ARPA-H would launch 1,000 drugs, Eric Lander promises NIH advisors

Francis Collins: We’re about to hear from Eric Lander, who obviously was an incredibly important contributor, more than any other part of the team, to the success of The Human Genome Project. 

Along with him is Tara Schwetz, who as I told you yesterday, is now on detail from her role as associate deputy director of NIH to OSTP, and specifically working on the topic that Eric and Tara have come to speak to you about, which is this proposal of a new component of NIH called ARPA-H.

I think most of you know Dr. Lander, and I will not use up a lot of his time with a long involved introduction, other than to say he has been confirmed, as of two weeks ago, as the White House science advisor, the director of OSTP, and a member of the president’s Cabinet, and we are all really glad, for the first time, to have a science advisor who is a biologist, and to bring with him all of the vision that he has been famous for, for several decades. 

So, Eric and Tara, thank you for coming to join the ACD. We are all in rapt attention about what you’re going to tell us about ARPA-H.

Eric Lander: Thank you, Francis, and first, I have a big, big thanks to Tara, who has been working full-time on this project. I think our goal is to start a discussion with the ACD, and in fairness, Francis has been very involved in this, as has Larry [Tabak, principal deputy director, NIH]. 

And so, when we turn to general discussion, they’re as much involved in the thinking, but the thinking is evolving, and input from the ACD I know would be very helpful to Francis and Larry, but also very helpful to me and Tara.

So, I’m going to quickly, in a modest number of slides, describe the conversation that has been going on around ARPA-H. Tara will elaborate on some of the practices and cultures that DARPA brings, and then we’re throwing it open to discussion. 

We are not asking for anything from the ACD, other than your input and ideas, so there’s no proposal on the table, there are ideas all over the place here, to talk.

The first thing to say is, we have an amazing biomedical ecosystem in this country. It is just extraordinary, especially for those of us who’ve been around for maybe decades, to watch this ecosystem go from some early ideas about what you could do with molecular biology to an incredibly vibrant biomedical research ecosystem over the last four years, and a vibrant commercial sector.

I’m not bringing any news to you, but I still feel obliged to say that these two gems, fundamental biomedical research and the vibrant commercial ecosystem, connect in this virtuous cycle. 

You know that there are many funders of biomedical research, but, by far, the dominant funder in this country is the NIH, and it produces knowledge that is of public good. And that’s what’s so powerful about that knowledge, is it’s available to everybody. Often it’s knowledge that you couldn’t appropriate anyway. No company’s going to go do this, because it’s knowledge knowledge.

It’s not a patentable invention, but it underlies, for example, oh, every therapeutic that ever gets developed. We can’t find a therapeutic that can’t, in its parentage, trace back to fundamental discoveries that come out of the NIH. 

It’s not to say that companies don’t make amazing contributions, it’s just, without this basic research contribution, we wouldn’t have what we have, and it’s been breathtaking what’s been going on, and it just keeps getting faster and faster.

The same thing’s got to be said about the vibrant commercial sector. There, companies are focused primarily on R&D, and marketing of specific products to bring systemic therapies to patients; therapies and devices to patients. And for lots of reasons, significant capital is available to that sector to develop products if they can generate profits. 

They’re not in the basic research field, but if they can produce products that really make a difference and can make a profit, there is funding to be had for startup companies and large companies. And there are 8,000 medicines that are reported to be in development, which is extraordinary, including 1,300 just for cancer alone.

So, with all that, why think of anything else? Why isn’t it all done? And I think for many purposes, this is all you need. 

However, there are at least two reasons why we should always be thinking that even though we have the most amazing biomedical enterprise ever created, why couldn’t it be better? Maybe it’s not doing all the things it still could do, despite doing lots of amazing things.

One reason is, events in the last decade or so have made clear how much is possible. Cancer immunotherapy is a perfect example of that. It wasn’t very long ago there were no cancer immunotherapies, and now there’s such an explosion of work on cancer immunotherapies, a whole new area has opened up.

But the most recent example, of course, is we have all sorts of vaccines of many different stripes and flavors for COVID-19. They’re all remarkably efficacious at avoiding serious disease and death, and is allowing this country to begin opening up back again—toward the barbecues, and normal celebrations, and things that we celebrate on the Fourth of July. We’re going to be back in many ways to doing that. 

We’re making really substantial progress, and there are many, many people who’ve contributed to this, from essential workers, to healthcare workers, to the people who drove people to get vaccinated, and set up clinics and all that.

But underlying it all, we really have to remember that there’s a tremendous amount of science that has contributed to this, too. And it was not even a year and change ago that everybody was saying, “Well, you can’t make vaccines in less than four years. Never been done before.” 

But it turns out, it can be done. And it makes you ask, “Well, what else can be done that we might not think could get done, if we can organize with the degree of organization and urgency that have been applied in some of these cases?”

So, what can we do to fully realize the promise of these advances in biomedicine by dramatically accelerating the pace of breakthroughs? What can we do to transform how we prevent or cure cancer, infectious disease, or Alzheimer’s disease, or other diseases? 

Or very importantly, to transform other aspects of the healthcare system: access to healthcare, healthcare equity, quality, reduced health disparities. There’s a whole set of questions, and don’t get me wrong, basic research is foundational to everything. 

But the question is: Is there more that can be done, or are there things that are missing from the ecosystem?

And there are. There just are. Sometimes really great ideas, let’s say like making a vaccine for COVID-19, don’t fit in so well with either of the two pieces. Telling a study section that you’re going to build and make a vaccine against COVID-19—I’m picking on that example, but for example, there are some things, which in our standard, basic research paradigm, are just too risky to undertake. 

Or the cost is just too large for the risk. Or the time frame is just too long for the mechanisms that exist. There’s no criticism of those mechanisms. Those mechanisms in basic research are the foundation of all of this revolution, but there are some things that don’t fit in so well with typical mechanisms.

Sometimes projects will be very important, but they’re just too applied for a purely academic focus. Sometimes they may involve complex coordination amongst multiple parties of the sort that simply writing a lot of subcontracts and a grant isn’t going to be the right solution. 

They might need more coordination, even, than some of the current mechanisms do. Sometimes companies are not going to invest in something because the near term market opportunity is just too small to justify the investment, even though a long-term opportunity might be huge.

And then in the extreme case, something might be a huge platform that is just plain too broad for any one company to do alone, because they can’t actually derive the full benefit of it. 

Economists taught me some time ago that a private actor will invest in proportion to the benefit that they privately derive. But some of these kinds of projects are so broad that no one private actor can reap the full benefit, and therefore the economists would say, they will underinvest compared to the social optimum.

Now, I don’t think it was this apparent 10 or 15 years ago that there were lots of projects like this that had multiple of these characteristics, but it’s becoming clearer and clearer that there are. Now we’ve seen them, even going back to The Human Genome Project.

The Human Genome Project was sort of like this in many ways. And it did get done, but it required creating a whole new center and institute at the NIH to do it. And so, you can pull these things off, but there’s not an easy way to do it. There’s not a standard way; they’re out of a process. 

And it makes you ask: are there a lot more things that fall into this category and could be done?

Well, President Biden has, for some time, during the campaign and recently in several speeches—most recently in his joint address to Congress—said that, and I’m just going to quote because it’s a great quotation from his wonderful speech, “We will see more technological change in the next 10 years than we saw in the last 50 years.” I don’t think anybody here disagrees with that. “And we’re falling behind in some ways in that competition, and the NIH should create an Advanced Research Projects Agency for Health.” 

Well, that, and I can think of no more worthy investment, and I know of nothing more bipartisan. So in the president’s budget, there is a request from the president for $6.5 billion available for spending over three years, to create an ARPA-H.

It’s an exciting idea. It’s based on the idea of DARPA, and many of you know DARPA. This is a defense department entity that was created shortly after Sputnik to be able to undertake breakthrough research, do things differently, and it has led to some remarkable products and achievements. 

Things that you’ve heard of, like the internet or GPS, or the early work on self-driving cars, and which actually contributed to mRNA vaccine development. So, there’s a whole lot of things that DARPA has done, and it makes one ask: are there lessons to be learned from a complementary approach to funding things? 

Again, no criticism of the way we fund things today, other than there might be additional things that could be thought of—and DARPA is the inspiration for the president’s idea, which he’s been talking about for quite some time.

This is not an official mission statement. This is a mission written out to provoke discussion, a little bit inspired by the way DARPA talks. To make pivotal investments in breakthrough technologies and broadly applicable platforms, and capabilities, and resources that have the potential to transform important areas of medicine and health, and that cannot readily be accomplished through traditional research, or commercial activities. 

And there’s an emphasis on building platforms, capabilities that might launch 1,000 ships, 1,000 projects, for example, because many things can we do with those. But at the same time, there might be instances where there are things that are very important but just plain uneconomic to do, but would be extremely important to the system. That’s kind of a rough idea of what’s being talked about.

I’ve written here a bit of a summary of what I’ve said already, but the focus is to find ways, platforms, methods, to revolutionize prevention, treatment, cures. It’s going to build on all this foundational research, but carry it to the next step. Convert use-driven ideas. 

We often think about curiosity-driven ideas, and they’re very powerful, to ask “why” questions, but use-driven questions are also powerful. How could I do something that would be useful in some way, into tangible solutions much faster than you might have thought possible?

Speed is an element here. Risk is an element here. So making high-risk investments in these broadly applicable platforms, capabilities. Fostering breakthroughs across various levels, from the molecular, to be sure, but also to the societal, because breakthrough platforms might equally well happen at the societal level, and drive them to the point of adoption, where industry will pick them up, or someone else will pick them up. And as I said, overcome market failures.

I think there are lots of areas that are ripe for this, and certainly this has been an idea that has been suggested by many people over the last several years.

As an example, imagine that we had molecular ZIP codes, some way to target your favorite drug to your favorite cell type. And I don’t just mean get something to deliver, because we know how to do that. I mean pick any cell type, and know that you have a ZIP code that you can attach to it that will get it to that cell type with a very high partitioning.

Well, suddenly that has huge effects for eliminating those serious side effects that are due to the effects of the drug in a tissue other than the target you want. So, targeting things to specific cells—for example, tumor cells is the most obvious case—would give you a much bigger therapeutic index that you could have any other way. 

But the idea of someone writing a grant, or starting a company to say, not just, “I’m going to try to target the pancreatic beta cells, but I’m going to try to get a whole post office that can target to every cell type.” I don’t know that this would fare well. 

I wouldn’t think it was a great idea for a foundational research project in the study section, nor for a great business plan, but a platform like that could launch 1,000 research studies and 1,000 drug applications.

But similarly, going from the molecular, we could ask about holistic interventions to eliminate racial disparities in maternal morbidity and mortality and premature births. There are systemic solutions to that, that may require scale, for example, or coordinated complexity that won’t fit in either of the two models as well. 

So, I want to say molecular to the societal, and there are other examples, and you’ll think of your favorite examples too. How do you fill in this middle bit? Well, a lot of it is cultural.

You’ve got to be able to seek bold ideas and people. That means you have to be welcoming to bold ideas, to risk. You have to be nimble. Saying three or four years later, “Yes, after four revisions this is a really interesting project you should try it,” may not be the best thing to do here. 

So, nimbleness, openness, transparency, failing fast, and then taking responsibility for including the responsibility that it’s okay sometimes to fail. But maybe not always is a good amount of time to fail, but sometimes.

The idea that’s proposed in the president’s speech is to have this entity coupled closely to the NIH, and part of the NIH, so it can draw on the vast knowledge and expertise, and activities of the NIH, and that’s why the president in his speech put it that way.

But, it should be different. Having yet another institute is not going to really answer the need, the opportunity. It should be something with a unique culture, a unique organization, and an independence to really take risks and do things in different ways that can complement.

If you look to DARPA—and Tara will say a little bit more about this—DARPA brings interesting examples of culture. They take on time-limited projects with goals and benchmarks, and accountability. They have a very different project review process and project selection process. 

They assemble portfolios of projects, rather than each project being chosen individually. They’re reviewed individually, and they work with groups across academia and industry and governments. You could even imagine them working with the ICs at the NIH. There’s no reason why performers might not be there.

You’d want, again, looking to DARPA, DARPA has a director with technical leadership skills, ensuring that the organization is serious about risk tolerance, urgency, innovation. And they recruit program managers for very short terms, three, four years, and then they turn over again, going back to their homes in academia, or industry, or government. They have individuals, they have a budget, and they have broad autonomy to drive transformational change. 

You sure wouldn’t want to run the whole NIH like this, but maybe you should run a little bit of the NIH like this. And you’ve got to figure out: How do you protect the distinct culture, and how do you get the cross-fertilization benefits of doing things different ways?

Then, equity is very important. I want to underscore that thinking about simply molecular platforms, this is an incredibly important obligation and opportunity. Some projects, for sure, should be focused squarely, wholly on addressing health equity, because I think there are amazing things you could do to address health equity if you think about building new platforms. 

But all projects should consider their impact on equity, because even when you might be developing a cancer vaccine or something else, you should be thinking pretty early on, that if this thing is going to go into use, well, what are the implications for equity? Who’s it going to be available to? Who will it serve? Who won’t it serve? 

And then, of course, in building an organization of any sort, ensuring the equity in staffing is very important, because who’s around the table is going to affect these other things, like what you’re going to think of to work on.

Conclusions: There are a lot of areas that are ripe for some transformation with the right support, the right collaboration.

A new entity, such as the president has proposed, could help unlock some of those opportunities, but designing it in such a way that it really will deliver on those kind of goals requires a lot of care and thought, and Congress will be very much involved in it, but I think the ACD is a perfect group to be thinking about, what are the elements to make such a thing a very important complement to this amazing biomedical ecosystem?

That’s it. I promised not to be too long. So, Tara, over to you.

Tara Schwetz [assistant director for Biomedical Science Initiatives, OSTP, EOP]: Thanks, Eric. Nice to see you all. I’m going to provide just a little bit of high-level background, as Eric mentioned, on DARPA, and more broadly this ARPA model and how it operates, with the full caveat that I’m not an expert on DARPA, although I have had the opportunity over the last couple months of meeting with many of them.

But I thought I’d outline a little bit of what I’ve learned for those who are less familiar with this model, and hopefully it will provide some context and insight into what Eric just presented, and how we’re thinking about ARPA-H.

So, as Eric mentioned, DARPA was launched in the wake of Sputnik, and it had a singular mission: to make pivotal investments in breakthrough technologies for national security. And it’s had many successes as was mentioned: the internet and GPS, of course, are two very distinct ones. But it attributes its success, at least in part, to its distinctive organization and culture, that really allows for innovation to thrive through autonomy.

It’s defined by its willingness to take risks, its flat and nimble organization, whose work is driven by a lean workforce of program managers. They’re often recruited from industry or research universities. And they bring with them these really bold, risky ideas, and are encouraged to pursue them. 

They’re also given the freedom to do so by providing them with significant resources. And they help mitigate the risk that’s associated with that by pursuing multiple different approaches to achieve a singular goal. So, they’re empowered to promote collaboration and integration across their performers, which is, they term their awardees. 

And they can make decisions to modify funding if a project is underperforming. They can also encourage or recommend, I guess, projects to be combined into one singular project.

Program managers come for limited terms, just a handful of years, and they manage projects that are also time-bound. And this really generates a sort of sense of urgency that contributes, as some folks say, to this DARPA magic. 

They utilize a very different review system than what we’re used to across NIH. Their proposals are reviewed, but they’re not ranked. They are done so on a competitive basis by experts from across the federal government. 

And the program managers have the authority to select portfolio projects that are intended to achieve a quantifiable program goal. So, they’re really aiming for the best compilation of projects to achieve that goal.

DARPA’s able to do this because they make use of broad, flexible hiring, procurement, and contracting authorities that they have at their disposal. And there’s some fundamental principles that I think will be important to mimic if we want to be successful in establishing our ARPA for Health.

But we know and acknowledge that it’s not a perfect model for biomedical and health research. And so there will need to be some obvious tweaks, because biological systems are really complex, and often more poorly understood than many of the engineered systems. And they interact in a really complicated world of systems and users, and with human behavior and social factors. 

So, that’s not to say that some DARPA projects don’t face similar challenges, but it’s important to recognize and consider the different environments that these bases operate in.

So, we have a lot to learn from, and build off of DARPA, and other ARPA programs across the federal government, and look forward to continue to learn from them, and also hearing from the community about this model, as well as the challenges and opportunities that are envisioned for ARPA-H.

Lander: Thank you, Tara. Francis, over to you.

Collins: Thank you, both of you, for putting forward the framework here, and I think we’re going to be very interested to hear from ACD members. Just a couple words here from me, in terms of the promise of this.

I think the experience of the Genome Project, as Eric has mentioned, was one in which I began to realize there are things that are really important to get done, that perhaps our normal framework doesn’t necessarily make it easy. And we had 13 years to get that one done.

And then, more recently, as projects like the Brain Initiative, we also have been interested in flexibilities. We’ve been learning, as you have heard in the course of this ACD presentation, how to use this mechanism called Other Transaction Authorities in places where we needed to really move quickly.

And certainly, over the last 18 months with COVID-19, to see what was possible, and had to be possible for the development of safe and effective vaccines, for setting up therapeutic trials at record speed, and for developing this platform, this innovation funnel for diagnostics has certainly been a wake-up call for all of us about what we might be able to do—even if it wasn’t a crisis pandemic, to speed up the kinds of developments that will provide great advances for people who are still waiting for those to happen.

Let me be clear. This is not the way that you would do an awful lot of what NIH needs to do, and nobody should feel threatened that somehow this is an attempt to downgrade the importance of investigator-driven, hypothesis-driven basic science. That will always be our foundation. 

But there are times where we could make things happen twice as fast than they otherwise would, and with high quality, and with the opportunity to speed up clinical benefits. And as we’ve done with COVID-19, why wouldn’t we want to do for cancer, diabetes, Alzheimer’s, hypertension, rare diseases, common diseases, all of those things, which have components that are ripe for this, at this exceptional moment of scientific history, where technological advances are just waiting to be pushed forward.

So, I’m very bullish about the potential here to try to make this happen, but also sobered about the challenge, and the need for a real effective strategy, both in terms of how it would operate and how you manage the culture change part of it, which is going to be very significant. Otherwise, it will flounder.

As Eric said, this is just going to become another typical institute at NIH—as much as I love all of them, that will not be the culture that we will need for ARPA-H.

Roberta Diaz Brinton [director, UA Center for Innovation in Brain Science, The University of Arizona]: So, to dovetail, actually, on what you’ve just spoken about, Francis, is the culture change. Having been involved with DOD on projects, the culture is quite different, and the project management is quite different. 

And so I’m wondering whether this kind of DARPA culture translates entirely well to the ARPA endeavor. I love it, it’s part of what we’re doing here in the center. 

It’s … yikes, there are timelines, and you got to get it done, and it doesn’t matter. So, it works, sometimes well in engineering. Go. Ideas.

Collins: I totally get that, and I think it would not work well in many other parts of our current organizational structure, and that’s why this needs to be a separate division, it needs to have a director who is brought in who has that kind of cultural mindset, and who is not coming in to sort of spend the next 25 years running this. 

This ought to be a shorter term effort for just the right person, who’s got lots of vision, and is sort of familiar with the Heilmeier Catechism, about how you decide to pick projects that are a good fit for this, and recognize that a lot of projects really aren’t.

And then, the project managers also are very different than what you would call our current program staff. There will be some program staff at NIH that I think have this capability of making this switch into a very different kind of mode of operating, but I don’t know that will be true of a lot of them, and a lot of these are going to need to be folks that are recruited, I think, from outside, who come for two, three, four years, but not longer than that. That’s been a very successful approach.

And you may get a flat organization. You don’t have a big complicated bureaucracy with lots of layers. Again, I think we’ve learned a lot by what Tara’s just described here, in terms of the DARPA model. And it doesn’t completely fit, but a lot of those lessons could be adopted.

Then, you don’t want to constrain it. You want it to be able to really operate nimbly, using not just Other Transaction Authority, but also it’s going to have to have its own very nimble contracting capabilities over multiple years, otherwise it’ll get all snarled up in those kinds of slowdowns.

Collins: So, I think we can see how to get there, but to just take what we’ve got now and say, “Okay you, become ARPA-H,” that’s not the way it’s going to work.

Brinton: One of the things that we’ve done here is try to build this hybrid model of a biotech in an academic environment. Sounds good, but it does have its challenges. So I’m very excited about the ARPA idea.

Mark Dybul [professor of medicine; co-faculty director, Center for Global Health and Quality, Georgetown University]: I was actually privileged to be on the Independent Panel for Pandemic Preparedness and Response, which was mandated to look at where we are, and what worked, what didn’t work, and what we should do for the future.

And one of the key recommendations is creating something exactly like you describe for pandemic preparedness response, but to have it established on an ongoing basis for those regular investments, so that we have things ready to go the next time. So, this is fantastic, fitting into that.

The second thing I would mention related to that, well, second and third, that I think we learned in part from the response to COVID, was first the global engagement; that no one country did well, and the remarkable intersection of scientists and companies from across the world was pretty remarkable. 

As you know, the European Commission’s trying to set up a BARDA-like mechanism. I’m wondering if it’s already in your thought processes that this place should be linked very much to other places around the world, so that there’s that global interchange.

And the third piece, I think we learned also from COVID, is the importance of the private sector being at the table from the beginning, which picks up a little bit on what Roberta said. 

Looking to the future, I personally wouldn’t be looking at Big Pharma, I’d be looking at what Big Pharma looks at as its incubator, which are biotechs, where they buy up all their products from academics and from biotechs. 

So, how to involve, not just by having people secunded, but how to engage and make sure the vision is out there to bring in and to welcome biotechs to apply to be a part so that they’re driving that engine of ingenuity and innovation.

Collins: Those are just great points. Eric and Tara, the idea of having a global health connection. Has that already been on the table, because I think that makes a lot of sense?

Lander: There’s already a lot of discussions, as Mark is indicating, both around pandemics, where I think there’s shared interest by a number of countries in taking on the kind of things that the IPPPR was recommending. I know that the U.K. is creating its [Advanced Research and Invention] Agency. 

So, I think everybody’s trying to make sure they create things right in their country, but they’re all looking at each other and saying, “Yeah, there are really a lot of opportunities to take on things together.”

It’s no small task to really get a culture right, and an organization right, and not disturb all the great things that we already have going, but we should be looking to the fact that as we build this, there will be allies to be coordinating with on some of the most exciting of these areas.

Collins: Yep. I do think the experiences we’ve had, both with vaccines, where of course we had to work closely with companies, mostly large, but also with the diagnostic efforts, where of the 32 technologies—that we were able in a sort of mode where NIH took on the role of being venture capitalist, which is kind of like DARPA—those 32, I think 28 of them were small businesses that had the chance to do something they couldn’t have done otherwise by this program. There was a lot of failure there too, but we learned a lot from that experience.

Dybul: Sorry, but if I could just follow up, Eric. I totally get that it’s complex to build something in the U.S. government and focusing there first, but having run and started both [President’s Emergency Plan for AIDS Relief] and then transforming The Global Fund [to Fight AIDS, Tuberculosis and Malaria], there’s something different when you have the global interchange that can actually, I think, help you create a better U.S. government agency. 

And that might just be from those experiences, and I might be wrong, but I know how complex it is in the U.S. government. I think there could be an opportunity to blend the two.

Lander: I think those conversations are kind of happening, but it’s good form for us to think about the design problem within our borders, but you should imagine that there’s a lot of exchange already happening and a lot of cross-influence. 

A lot of people look to the U.S. for models. The U.K. is building an ARIA based on an ARPA model, so there’s a lot of conversation going on. I think more and more that conversation will get exposed over time. 

So, I didn’t mean to suggest it was purely serial in order, I was just, for the purpose of this one meeting, suggesting that focusing on the design problems within the NIH was probably enough with the hour allotted. That’s all.

Francis Cuss [Retired executive vice president, chief scientific officer, head of Research and Development, Bristol-Myers Squibb]: It’s very exciting indeed, of course, and one seeing what DARPA has done is terribly imaginative. But I have a sort of test set for a really difficult problem when you set this up, which I have been wrestling with peripherally for some years, and that’s the problem of antimicrobial-resistant drugs. 

And there, there are roadblocks in discovery and development, in the regulatory space, the policy space, the commercialization space, the international space. I don’t know whether that’s a big enough thing for you to want to bite off, but I think we can all agree how important it would be. 

And it’s almost too big for the NIH, I would say, or at least it needs someone who can dip into other parts of government. It’s not just a scientific problem. So, I just give you that, it helps me to frame how one might solve very big problems.

If that’s what you’re trying to do, I think that’s fantastic, and it would be extraordinary if you could bring together the people who could actually solve on all those different dimensions.

Collins: As I recall, Eric, you were a co-chair of [President’s Council of Advisors on Science and Technology] that wrote a very significant piece about [antimicrobial resistance], so maybe you have some thoughts about this.

Lander: Well, I want to say this is very close to my heart because in the PCAST some years ago, we did take on many of the opportunities and desperate needs, so I care a lot about this.

I think an ARPA-H taking on, say, all of AMR as a project is probably too big a scope for a project, but within AMR, I think there are a lot of platforms that could be made that might lead to scores of antibiotics, or challenges like: 

How do I reactivate an antibiotic for which there is resistance? Are there other agents that can now make organisms susceptible again to them? I think there are the challenges around host metabolism, where you’re not going to encounter resistance.

So, I can imagine an ARPA-H having four or five audacious projects in which, within those projects, it would say things like … and this is going to do no good unless there’s a regulatory path for it at the end, so we better start thinking about it.

And so, one of the bullet points was about the complex coordination, so, within antibiotic resistance, I think there would be some amazing projects that should be looked at holistically that would give rise to many agents. 

And that would be an exciting thing for me, because I think AMR is exactly the sort of thing that is hard for a company and hard for a research grant, even a program project research grant. It’s not scoped as you describe. So I’m very excited about that as an instantiation of the concept.

Schwetz: I was just going to add one thing, to piggyback on what Eric was just saying. I think ARPA-H could be viewed as an abler and a catalyst; right, so, facilitating collaborations across the federal government to tackle an area. 

So while ARPA-H might only be focused on one, or two, or four, or whatever the number is of the approaches to solving problems in one’s specific area, like AMR, other agencies might come in and be able to coordinate with ARPA-H and build off what’s being supported there, to fill in other gaps and to accelerate revised policies or processes.

Cuss: I think that’s a fantastic point, and I suppose, Dr. Lander, I’m hoping that you have a very big voice and you sit at a very big table. If you are passionate about AMR, you would have that opportunity to encourage others to do the things that would make the antibiotics coming through actually be bought by someone, and paid for, and all that stuff.

Lander: I am passionate about that topic, and I will do what I can to overcome my shyness.

Collins: We have confidence that you will somehow manage that.

Shelley Berger [director, Epigenetics Institute, University of Pennsylvania]: That was really exciting. It’s a tremendous program. I’m curious about your thoughts, Dr. Lander, about basic research in this program, compared to more clinical, the sorts of things we’ve been talking about so far. 

So, I just wondered, maybe a little bit about tagging on what Roberta mentioned, how to fit basic research into this sort of paradigm. And to give you an example, maybe something you’ll know about, we’re both interested in, I’m sure, is genome architecture. We don’t really know what the significance of it is, functionally. It’s a big question in epigenetics. 

And here’s a big question, and this would be a great question for this sort of program, but how do you fit timelines and this sort of thing? I can’t quite imagine it, so I wondered if you could speak to any sort of example of a basic research project and how it would fit?

Lander: It’s a great question. I want to be careful to say, I think the vast majority of research should go on through the standard research pathways, where you do foundational research with R01 investigator-initiated grants, because I still think that’s the most powerful system that’s ever been invented for discovery.

But I think the premise was there are things that don’t fit in the existing buckets, and I think there are aspects of basic research that will not fit in those buckets, so that’s the test. If you can do it through a proper mechanism that already exists, you should. And if it just has to be a little bigger, then you should. 

But where it meets one of those tests, it requires complex coordination or it requires something too big for any one group to do, or things like that, those are the kind of things, and something that was a foundational research project that fit in there and that might lift 1,000 ships, that would be great.

So, I think about the ZIP code addressing. I imagine there’s a lot of interesting basic research that has to go on to do that. Where are you going to get those ZIP codes from? Well, I’m joking around about ZIP codes, but I don’t know that there’s going to be a unique protein on the surface of every single cell type that marks that uniquely. 

So, it might be that you need to manage to get something in by virtue of three markers, distinctly marking that cell type. What does that mean? Am I using three separate routes of delivery that reassemble in the cell type? 

Well, then I just got to demonstrate that proof of principle, and that’s a really amazing foundational project. Or am I going to somehow manage to detect the coincidence of those markers on the cell surface?

Or am I going to put something in, and it’s only going to get activated because, well, it got in because there was something on the surface, but it gets activated because there’s a particular program running in that cell.

And so I think there are things that are foundational that are use-inspired that may well fit. But the one thing that I’m absolutely confident of is all the examples that we’re making up, and putting on slides, and Francis is seeing, will be four-fold less interesting than the actual examples when the thing gets going. 

So, we’re kind of pulling stuff out of the air, as these are plausible examples, but once you get people thinking this way, I think the next generation of researchers are going to come along and refactor and reinterpret this in really creative ways, and that’s what excites me.

Collins: I’m totally with you on that. Yes, absolutely. And I think also that we do have other components that can fit in spaces that aren’t quite right for ARPA-H, but aren’t necessarily just right for an R01. Think of the [NIH] Common Fund, for instance, in that regard.

Barbara Wold [professor of molecular biology, Caltech]: I’m interested in scale. The scale of some of these problems, and the ability to go large is really exciting, totally seductive, and optimistic. On the other hand, the scale factor for how you run a DARPA, which is, how many actual program officers do you have, it is very flat. 

And so the question is: is the scale right, at the way that it’s pitched right now? Might it even be a little too big out of the gate to build a culture of that flat sort?

Lander: Well, I think the ramp is a very important question, which is why it’s, I think, very wise that money was put out as three-year money, because I think trying to build it as a step function, operating on day one at scale, is not a good idea, it’s not the way to build culture. On the other hand, doing it at such a low level that you can’t take on ambitious things isn’t good. 

So, amongst the many open design questions, and I think it’s important that there’s a robust discussion in Congress, in the biomedical community, amongst those are—what’s the right ramp to get this up? There’s a learning curve to figure out how to translate the concept to biomedicine.

Wold: Right, and you only get to start once, so if the culture veers badly at the beginning, changing it’s really, really hard. Of course, I’m not telling you anything you don’t know. 

I did, though, just have observed at close range a really terrific experience with the relatively new non-governmental version of this, which is Wellcome Leap. 

And I’m sure that you’re probably chatting, but following what Mark and others have said, the idea that there are some things that really are intrinsically global and that doing the dance is probably really going to be worth it. Wellcome has been in your dance before.

Lander: And Wellcome Leap is run by a former DARPA director.

Wold: I am well aware, and [Wellcome Leap CEO] Regina [Dugan] is exactly the kind of talent that one wants, which is very rare.

Lander: And there is a great set of conversations amongst the many groups that are thinking this way, including Wellcome Leap and others, so I think there will be many good conversations about how to build these kind of cultures. It is the right time for this.

David Glazer [engineering director, Verily]: Actually, following up on Barbara’s question, I had the chance to spend a very little bit of time with Regina when she was at Google in between DARPA and Wellcome Leap, and one of the things that I learned from her that you mentioned in your presentation—but it surprised me as to how central it was—was the timeboxing. 

And that idea of timeboxing was not just an operational thing, which is how I initially thought of it, but it was a cultural foundational thing because of how … you’re getting it, so you’ve heard all this. 

What I’m curious about is: how much have you thought about how that should be applied to ARPA-H? Do you think you can take exactly the rigid and set time lengths that have worked in the other ARPA-ish things, or do you think they will be different in bio, or TBD?

Lander: Having timeboxes is important, but one size fits all will work less well in biomedicine than in a pure engineering thing, because there are some things that are clinical experiment, testing things in humans. 

You’re not going to get that done in the same period you’re going to get certain other things done. So, everything should have a box, but it should be the right-sized box for that challenge.

Glazer: Yes. And then I guess my follow-up would be, that sounds right to me, and I expect that will have surprising cultural implications also, because of how central it was when it was one size fits all.

Collins: Yes. I imagine you’re right. Kristina, you apparently have some experience that might be relevant here.

Kristina Johnson [chancellor, The State University of New York]: Well Dr. Collins, thank you, and thank you, Eric. When I was under secretary in the first Obama administration under Secretary [Steven] Chu, we had the opportunity to set up ARPA-E. 

And many of the things that Barbara, you just mentioned, David you mentioned, are exactly right on.

And so, the first thing we had to do is sort of define, okay, how does this differ from the Frontier Centers and the Innovation Regional Hubs, which were being envisioned at that time, and they’re having a comeback, fortunately. And then the fundamental work.

And I think that one of the basic things is we learned that we pretty much had to move it out of the [James V.] Forrestal [Building] and up the street a little bit. And Barbara, what you said is right on. It’s that first director that we hired and recruited who then established the culture, and from there, the program manager’s just really high-risk and big-impact programs.

So, anyway, there’s anything that we can provide you from our experience at ARPA-E, I’d be very happy to do that. But very exciting, and see whatever we can do to help you succeed.

Lander: Thank you for that offer, which we are going to take you up on with the greatest enthusiasm. Thank you.

M. Roy Wilson [president, Wayne State University]: Thank you, and thank you Dr. Lander. It seems to me that this group is very supportive and even excited about this, and I tend to fall into that camp also. But I have been part of other discussions with other research circles, and there is skepticism out there by some people, and some people with very big voices in research.

And I’m just wondering if you have any plans on trying to move these skeptics, or just moving forward without worrying about it?

Lander: Oh, no. I think the purpose of coming to ACD and having other discussions is because this merits discussion. I know there’s some skepticism around: will this take away funds from the poor biomedical researcher?

Wilson: Right.

Lander: I would be deeply opposed to that. I think that would be a terrible thing. But then again, I remember when the Genome Project was viewed as a threat to take away funds from biomedical research and everybody was very concerned if we did that, it would contract. 

The Genome Project made the rest of biomedical research, I don’t know, twice as productive, or something like that. So, one wants to take account of that, and make sure that there really is no harm done in addition to the core ecosystem that we talked about that’s the foundation for every bit of progress we’ve had. That’s a skepticism that I think we have to address by making sure that doesn’t happen.

Then there are, I think, lots of really important questions—you’re asking them too and others are asking—how do we design it so it really succeeds? And if one is not skeptical about all the ways in which something might not succeed, and being analytical, be prepared to name those things, like culture is hard to change. 

Well, maybe it should be at some distance. It’s got to be protected in some way. DARPA sounds like good lessons, but do they really fit?

I’m sort of a real skeptic about everything I hear, and I think that’s what’s great about this community is, the only way it’s going to work is if we all go in skeptical so we’re going to find the problems in advance, and design it to succeed. So, I don’t think any of this should proceed over objections. I think it should embrace all the worries, because it will make it better. 

When Francis and Larry said could we spark a bit of a conversation here, this is meant to be a spark and initial conversation, that I tried to say at the beginning and I’ll repeat right here at the end, this is looking for input from a community because all these projects get better when they go that route.

Wilson: Thank you.

Collins: And this is one of what we hope will be many discussions with the scientific community, and Tara is working hard on designing some of those stakeholder gatherings so that there’s more information. I think there’s still a lot of people going like, what is this? 

And we need to get that information in front of them that’s accurate. There are various people telling you what it is who don’t actually know. Be careful of those. And we will have other citable objects very soon that I think will help with the actual reality check of what we’re trying to do here.

So, Eric and Tara, thank you so much for coming to join us here at the end of our two days of intense discussions. This was a lot of fun, and I do hear a lot of excitement from this group, as well as some skepticism and I think that’s where you are, and that’s where I am, that’s where Larry is—and as you say, that’s what will make this even better. 

Nobody’s sold yet that we have it exactly right, but we have a pretty interesting shape, and now it needs some of those details to be filled in, and that’s what we’re going to be doing.