I want to reboot the conversation about this patent, and, crucially about patents in general. Let me start by saying that it was not my intent to single out @anshulkundaje, as the problems with this patent are widespread across all of academia. But it is a useful case study. https://twitter.com/mbeisen/status/1199348013077917696

There are four overarching issues to discuss. 1: Should there shouldn't be patents in academia? 2: How patents have corrupted universities. 3: Why the arguments for academic patents are flawed. 4: And how this patent manifests these issues.

1: Should there be patents in academia? Virtually all academic research is funded by public or public-minded non-profits. The results of such research should as a matter of principle belong to the public, unless an iron-clad case can be made that patenting it benefits the public.

The arguments put forth for academic patents benefiting the public are a) they are required to productize inventions; b) they stimulate economic activity and thus benefit public indirectly; c) their revenue funds additional research; d) they incentivize researchers.

I do not think any of these arguments stand up to scrutiny. While there are arguably some cases where the transformation of publicly funded discovery into a product that benefits the public was aided by or even required patents on said invention, such examples are very rare.

Research looking at whether academic broadly promote the development and commercialization of products that benefit the public is at best, equivocal - and is certainly inadequate to justify of a system with so many other negative consequences.

And the way universities tend to license their IP proves the point. There has been a broad move away from exclusive licenses in which one company pays for exclusive use for IP and towards non-exclusive license in which anyone who wants to use the tech pays for that right.

There is no sane argument for how non-exclusive licenses promote the use of IP compared to placing the IP in the public domain. Such licenses are paywalls on ideas - they either cost companies money (and therefore slow their activities) or the prevent them from being used.

Now of course companies should be paying back into the system that generates the ideas they use. But the most broadly effective way to do this is by taxing economic activity after it has happened, not by placing upfront roadblocks, which non-exclusive licenses clearly are.

Ask any company that wants to use CRISPR in their R&D efforts whether it's better for them to have to pay licenses to all the parties who have claims on its IP compared to if it were all in the public domain.

The argument that academic patents stimulate the economy broadly is also faulty - it basically comes down to "the use of academic patents in industry has grown as more academic IP was patented". An uncontrolled observation that is pure correlation with no evidence of causation.

So even though the system isn't helping inventions get productized to the public good, maybe it's helping universities by bringing in money that promotes research?

This was certainly true at Stanford, while the department I was in as a postdoc was funded by cloning patent money. And my department at Berkeley benefited from revenue from Jim Allison's immunotherapy work and stands to do the same from Jennifer Doudna's CRISPR discoveries.

But, these are the exceptions not the rule. A few universities like Stanford make reams of money from the IP operations, but must struggle to even break even. In general it's a case of the rich getting richer - and it's hard to make a case that IP revenue truly boosts research.

While it's hard to find the good in the system, it's easy to see the bad.

I've written previously about how patents are corrupting academia http://www.michaeleisen.org/blog/?p=1981 . And I encourage everyone to read University Inc. by Jennifer Washburn - https://www.amazon.com/University-Inc-Corporate-Corruption-Education/dp/0465090524 a fantastic book that goes into these and other issues in detail.

While there are exception, by and large university IP offices are structured and act to maximize the revenue they generate for the university - indeed they are arguably compelled by the horrible Bayh-Dole Act to do so. Is this really what we want universities to be doing?

All you have to look at is the saga around who own CRISPR to see how the quest of universities to turn inventions made by scientists who work their into lucrative patents has corrupted not only the way academic IP is used, but the way academic research is conducted.

There are, in the CRISPR saga, countless examples of research planned, experiments done and sharing of data with colleagues organized not around what is good for science, but what is good for patents. And universities are fully complicit in this process, often even requiring it!!

It's depressing how often when I talk to colleagues about some cool new thing they've invented or discovered the conversation shifts to patents. I hear this everywhere. It's become a deeply ingrained part of the process of academic research.

I obviously can't know peoples' real motivations, but at no point in any of these discussions has the expressed motivation for heading down the path of patenting been to make sure the invention/discovery is used in ways to benefit the public. It's always about money.

And I get it. Many academics - especially trainees - are poorly paid and I don't begrudge them the desire to make a few extra bucks. And I claim no moral high ground here. As a graduate student and postdoc in the 90's, I received licensing revenue from software I wrote.

But as nice as that money was at a time when I desperately needed it, I still think it's nuts to argue that supplementing the pay of scientists with revenue from IP they develop justifies patenting IP that should belong to the public.

I certainly don't think the potential for patent revenue motivates many academics. So we can't make the case on the grounds that the public benefits from improved research outputs. Indeed, to me the most tangible outcome of this whole mess is the stymying of academic research.

It wasn't that long ago that if I as an academic research wanted a method or reagent from a colleague, all I had to do was ask and, unless they were worried I'd scoop them with it or something, they'd stick it in the mail and send it to me.

(more coming but stupid Twitter limits the number of posts you can include in a thread at one time)

What happens today? If I want a fly line from some other lab, they are required to consult with their lawyers. I have to sign an MTA that gives the university who's sending the reagent rights to any IP I might generate with it. And this MTA has to be looked at by our lawyers.

Sometimes the process is quick. Mostly it's not. And sometimes there is no resolution. The system serves ZERO useful purpose - all it does is reapportion who benefits if and when IP is generated. But the costs in terms of time and money - and research progress - are staggering.

That such an insane MTA system whose only effect is to slow research exists is a pure function of the desire for universities to maximize the value of their IP portfolio.

So, tl/dr, academic patents are horrible. They do not benefit the public in any direct or indirect way, they corrupt academia, and their most tangible impact is in slowing the research process.

Now, on to the patent that started all of this. Again, I want to emphasize that I am using this as a case study. Most of the things I am going to point out are true of literally tens of thousands of other academic patents. So please take this as criticism of process not people.

First of all, there is the patent itself. Claim 1 is so insanely broad that, if granted, it could be construed to cover literally all of the application of computational genomics to medicine. https://patents.google.com/patent/US20190287654A1/en

I know that this is how many patents are written, that lawyers try to get the most expansive claims they can, and that such claims are rarely - though far from never - given. But the fact that others corrupt the patent system is no excuse for universities to do the same.

Indeed the fact that university patents tend to look indistinguishable from company patents is another argument against the laughable notion that universities are creating such IP to promote the use of their inventions.

The potential damage from such claims being granted is immense - it could stymie entire fields of research for years and cost millions of dollars to resolve. As academics working for the public we should not allow even the potential for this kind of thing to happen.

At the very least, all academics who choose to patent their work should make sure that this kind of absurd language doesn't end up in their patents, and I commend @anshulkundaje for saying that he will talk to Stanford OTL about removing it from this one.

But even if you delete claim 1, and ignore the other places where there are overly broad claims (other parts of the patent cover fairly generic aspects of computational biology) and you read the rest of the claims of the patent narrowly, what is the point of patenting this work?

I think the work is cool and all, and I can clearly see how it would be useful. But I could say the same of literally thousands of other papers with similar novelty and utility published just this year. That's not a criticism, rather a testament to the productivity of the field.

Do we really want all of these works to be patented? Who on Earth benefits from patenting every new and potentially useful method an academic devises? What is the endgame here?

Imagine if every incremental advance in algorithms for computational genomics were patented, and that non-academic researchers had to obtain and pay for a different license from a different university for every step of every analysis they did. It would cripple their work.

And this, ultimately, is the real problem. It's not that @anshulkundaje's patent on its own is so destructive - one can always pay for or work around one patent. It's the collective damage of every similar type of work being similarly patented that is the real threat.

And it is to avoid creating a science dystopia where every quanta of academic output is protected IP and thus where nothing is free or simple to use that I call on all of my colleagues and all of academia to never patent anything.