AMP 2018: An Interview with Dr. Jonathan Rothberg
Dr. Jonathan Rothberg is a man who needs little introduction. A serial entrepreneur, he has pioneered and then revolutionised next-generation genetic sequencing, as well as creating a world-first handheld medical imaging device retailing for just $2,000.
First awarded the National Medal of Technology by President Obama in 2013, Dr. Rothberg is now receiving the Association of Molecular Pathology’s (AMP) Award for Excellence in Molecular Diagnostics Services. We spoke to him about his many achievements, his proudest moments, and the future of his fascinating field.
FLG: You founded your first genomics company, CuraGen, in your basement in the ‘90s, and since then you’ve built up a tremendous career. Now, you’re being presented AMP’s Award for Excellence in Molecular Diagnostics. What was it that first drew you to science and sparked that interest?
JR: It goes back to growing up. I was lucky: my dad had a chemistry lab in our basement, so from grade school onwards, I was able to do chemistry and electronics at home. My mom would encourage me to pursue science and my dad would tell me about the power of engineering, so I was exposed to it all right from childhood.
Probably the most formative thing in my life was my older brother Henry buying me a personal computer, a TRS-80. At the time, it was considered outrageous – my parents couldn’t believe he spent $900 on a computer for his younger brothers!
So it was a collection of my older brother and the computer, my dad and the chemistry lab in the basement, and my mom encouraging me to do good science.
FLG: You’ve been involved in some of the biggest genomics companies of the last two decades – CuraGen, 454 Life Sciences, Ion Torrent, etc. Through these experiences, what do you think are the most important lessons you’ve learned about operating within a large, successful biotech?
JR: I would say the most important lesson wasn’t how to operate, but it was who to surround yourself with. Any time that I’ve been successful, it’s because I hired somebody who knew better than I did at something and who was able to tell me where I was wrong, and I was able to accept their solution.
So, you need to hire people who are better than you at something, but they also need to be people who can work as a team and accept somebody else’s idea. You can hire smart people, but they must be willing to look around the room, have a discussion, and go with somebody else’s idea. That’s the most important thing. You can talk about laying out a project plan, or program managers, but at the end of the day, you’re going to be moving fast, discovering things, and changing direction, and so you need people who don’t need the details and who are willing to work with other people’s ideas.
If I look back at each company I was involved with, there was always a moment where we had a problem that nobody else in the world could solve. For example, with next-generation sequencing, no one knew how to prepare a sample, or how to sequence in a massively parallel way, and those were questions we wanted to answer. You need to have people who can solve those problems. I’m pretty sure that Steve Jobs was successful because before he was 21, he met Steve Wozniak. Steve Wozniak used to redesign computers that had originally been designed by over a thousand engineers, but as an individual, he could redesign them with half the parts. So, Steve Jobs knew that extraordinary people like that existed.
All of my success comes from getting people like that and having teams that are willing to go with someone else’s idea.
FLG: A group effort then?
JR: Yes, it’s a group effort, but it’s based on extraordinary individuals who are willing to take turns running with somebody else’s idea. If you have five smart people who meet up and agree on something, but then all go their own ways, it doesn’t work. You have to have people who are smart enough to know when somebody else in the room has a better idea.
We actually have a rule in our meetings that you’re not allowed to say that somebody else’s idea is bad. You’re only allowed to say that you have a simpler idea that can be done faster, or scale more quickly. That’s key. And then, once you leave that room, you have to be willing to go with somebody else’s idea if you’ve agreed on it.
It’s based on extraordinary individuals, but you take turns being the extraordinary individual.
FLG: What are you most proud of from your time in the industry?
JR: I’m most proud of what I call having forked history – that is, having made history go down a different path. When we invented next-generation sequencing, we really changed the way that science was done. Once we had a way to sequence DNA that was thousands or hundreds of thousands of times faster and cheaper, then we were able to do good science. I was able to cold-call Svante Pääbo and sequence the Neanderthal genome, for example. Or being able to do the first precision medicine diagnostics – we had a biopsy sample from a patient and we did the world’s first deep sequencing, which ultimately changed the medicine that the patient was being given.
So, that’s what I’m most proud of: making a technology that became an industry, and then the new science that people were able to do with it, either science that we kicked off like the ancient DNA sequencing, or science that other people did because they had a tool like this.
In terms of technical details, two ideas that we came up with for 454 have been the basis of the whole industry. The first was the idea to sequence without bacterial cloning, and now that’s how all sample prep is done. The second was that instead of Sanger sequencing, where you have a piece of DNA that you chop up to determine what it’s made of, we were going to do sequencing by synthesis. That’s the basis of all next-gen. So, at a lower level within next-generation sequencing, I’m proud of that.
And then it’s the science that we’ve done, like the Neanderthal Genome Project. There was another project we did once where there was a transplant patient and everybody who received organs from that one patient died. So, we sequenced the organs to find out what was happening. It turned out that there was a virus killing the patients: because they’d received a transplant, they were being immunosuppressed and it meant that something that would normally have been benign was killing them.
I do think that next-generation sequencing has changed the course of history.
FLG: You were awarded the National Medal of Technology by President Obama in 2013. What was that like?
JR: It was fantastic. There’s a funny story there too: I got a call to tell me that I’d won the National Medal of Technology, which is personally bestowed by the President, and that I could bring two of my family members with me to the White House. Now, I am one of seven children, I have five children, and at the time, my mom was alive. I told the handler that I couldn’t pick between my children and she said okay, so instead of two guests, I had myself and five kids.
Then, there was my mom and it was super important that she come. She’s my biggest supporter! So, I was up to seven people. Then, I told them that I couldn’t pick between my siblings and they said, “Oh, of course, you can bring your brother or sister,” and I said, “Well, I’ve got six of them!”
Fourteen of us showed up at the White House. True story!
It was so great. You see President Obama and he fell in love with my two youngest kids. He told me that I have the most beautiful children and then, even though there are a lot of guests at the White House, my two youngest kids got a personal present from the President. He sent over a goodie bag with signed M&Ms and other candy and immediately, my kids ripped it all open to eat it.
He was super gracious; it was a really incredible experience. President Obama put a medal around my neck, my mom gets to watch it, and I don’t think there’s anything better, but there’s a little twist to it. I got that medal exactly ten years after I got fired for inventing next-generation sequencing.
When I was at 454 and we developed next-gen, we had started an industry, but the board didn’t think that the hype around sequencing was going to continue. The Human Genome Project had been done and they thought that the fanfare was going to die down, so they sold 454. We’d sold $100M worth of equipment and we had a monopoly, but they sold it for what I think was 20 cents on the dollar to Hoffmann-La Roche. I woke up out of a job, and I had to start over again. Then, ten years later, I get a medal from Obama for the exact same thing that I got fired for! I was the same person and it was the same thing, but instead of getting fired, I got a medal.
FLG: What do you think has been the biggest or most important development in Molecular Diagnostics in the last 20 years?
JR: It’s a couple of things. It’s the ability to sequence large amounts of DNA, and it’s the ability to analyse that data. You needed both for next-generation sequencing. When you do that, you can close the loop and make those correlations.
We’re just getting started there. There’s a revolution in AI and it’s real this time, around deep learning, and in combination with DNA sequencing, we can really start understanding complex genetic disorders for diagnostics and therapies. It’s that combination of our ability to collect lots of data with our ability to analyse it effectively.
FLG: You’re now working as the founder and CEO of the Butterfly Network. How did Butterfly come about?
JR: All my genesis stories are identical – I do these things because of somebody that I love. In the case of next-generation sequencing, my son wasn’t breathing, and I wanted to understand why he wasn’t. I saw a magazine that talked about a computer chip, and I said, “Hey, I should miniaturise the ability to sequence and put it on a chip.” That then became 454 and next-generation sequencing. I didn’t want to know what we all had in common, which was the goal of the Human Genome Project. I wanted to know why my son Noah wasn’t breathing, and so I came up with next-generation sequencing.
For Butterfly, my daughter has Tuberous Sclerosis. It’s a genetic disorder and she has growths in her kidneys that I wanted to be able to monitor and non-invasively stop. I’d been sponsoring clinical trials at centres that used MR to image and then used high-intensity ultrasound to treat the condition, but they’re $10m centres. It’s too expensive and it’s too slow. You go in and they take an image every eight seconds, then try to direct energy at the right sites, but it wasn’t precise. It’s as if you sent something to your printer, but it takes two weeks, the printer costs $1m, and then it comes out looking like your kids drew it with crayons.
So, the idea with Butterfly was that if I’d just put DNA sequencing on a chip, I figured I could put imaging on a chip too. Right now, Butterfly has launched the world’s first hand-held, whole-body scanner with the broadest FDA approval ever for this type of device. It’s just an imaging device, so it can’t treat conditions like Tuberous Sclerosis, but to me, that’s a stepping stone. We’ll have to do more clinical studies, of course, but with a bit more energy, you’ll be able to not only see things, but be able to do non-invasive surgery.
I figure that any time you put something on a chip, it changes history. When they put computing on a chip with microprocessors, it gave us modern computing. When they put a camera on a chip, it put Kodak out of business and a camera on the cell phone I’m talking to you on. When we put DNA sequencing on a chip, we changed the world with next-generation sequencing. So, I decided to put imaging on a chip.
FLG: How do you think Butterfly will change the way in which people gather and control their own healthcare data?
JR: It’s pretty simple. Around the world right now, there are 40 million people who are part of the healthcare system and only 1% of them have access to the ability to image. Our goal is to give imaging to the other 99%: an ultrasound cart costs, on average, $110,000, but you can go to our website and buy whole body scanners for $2,000.
And if you give imaging to more people, it’s more than just making it lower cost – you have to make it easier. So, we have a low-cost, $2,000 ultrasound probe, but we will also have artificial intelligence that allows anybody to use it. My son, within ten minutes of picking up our probe, was able to look at his cousin’s heart and measure his ejection fraction, which is something that you can bill for $1,000 in the United States.
Our goal is simply to democratise. If you look at everything that I’ve done, that’s really it. Sanger and Gilbert won a Nobel Prize for DNA sequencing in 1980: people were able to sequence already, I just made it thousands and then millions of times cheaper. People already knew that they wanted to sequence, I just democratised it.
The demand is important. For Butterfly, we knew that people wanted to image. People have had ultrasounds for 50 years, they’ve just been built on hardwired crystals where any time you want to image something, you have to have a probe that matches it. Ultrasound carts will have three to six probes that you have to plug and unplug, and vibrate at different frequencies… What we did is just put all that on a chip. Just like a microprocessor, just like a camera, just like DNA sequencing. But the important thing is that I didn’t have to say, “Hey, it’s a good idea to image.” People already wanted to do this. When we turned on our website, we had 40,000 orders because the demand was already there, I was just making it 50 times cheaper. If you make something cheaper that people already want, it expands the market greatly.
People managing their own data is something we think about a lot. I told you that our machines come with AI. The AI can, for example, measure the ejection fraction of your heart, and we’ve trained it by looking at hearts and having experts annotate the data that we feed into it. Then, every time our machine is used and somebody corrects the data, it gets better.
But I call that a short loop. The AI will look at the heart and then get better at measuring the heart. If you have a nodule in your thyroid, then it can look at your thyroid and you can better teach the machine to identify anomalies in the thyroid. I think the excitement comes over the next five years when all of us have our medical records on our phones because then, when you use a Butterfly device, you can track what’s happened to that patient’s thyroid over time. The thyroid may have been biopsied, or sequenced, and they may have noticed, for example, that the BRAF gene has a mutation that made it cancerous. Now, you can go back and train the analysis software to look at the images and see if there are differences between the images of thyroid nodules with and without BRAF mutations. I call that the long loop and it’s what I’m most excited about.
When we all have control of our own data, we can share that data back to people who are collecting specific information to make diagnostics better and better. We’ve seen this happen with sequencing, where people are sequenced to find out what disease they have and how they respond to medicines, and I think it’s only going to happen more over the next five to ten years. People are going to be handling their own data on their phones, and it’s going to be really easy to opt in to data sharing when they have more data collected, which is going to enable these long loops and make these things much more powerful.
I think it’s going to happen because both Apple and Google, who control the two major phone operating systems, have applications for organising health records. As people go to their doctor, they’re going to be able to ask for their health records back on their phone. Individuals are going to drive it. That’s the real revolution. Once it’s on the cell phone and they go to a doctor, or pick up a Butterfly device, or have their sequence done, this will happen naturally. I think that we all have to opt in and build those medical records, and then be willing to share those records when they have their next diagnostic or yearly physical. That’ll create the long loop.
I’m confident that it’s just going to happen because people are going to want their medical records with them, and the only way to carry things with you, at least for me, is to have it on my cell phone.
FLG: How do you think that we can encourage data sharing the way that we would need for that type of system to work?
JR: I know what I want to do at each of my companies. I have an incubator called 4Catalyzer, which has four companies, the most mature of which is Butterfly. Each of them is creating a device that nobody has done before, and each of them wants to implement this virtuous circle where data is collected in the cloud, the machine analyses it, and it’s combined with either short loop information or longer loop information.
It will be driven by the companies that have the devices people want. Forty million people want a Butterfly device and every doctor or consumer that has that device could have a dialogue box pop up that says, “Hey, please share that medical record you’ve just collected with the patient.”
A lot of this is going to be driven right to the patient. The first next-generation sequencing machines were half a million dollars and went to some big labs. The next ones were $50,000 and could go into small clinical settings. In the next five years, you’re going to have a sequencing machine by your toothbrush and you’ll be able to say, “Yeah, I’ll participate in a study.”
It’s funny, because when I tell people this, they say that it doesn’t sound realistic, but my Chief Medical Officer reminds me that at one point, blood pressure cuffs were only in hospitals. Now, over a hundred million have been sold and people have them at home. Once, thermometers for taking body temperature were only available when you went and saw the doctor and now that’s at home. Glucose meters are another; you can have them at home now too.
Butterfly, as I told you, is selling to 40 million people in medicine, but our goal is to get the FDA to allow us to sell it directly to consumers. Once that consumer has it, we can ask them to share their medical data in a confidential way so that we can learn, for example, about whether a thyroid nodule is benign or malignant. It’s getting that question in front of people. In my experience, 80% of people asked to participate in a study, whether it’s at the dentist or another medical procedure, will say yes.
FLG: Is there anything else you would like to add?
JR: I would say that this is an amazing period in history. I have always been a serial entrepreneur, but with the breakthrough in our understanding of medicine coming from next-generation sequencing and deep learning AI, I became a parallel entrepreneur. I’ve set up an incubator with a number of companies, and I think that this is a unique period of history with a lot of low-hanging fruit. If you’re the first company to use AI with pathway analysis, or genetic analysis, or drug development, you’re going to make fantastic progress. If you’re the first company to use AI with ultrasound, you’re going to make fantastic progress.
It’s an amazing moment in history to combine breakthroughs in medicine with this breakthrough in artificial intelligence, and for me, personally, that’s why I set up my incubator. It’s too exciting a moment, with the convergence of these two revolutions, to do just one thing at a time. At the turn of each century, it seems that one of these fantastic opportunities happens and so it’s a great time to take advantage of it.