com/science/archive/2017/04/why-uc-i...

"The UC’s filing covered gene editing with CRISPR, but it did not specifically cover eukaryotic cells—complex cells that include animal, plants, and humans—where all the lucrative applications are likely to be. The Broad’s patents did cover eukaryotic cells, and its lawyers argued that getting CRISPR to work in bacteria or free-floating DNA did not automatically mean it would work in the more complex environment. Hence, there was no inference—the decision that UC is now appealing."

https://www.statnews.com/2017/02/15/crispr-patent-ruling/

"The patent board said in its decision that the achievement of the Broad’s Feng Zhang in inventing a way to use CRISPR to edit the genomes of mouse and human cells “would not have been obvious” from the invention by Doudna and Charpentier “because one of ordinary skill in the art would not have reasonably expected a CRISPR-Cas9 system to be successful” in those higher-order cells."

The tennis ball summation is perfect.

However, you can go further with it. UC Berkeley has the right (or they will when their patents issue) to exclude Harvard/MIT (and anyone else) from manufacturing tennis balls. Harvard/MIT also have the right to exclude UC Berkeley (and anyone else) from manufacturing yellow tennis balls.

So UC Berkeley and Harvard/MIT will have to have a sit down and do a deal in order for anyone to get any yellow tennis balls manufactured. It's in their best interests to do this deal but first there will be some legal preliminaries.

(NB: we are ignoring Design Patents here.)

I think it would be really interesting if the Supreme Court stepped in an invalidated the patent based on 'natural function' of the Cas9 gene. Not that I expect that to happen just that it would be hilarious if it did.

Berkeley applied for a patent on tennis balls, an application still pending. Harvard/MIT later applied for a patent on yellow tennis balls and paid a fee to accelerate the review of their application. The Harvard/MIT patent was issued first. The recent PTAB trial found that the Berkeley patent application did not interfere with the Harvard/MIT patent, but this does not mean that the Berkeley application will not issue, nor does it mean that one could practice yellow tennis balls (CRISPER/Cas9 in eukaryotes) without licensing both from Berkeley and Harvard/MIT.

As to how Harvard/MIT won, part of the PTAB decision relied on general statements made by one of the Berkeley inventors as to the challenges in dyeing tennis balls yellow (getting the CRISPER/Cas9 system working in eurkaryotes) and so demonstrating that the Harvard/MIT inventors had to made an inventive step. That is a pretty profound and worrying development with respect to academic freedom.

Why is it a worrying development with respect to academic freedom? I think I'm getting the sense -- statements you make about how difficult some piece of work is can be used as evidence that the work is difficult, and thus as evidence that someone who does that work has done something inventive. But it doesn't seem like a particularly unusual revelation. Sometimes true things will be to your adversary's advantage. That's always been true.

I presume parent is worried that this will lead to less open exchange on the academic marketplace - you would not want to have public discussions about anything related to the work you are planning to patent, out of fear you might sabotage it.

That cuts both ways, in that open academic exchange can also be evidence that strengthens your own case. Although I guess the real risk is that lawyers would get end up involved in that process.

If something is "difficult" to do, it likely requires considerable capital investment. Shouldn't the commercial returns go to the entity that made said investment? It is important not to mix up the concepts of discovery, invention, and patent, in this case. It's worth noting that, per my other thread here, this all has been heavily adjudicated by the Supreme Court, with much resting on the "expectations of success." It seems that in such matters the nitty gritty details of the particular situation matter, and it can be hard to discern innovation from discipline. In actuality, courts appear to have sided more often with parties that have "reduced to practice" the claimed IP.

There has been tremendous lobbying, pressure and marketing from the Broad Institute on this issue.

For instance, see: http://www.michaeleisen.org/blog/?p=1825

I have personally talked to a student at the Broad Institute who told me a number of people there do not agree with the Broad positions (namely the Cell article and Lander's statements).

Disclaimer: Also a graduate student at UC Berkeley (not in Bio).

It is a tragedy that all this time and effort is being spent to create a temporally bounded monopoly which, in the long run, will likely hinder the use and development of CRISPR. Why is it not possible for patents to recognize multiple, independent sources of innovation and to award intellectual ownership of the ideas to all deserving parties?

Australia's patent on Wi-Fi show that patents held by the education sector can be big money for Universities. But in the article they're talking about how these two schools have already spent tens of millions on this. That's money that could have gone to research, scholarships, grants, and reducing the insanely high tuition fees students have to deal with.

I hate this model of how Universities, instead of seeking to help each other and help the world benefit from open innovation, instead are willing to pay huge legal fees so that one of the two gets a massive payoff while the other is stuck with a large legal bill.

There's another good example: H264, and the related audio patents. The Fraunhofer Society has funded years of university research just with the license money from MP4 and similar. Always keep this in mind when choosing between h264 and vp8/vp9.

> Always keep this in mind when choosing between h264 and vp8/vp9.

That's worded like you want to guilt me into approving of the actions of the patent pool. I'd rather not. Give me a link and I'll just donate $2 to them while strongly supporting AV1 over H.264/H.265.

They don't accept donations, they only accept tax funding or money from patent licenses.

Processing donations would be far too much work for such a research organisation.

And if you want to avoid giving away your right to sue for patent infringements, you'll want to avoid VP8/VP9 anyway, as the patent grant is similar to Facebook's and Tesla's patent license, and does end as soon as you sue.

Editas, a biotech company, is funding Broad's legal fees [1]. I had thought that another startup was funding Berkeley's side, but am having trouble finding a source.

[1] http://www.sciencemag.org/news/2017/02/how-battle-lines-over...

> Why is it not possible for patents to recognize multiple, independent sources of innovation and to award intellectual ownership of the ideas to all deserving parties?

Why is a temporary monopoly even necessary here? I've written about this before, but every time I see "patents" mentioned anywhere, I find myself thinking "Do we need this system?"

I asked on HN a while back for good examples of the patent system working as it should. The examples I got back were the Super Soaker guy and a sprinkler system someone's uncle invented -- but that's such peanuts compared to what would be possible if money fighting this case were spent on CRISPR researchers collaborating across organizations instead.

The arguments in defense of patents are so weak and yet so vehement. Delaying the cure for cancer is not a fair exchange for the Super Soaker guy getting rich from artificial scarcity in a child's toy category.

When Feng Zhang spoke at my school, it seemed that the use of patents was to prevent unethical modification of human embryos, but that hasn't seemed to deter scientists in countries where IP is not recognized (China).

If that's a genuine ethical issue then patents would be woefully inadequate at preventing it, since they would only provide a time-limited block. To actually fix things you'd need actual regulation, at which point the patent goes back to being bad.

> Why is a temporary monopoly even necessary here? I've written about this before, but every time I see "patents" mentioned anywhere, I find myself thinking "Do we need this system?"

In biotech and medtech no one will invest in your startup, even at the angel level, unless you have a sound IP strategy that includes patents or a very well-considered trade secret plan. The time from discovery to translation and commercialization is long, risky, expensive, and too difficult to predict to permit investment without some sort of period of exclusivity. If someone can clone a product as soon as it is developed and on the market, how could a company recoup development costs?

I've had this conversation before (not with you, but re: this argument) and I think you're right, it's hard to make certain areas of biotech work without having the ability to recoup costs (today handled with a temporary monopoly on the product, keeping gross margins very high).

One answer is "if you abolish patents a lot of science research will have to come from privately or publicly funded research" but this is pretty extreme and would be a tectonic shift in the biotech industry.

I'm not advocating an extreme solution like abolishing patents. But I think the pendulum needs to shift very hard in the other direction re: patents in general. Like, how about not having one court in Texas decide it all? Or "you have to sell the product to keep the patent and you can't just pretend to sell it". Or you can't patent software unless there's a really really good reason (like recouping huge development losses)

I feel like this only underscores the ethical problems of for-profit medicine. The argument, if I understand correctly, is essentially that there's no other way to fund large innovations in medical treatment in a free market than to have a patent system or some kind of artificial exclusivity. To me, this doesn't answer the question of what you're supposed to say to a dying person who can't afford to save their own life when a drug company unfairly sets their prices too high.

What do you see as the benefits of having a patent system?

Recognition is one thing. In the case of the MIT CRISPR patent, US8697359, that patent recognizes Feng Zhang as the inventor.

The patent also recognizes a LOT of prior art, in particular, publications from UC (A CRISPR Approach to Gene Targeting, ...) and UC patent applications (WO2016011080A3, ...).

Assignment, however, is another thing entirely. That patent has been assigned by Feng Zhang to The Broad Institute (MIT).

https://www.google.com/patents/US8697359

So pretty much the world is already the way you want it to be. Multiple researchers are getting recognized at multiple levels for their work. But know that there's a huge difference between recognition and assignment of ownership. For example, Transmeta did some innovative work in low power x86 emulation. But I don't think that anyone recognizes Intellectual Ventures for their now owning Transmeta's patent portfolio.

My apologies, I believe, after your explanation, that my intent was to use the word "assign" instead of "recognize".

Thanks for taking the time to explain the difference.

So if you are going to have a intellectual property system, then someone has to own that property. That starts with Inventor and then proceeds to Assignee (which never changes) through Legal Events to the current owner.

If you're gonna have a property system then you have to have ownership. You can have multiple Inventors and, failing Assignment to a third party, then each Inventor will have equal rights. Also, you really really have to declare everyone because if you don't they can come in later and claim ownership regardless of later transfers.

So there can be multiple ownership (but VCs are gonna hate that). Also, with the tennis ball vs yellow tennis ball example, there's a 'form' of multiple/conflicting ownership. Berkeley can't manufacture yellow tennis balls and Harvard/MIT can't manufacture tennis balls at all. Unless they both come to the table and agree.

> will likely hinder the use and development of CRISPR.

What is your thought process leading to that conclusion?

Sure, my thought process is as follows:

Hindering the development is the increased caused of licensing from a monopoly instead of a duopoly where they would need to compete. Higher cost of development hinders the use, as we are talking about a tool, wherein its usefulness is linked to its use.

Have you considered that the licenses made need not be broad, and that university licenses include milestone requirements to maintain the license?

I have not put significant thought into if making patents more specific would help the current situation. It could help or it could make relatively trivial derivative works a larger problem. That is actually some of what this article is about, wherein one group describes a tool in one system and another group describes it in a separate, more lucrative system. I'll leave the argument of if translating expression from prokaryotes to eukaryotes is trivial, but I am convinced the translation was significantly easier than the discovery and description of the system in the first place.

That's it. I will patent the process of picking up rocks in order to throw them at things.

I think it's a great achievement to discover the CRISPER/Cas9 system, but just as a rock - it was just laying there.

It's absurd that genes can be patented, and the ethics is mindbogglingly disturbing - I could understand procedures and methods for using or analysing genes, but the actual code as found in nature?

For those of you coming late to CRISPR and this debate - it's covered in a Radiolab podcast episode from February http://www.radiolab.org/story/update-crispr/

This is an update, but opens with the original CRISPR report from some years back -- featuring Jennifer Doudna from UC Berkeley.

After having spent several weeks studying the case law and patent procedures, I have come to the conclusion that Doudna/UC were naive and sloppy in their original approach to the CRISPR IP and public positioning. Morality aside, I believe Broad has the stronger U.S. claim, at least legally speaking. It's typically not the court's position to rectify the mis-steps of others. The processes are what they are, and the law is what it is, and the onus is on the claimant to properly protect themselves.

I'm confused about how either MIT, Harvard, or UC Berkeley owns the IP. Ostensibly the research was funded by the NIH or NSF, not those institutions. Is it really the case the the universities own the IP of NSF-sponsored grants?

Patents are awarded to inventors not funders. Grant agreements rarely require the transfer of patent rights, but employment agreements often do.

If it's a US federal grant from SBIR for example, even when retaining title to the patent, the researcher must grant a nontransferable license to practice the invention to the federal government.

Bayh-Dole Act. Since 1980, IP produced by government sponsored research is owned by the inventor, not the government.

https://en.wikipedia.org/wiki/Bayh%E2%80%93Dole_Act

That's for Federally funded research.

The definitive way to find out who owns the patent is to look at the Original Assignee and then the Legal Events to find out who the most recent owner is:

So Transmeta was the Original Assignee for US7100061 but Intellectual Ventures now owns it.

https://www.google.com/patents/US7100061

How can they patent this? It's been occurring naturally for millennia.

General Electric fought to patent life.

New Zealand banned software patents.

I feel like we're going to see similar analogies between bioengineering and software in the future.

They're not patenting the biological process itself, they're patenting how they can hack it.

Call me crazy, but I somehow doubt all of this fighting over profits/bragging rights/what-have-you will speed the rate at which CRISPR can be developed into a technology that actually saves lives.

Required reading if you want to really dig on this - https://acts.uspto.gov/ifiling/PublicView.jsp?identifier=106...

Also - see here https://www.law.cornell.edu/supct/html/04-1350.ZS.html and here https://www.pillsburylaw.com/images/content/6/5/v2/65484/Whe... Would definitely read the Transcript and Decision on Motions thoroughly from the prior posted link. The weight given to contemporaneous statements regarding "likelihood of success" is paramount. Bottom line - UC muffed the punt, imho, and their counsel didn't make things any better by taking the tack they did. Broad/Zhang drew a straight bead on the commercial IP from the get go and outmaneuvered Doudna/UC from what I can tell. Appellate court could certainly disagree, however, I believe they have a tough hill to climb in light of the record. From a scientific discovery perspective, it seems Doudna/Charpentier are likely to garner the lion's share of credit and accolades, but from a commercial perspective, I do believe Broad will come out on top. Doudna/UC is in a tough spot to claim that it was confident that CRISPR would work in eurkaryotic environments, and Broad had the jump with a better lab setup (and Doudna being crystallography focused). I also believe there is a bit of high level geo-politics at play here considering the Max Plank Institute and (Berlin) Univ. Vienna's involvement with the UC claim.

Reply on news.ycombinator.com