This transcript has been edited for clarity.
Eric J. Topol, MD: Hello. This is Eric Topol. I'm editor-in-chief of Medscape, and it is a special privilege for me to bring Manu Prakash to the Medscape community. Manu is a computer science innovator of our era. Manu, welcome.
Manu Prakash, PhD: Thank you for having me.
Early Invention and Education
Topol: For a bit of background, am I right that you made your first microscope when you were around age 7?
Prakash: Yes, 7 or 8. It's old memories, but I remember not having access to a microscope and that bothered me. My brother was upset because I didn't understand that you needed specific lenses—or any lens—for a microscope. The lens [I used] happened to be his from his eyeglasses. I still remember that time.
Topol: You graduated from the Indian Institute of Technology in computer science and went on to Massachusetts Institute of Technology (MIT). There is a funny story from when you were at MIT. They would not let you graduate at first because you had so many library books that were overdue.
Prakash: Yes. Now it's funny, but at the time it was a little bit scary because I had several thousand dollars in fines. I actually met my wife from this because she was the one who gave me a loan for the fines.
Topol: I think that was a good investment on her part.
Prakash: Yes, it worked out.
Foldscope
Topol: Then you moved on to Stanford where you are a biophysicist in bioengineering. You have a great lab and have been remarkably productive. It's actually very hard to keep up with you on your inventions. The first one I thought we would get to is the Foldscope. Can you tell us about that? About 5 years ago it was published in PLOS One. I think it's the most cited paper in the history of PLOS One, maybe PLOS completely, and it's changed the world. Tell us about Foldscope, and are you still "Foldscoping"?
Prakash: I'm so glad you use "Foldscoping" as a verb. One of our journeys is to make microscopy so accessible that you would [talk about it at] the dinner table. It started as the idea that everybody deserves to have access to curiosity and access to the microscopic world. When you talk about medicine, it's very hard to talk about these complex ideas, like germ theory, if you have never, ever seen them through the microscopic world.
Foldscope is literally an origami microscope—you fold it together with a piece of paper. I did this work with my very first graduate student, Jim Cybulski. Sometimes the first graduate students are very adventurous.
The paper was really about the technical capabilities of bringing cost and performance together, where we could demonstrate 700-nm resolution imaging at a price point of $1.00. We asked ourselves, if we can do this and write a paper about it, what would it take to scale? And in the 5 years since publication of that paper, I'm happy to tell you that just last month we crossed 1 million Foldscopes.
There are 1 million kids and people and doctors and researchers and veterinarians around the world who are exploring the world and documenting and collecting data. It's spread across 140-145 countries. It's been a real journey; it's been a social experiment.
Topol: Originally, that was the $1.00 version. Then you made a fancy version, too?
Prakash: Yes. The way we made the project sustainable is that any kid in the world can have access to it for $1.75 (so, still less than a cup of coffee) and it's the same price anywhere in the world. An advanced version has many other accessories, like cell sorters and a lot of microfluidics and other things that support the deployment for the basic Foldscopes. It's kind of a TOMS shoe model, but in the end it's incredibly important because much of what we have done with this has really been supported by everyday people.
We structured this to ensure that we can bring sustainability. There are distributors of Foldscopes around the world, like in Iraq, Syria, India, and Philippines—places that you wouldn't think would have access to these tools. People in communities literally teach and train others on local problems; it's pretty much driven by the passion of people who have passion for science and deeply care about sharing it.
We're a very small group. Even today, having made a million Foldscopes, there is only a team of three or four people at the core of it. So, yes, it's been a really remarkable journey.
Topol: It's not just that you did the innovation of making a microscope for less than $1.00 in parts, but also the fact that you democratized it. Initially, you were getting it out to 50,000 people in the developing world, and now a million. That is amazing. When The New Yorker wrote about you back in 2015, they said that in your house, the kitchen was a lab, the dining room table was a lab, the bathroom was a lab, everything was a lab. Is that still the case?
Prakash: I didn't have kids at that time, but now I have two young kids running around and they have their own ways of doing experiments. I think that resonates with me a lot. It's true that you can just look in your kitchen. The first time you see bacteria is such a profound moment. You could have studied as many genetic networks as you want, but [seeing bacteria] for the first time in your own arena makes you say, "Wait a second—I am surrounded by it." This is me, in some sense.
Scientific tools are like pencils: They need to be accessible and ready whenever we want them.
I deeply value that experience. I often say that I created Foldscope not just for other people, but also for myself. And it happens to be useful for others. We have discovered several remarkable cellular phenomena by just poking around with Foldscope. Last month, we described a new kind of intracellular communication in Nature —that discovery was made first with a Foldscope. I was in a pond in a swamp and I saw a cell with contractile capabilities, and it was just obvious from the first moment we saw it that it didn't make sense, so we had to study it. We spent several years studying that phenomenon, but I would have never stumbled upon that phenomenon [without the Foldscope]. That is a common story that happens with all of the Foldscope users out there. It's just like how we carry pencils; when you have a thought, you want to write it down. Scientific tools are like pencils: They need to be accessible and ready whenever we want them. You have to lower the barrier. Sometimes you tell yourself, "Oh, that is a stupid idea. I'm not going to try it." But it really is worth trying.
Topol: Yes. Have you even hit age 40 yet?
Prakash: I'm about to; I'm 39.
Topol: With what you have done by age 40, I look at you as kind of the current version of Antonie van Leeuwenhoek. It's like you're 400 years old or something, and it's just wild.
It wasn't just about the innovation or the democratization, but about trying to enlighten the world about the microcosmos and biomimicry. You have a quote about the microcosmos: "Plants, insects, tiny bugs under the sink, bacteria, day after day, accomplish things that no scientist anywhere in the world knows how to do." [The Foldscope] is a great contribution worldwide, and it's going to continue to grow, undoubtedly.
Paperfuge
Topol: Let me go on to your next big invention, the Paperfuge, which is the 20-cent centrifuge that goes 125,000 revolutions per minute. Can you tell us about that?
Prakash: Yes. I think Paperfuge is in the same space as Foldscope, where I have been asking the same question over and over again: How do cost and performance couple? In our day and age, we have this intuition to say that if you throw more money at a problem, you can solve it in a better way. That does not have to be true, because we bring an intrinsic bias.
Paperfuge started when we were implementing and doing clinical trials of schistosomiasis with Foldscope. I was in these little remote clinics in Uganda, seeing the need for both sample processing and sample analysis before imaging. Simultaneously, I had a moment where I saw a centrifuge being used as a doorstop because this place had no electricity. That was a very aha moment to say, "We need a powerful scientific tool, but it cannot depend on electricity. It has to be unplugged for it to reach people." And on the flight coming back, I made a list of all of these toys that spin—I am personally fascinated by toys. We analyzed many of them. For months, we did a calculation to show how good yo-yos are in transferring energy. They are very good, but they are not so good because it requires a certain skill to make a really good yo-yo throw. And while playing around with that process, we stumbled upon this toy which is called a button-on-a-string. Anybody can do this: You take a button, you put two strings in, you pull them back and forth, the disk spins.
When we started doing this, we realized that nobody actually understood how this simple toy works, and it happens to be the oldest recorded toy in the history of mankind, which I find remarkable. When we looked at the archeological data, 5000 years ago people had relics of almost the same object. Many cultures have discovered it again and again but have never understood it. That number of revolutions you quoted, 125,000, is powerful because we first arrived at that mathematically. It's the same story as with Foldscope: These types of problems are difficult, but you can tackle them with an engineering mindset and optimize them. In our current work, we are trying to exceed that by an order of magnitude. We want to break the 1 million barrier. We want to go faster because that would allow us to pull out more and more precise biological parasites in shorter and shorter time. Right now, pulling out a malaria parasite takes 5 minutes; we want to bring that down to 1 minute.
Octopi
Topol: That gets us to the most recent breakthrough, Octopi. You had a preprint a couple months ago and I'm sure it will get published soon. Can you tell everybody about Octopi?
Prakash: Yes. Octopi stands for Open Configurable High-Throughput Imaging Platform. It's also called Octopi because my kids really love octopuses right now. It is important that scientific tools cost less, but you still need high throughput. When we are in clinics in the field where I'm traveling, I can see a line of 100-150 patients waiting to get diagnosed for malaria; 10% of them might be positive, and in some places, even more. But the traditional gold-standard microscopy mode still requires between half-an-hour to an hour per slide, and rapid diagnostic tests don't actually match the specificity and the sensitivity.
As a lab technician, you are in this conflict where you want to do a good job on every patient, but there are not enough hours. I have met many of these people who work nonstop for 10 hours a day on a microscope, and it's just not feasible. With eyestrain associated with 10 hours of microscopy, we'll never be able to provide the kind of sensitivity [we need].
Octopi is essentially a mash-up between flow cytometry and microscopy. We built an instrument that, depending on the parts, costs between $100 and $250, but it has the capacity to image 5 million red blood cells per minute. For a single patient, we can image every single blood cell in a smear within 2-3 minutes.
We made a discovery that I'm very proud of and surprised that nobody else had noticed: Octopi does not just do microscopy; it does spectroscopy. When you look at a star and all the spectra of light coming in, you can tell the chemistry of the star. We applied the same idea to parasites. We realized, in a frugal manner, that by not choosing very tight bandpass filters or by not choosing any excitation filters at all and using a color camera, we could see a color change in the parasite compared to human cells with the dyes we were using. This is an accidental discovery I made in the lab with my graduate student, Hongquan Li. We have pinned down what is actually happening. These dyes that bind to RNA and DNA have different conformations, and it's the ratio of DNA to RNA that starts giving us the spectral shift. Malaria is usually diagnosed with 100x or 60x objective; we are doing the same thing at a 10x objective. This gives you a massive field of view, but by using spectral channels, we can identify it without the resolution.
That is a big win in terms of the size of the tool, the cost of the tool, and the time it takes, and now we have the capacity to start applying these spectroscopic ideas to even better dyes that will be able to tell other states of the cell, for example. I'm very excited about this tool. We have developed modules. Because it's modular, we have a malaria module and a tuberculosis (TB) module, but anybody in the world can make a module. If you care about a disease and you identified the speed you need to do this at and the cost, you can build the imaging platforms on top of it. The hope is that people share them. These are not black boxes, which is something that really bothers me about diagnostics. We have this notion that black boxes get fit, and then the lab technicians and the doctors go further and further away from the processes.
I was having a conversation with a physician and he said, "They took my microscope away." They did not want the doctor to have a microscope in their office because of liability reasons or something. Now you have to write a $500 test that's going to go to the lab tech and then, of course, the hospital makes whatever number of dollars they make. We can discover many things, but the way we perceive these sets of technologies and how we deploy them in the context of the need is important. If it's malaria, we have to think about the places where malaria hides, and if it's TB, we have to think about the places where TB hides. This is why I'm very excited about implementation, because Foldscope has taught us a lesson that you can't just make discoveries; you have to take the next step.
Artificial Intelligence
Topol: You have taken a lot more than the next step; you've taken multiple steps. One of the things that you have done with Octopi is add in artificial intelligence (AI) and neural networks. Some people have said that someday the microscope will be truly historical. You are already getting rid of the eyepiece. Can you give us your view about the dimension of AI and how that plays into all of this?
Prakash: I have a very grounded view about solving problems now, and of course, the future in 5 years is going to be very hard to predict because of the computation capacity. But from a perspective of now, very specifically, the AI platform is embedded right in the hardware in the machine itself. We don't depend on the Cloud, primarily because we cannot expect to have the Cloud—we can't even expect electricity. We can't upload data somewhere and have to have something that works in real-time. That has been very powerful. We are using hardware that was built and designed for real-time decision-making in automated cars, because the computers that will be embedded in automated cars are being produced at large scale now, so we have been able to utilize these graphics processors.
Technology is a blend of humans that provide the service with the technical capabilities. The features we have built allow the microscope to sort out all of the parasites, but it also still allows you as a user to screen through and see a few parasites yourself because you are learning with the microscope—and sometimes the microscope is learning with you.
One of the big things I am excited about that is not yet public is that we will be announcing the 100 Octopi Program. One of the challenges with AI is what data you train with. Real-time data coming from the field, from the context, is always a challenge because many times the hardware has switched, so you trained on one hardware and then suddenly it's a different hardware.
We are planning to deploy 100 Octopi at 100 clinical sites. For users to sign up, they have to agree on sharing that data real-time—not patient data but raw, de-identified imaging data. That allows people who are developing new algorithms to truly understand the complexity of diseases and many forms of disease that show up, such as with coinfections, when you have field commissions where the dyes are not stained perfectly, and when there are temperature effects.
On one hand, the goal for this is to build better algorithms, but on the other hand, we will be able to test the operation. I think the inspiration for this really comes from robotic telescopes. You can go online and log in to hundreds of telescopes that are operating currently on our planet, looking at the sky, but an analogy to that has never been created. Because Octopi is autonomous, the goal is for researchers, communities, nongovernmental organizations, healthcare centers to truly test the tool at scale but then have connectivity. A person in Kenya is struggling with the same sets of problems as someone in India; they are dealing with the same disease but there is often very little communication even though they have the same tools.
It's much more about training this community itself because there often can be a fear of technology. People wonder, "Is this going to take our job away?" or "What if I don't trust it?" Trust can only be built slowly with a network, and because it's a modifiable object, we want people to play with it to get a sense for what these new tools would look like. I kind of struggle with this because many larger-scale organizations have opinions that are made based on their biases, but nobody is willing to do medium-scale tests to really understand how people would react. We just got back from Uganda and we had these fascinating conversations about the lab tech looking at this machine saying, "Is this going to take my job away in 5 years?" I've been building and designing these tools, and there is this very deft conversation to have, that human conditions can be improved and human services can be provided by these incredible people with empathy and care. So the answer is, no, this is not about replacing people, but it's about making their lives easier because currently they don't have the capacity to deliver care at a level we would actually accept.
Implementation Challenges
Topol: I couldn't agree with you more about enriching people's lives with this type of technology so that they can be more human.
The triad of making something better, faster, cheaper—you've done that time after time. You rewrote the book on this. What are you going to do next?
Prakash: I try not to think about that because there is a lot to tackle. But I love implementation. I have these debates and arguments with many of my academic friends where, since the paper is written, let's move to the next thing now because some company is going to come and solve these problems. Many times in that transition, the ultimate goal for why the technology was invented is already lost. The mission is lost; the vision is lost.
It's very important to us that these tools reach the people that we designed them for to begin with, and that road is not easy.
To me, the blood and sweat in everything we do is the pain associated with implementing. We are about to start the Centrifuge Program, which couples with Foldscope, so it enables Foldscope to do more and then, again, similarly with the smaller-scale trial with Octopi. I get emails every day from Pakistan and Fiji and remote parts of the United States, talking about engaging in these communities.
As an inventor, that is enriching and what drives me. It's very important to us that these tools reach the people that we designed them for to begin with, and that road is not easy. I struggle with funding. People say they are willing to fund the next innovative idea but [not necessarily] when testing it at scale, where we can learn so much more about the disease. I've just been starting to think quite a lot about ecology as well, and this notion of planetary health and the health of our planet, which is human; but not all of it is human, and human health depends on its health.
Foldscope is used in agriculture and veterinary care, but also for environmental measurements, and I often think that the planet is such a large place that if you don't include communities, we will never be able to measure what's happening on our planet—be it diseases in humans, be it diseases in animals, be it loss of biodiversity. The next big challenge is to integrate many of these tools, bring more numbers of tools to have a societal revolution in measurement. I want everybody to take part in it. We have the capacity to do this and we have access because we live on the planet. So wherever you live, you have a unique context, but we don't yet have a common mission.
It's a much larger, longer-term challenge, but I am inspired by what has happened in the past. A community of people discovered the first Sputnik. When Sputnik was launched, nobody knew where it was, and one of the amateur astronomers across the world looking at the sky spotted it. Only then did the military infrastructure catch on to it. It inspired a generation of people who think about space because they were engaged.
I want society to move in this way, where consumption of knowledge and information is not enough for us to be informed citizens. We have to take next steps. These are big goals.
Topol: Right. If anybody could inspire the 7-some billion people on the planet, it would be you, Manu. You are a force. How many people are in the lab with you at Stanford?
Prakash: I have around 20 people in the lab, but I think that is because of much of the very basic biology stuff that we do. We work on some very basal animals that have been forgotten, which might be the root of all animals.
I don't do this alone, and when we engage people, we pass the ownership to local communities. Literally, while we are talking, a group in Iraq is running workshops in the Kirkuk area with Foldscopes, teaching and training people, getting samples back, and probably doing the very first biodiversity surveys in that area. This was completely driven by a single individual who felt the need, who associated with the mission, and then went on and actually engaged.
Much of what we have been able to achieve, especially with a tool like Foldscope, has been because of individuals we are empowering who are deeply passionate about science as much as I am. It's very hard to discover them because they are not sitting in universities, they are not professors. Many of them are like Mo [Pandirajan], who is one of the most prolific Foldscope users in the world. He was running a 12-kid school in a village in South India, and in the past 4 years, he has trained 97,000 kids in India alone. He's the most prolific contributor, and after English, Tamil is the most spoken language on our site because of one individual, because he inspired so many people. How do we find people like Mo in the world if we are not willing to give and share? Papers don't do that. You cannot discover remarkable people. Some days I'm optimistic, but other days I have my challenges.
Topol: I can't imagine you on a day that you're not optimistic. I can't thank you enough for joining us. The people who are listening and reading this transcript will have links to your incredible papers and profiles in The New Yorker and The Atlantic. We are going to be following you with tremendous interest because you are changing the world, and not too many people can say that. It's something to be proud of, and you're still young. Just think about the next 40, 50, 60 years. Watch out.
Prakash: You are very kind, Eric. Thank you so much. I enjoy what I do and I'm excited about engaging people, so our doors are open.
Topol: Keep up the terrific stuff. Thanks a lot, Manu. And thanks, everyone, for joining.
Eric J. Topol, MD , is one of the top 10 most cited researchers in medicine and frequently writes about technology in healthcare, including in his latest book, Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again.
Follow Medscape on Facebook, Twitter, Instagram, and YouTube
