Founder Stories: Alec Nielsen, Co-Founder and CEO of Asimov

Everything in our world is going to be genetically engineered, and I mean everything. Our food, our therapeutics, our materials, our manufacturing infrastructure, our consumer goods. My name is Alec. I'm the co-founder and CEO of Azimov. We're a company in Cambridge, Massachusetts that engineers living systems. Biology is a master architect at the most impressive of scales. This is the majestic sequoia towering hundreds of feet in the air. This tree built itself from scratch from the single-cell stage by compiling carbon out of the atmosphere and turning that into plant tissue. They're pretty science fiction sounding, right? They're solar-powered, self-healing, self-assembling, self-replicating, they're planet-wide. A lot of biology is like that, kind of ridiculous sounding at the face of it. Biology is not just a master architect at the large scale. If you zoom all the way into a single cell, biology is a master architect on the nanoscale as well. And so if you zoom all the way in, what is biology? It's hyper-advanced molecular nanotechnology. Truly, without exaggeration. So what we're looking at here is a single cell, and the cytoskeleton, which is the network that allows the cell to have shape, and to move. Maybe you remember the cytoskeleton from high school biology. This cytoskeletal network is built from proteins. Proteins are molecules that are encoded inside genes, and then they're constructed by cellular machinery with atomic precision. These things self-assemble and self-organize. They're dynamic. Every single cell on the planet has the ability to do computation, that is, perceive their environment, do information processing on those signals, and ultimately regulate genes in response. And of course, biology replicates. This cell has all of the molecular machines needed to make a faithful copy of itself. Biology is hyper-advanced molecular nanotechnology. As engineers, we want to harness that. For a long time, humanity's only ability to manipulate or modify biology was through techniques like selective breeding. And this started a long time ago, millennia. This is how we bred from wolves to dogs. This is how we developed large, tasty, nutritious agricultural crops. And so humanity's relationship with engineering biology goes way back. But then in the 1970s, there was a dramatic shift in the precision with which we could manipulate biology, and that was by engineering DNA sequences, engineering genetics directly. And so in a set of molecular techniques that were developed in the 1970s, we really built the backbone for how we do genetic engineering today. In the '70s and '80s, we built a technique called polymerase chain reaction, which allows us to make lots of copies of DNA. This is really useful if you're going to use that DNA for things downstream. We also developed the ability to cut and paste DNA, okay? We can take genes from one organism, we can remix it with other genes, recombining genes. This is what people mean when they talk about recombinant DNA or transgenic organisms. In addition to these molecular techniques that operate on naked DNA in a test tube, we can put these inside cells. A favorite analogy of our field is that DNA is kind of like software. It's code that you can program, and then when you put that custom DNA into a cell, the cell boots it up kind of like software. It's an imperfect analogy, and it often gets taken too far, but I think it's useful for capturing the idea that we can engineer DNA sequences and imbue organisms with new function. And so these techniques have continued to develop in the decades since, along with other techniques like the ability to write DNA directly using chemical synthesis. And so collectively, these are the types of technologies that we use for the physical process of genetic engineering. What have these techniques gotten us over the past 40 years? Well, here's a whirlwind tour of all of the highlight genetic engineering products that have been brought to market. In one slide. Starting in 1978, Genentech showed that they could produce insulin using genetically engineered cells. You know, previously, if you were a diabetic and you needed to take insulin to treat your disease, that insulin came from the pancreas of pigs and cows. Genentech showed that they could genetically engineer microbes to produce insulin in a bioreactor, in a vat. This turned the therapeutic manufacturing process from one that was basically a farming operation into one that was a lot more akin to brewing beer. The 1980s saw the development and environmental release of the first transgenic crops, notably the ability to engineer herbicide-resistant crops so that farmers could plant these, spray their fields with herbicide, kill all the weeds, and their plants wouldn't die. And it's hard to overstate, actually, the dramatic increases in productivity that this has driven. Driven globally. An application that a lot of people maybe don't realize is made possible with genetic engineering is cheese. Specifically, the process of coagulating milk into curds uses an enzyme called chymosin. We used to have to get chymosin from the stomach linings of cows and goats. Over 90% of the cheese in the world is produced using recombinantly produced chymosin. So if you like cheese, you have genetic engineering to thank. Okay, so we've talked about therapeutics, we've talked about foods, what about materials? So in the 1990s, another blockbuster biotechnology application was the development of cotton that produced its own insecticides. That is, it's making its own defense against bugs from eating it. And this, along with herbicide-resistant plants, are now continuing to climb. They are the dominant ways that people use these crops in agricultural settings. A kind of fun application in the early 2000s was the first commercial transgenic animal called the glowfish. This was a genetically engineered fish that fluoresces when you put it under a black light. You know, it swept the aquarium enthusiast world by storm, and interesting fact, it was actually illegal to sell and possess these fish in the state of California until relatively recently. And that actually laid the stage for other applications in food. The FDA recently approved the first GM animal that's fit for human consumption. It's the AquAdvantage salmon. It's engineered to grow year-round instead of just the spring and summer months, and that should be hitting the supermarket shelves imminently. The Impossible Burger. Consumers are clamoring for it, right? It's the burger, the plant-based burger, that bleeds when you bite into it. And that blood, which is what gives it its meaty flavor, isn't animal blood, it's hemoglobin that's produced, again, using genetically engineered microbes. And so there's been a lot of development, not just on the technology, but also the public adoption of genetically engineered technologies over the past several decades. And then bringing it back to where we started in therapeutics, in the past couple of years, there have been some really incredible advances in cell and gene therapy. This is the idea that we can put DNA into a human cell, in some sense patching the software of the genome to fix some critical defect, And these cell and gene therapies that the FDA has now approved are treating things like cancer, leukemia, lymphoma, other inborn genetic disease like blindness or muscular atrophy. It's the most exciting time in cellular and molecular medicine ever. So that's the whirlwind history. Where do we go from here? Well, I hope it's not too controversial to say that while those products are amazing, they've all been made using relatively primitive genetic engineering. Okay, we take a gene, we put it into a cell, and we try to make a lot of that gene. This is in contrast to what nature is capable of, which has this astoundingly complex system of thousands of genes that's sensing and responding to the environment. We want to achieve that. We want to rival nature in our ability to program these complex genetic functions inside cells. One day, we want to be able to beat nature at our ability to program these functions. And so as synthetic biologists, We think about building integrated genetic systems that have multiple layers, genes that act as sensors, so the cell can perceive stimuli like pH, chemicals, light, pressure, oxygen, convert those environmental signals into a signal that the cell can do information processing on, whether that be digital logic, analog computation, dynamical feedback control. We want to be able to build embedded genetic nanocontrollers to then drive the expression of actuator genes, These are genes that give rise to the phenotypic or behavioral outputs of the cell. Therapeutic responses, cell movement, metabolism are all examples of actuators. And so how do we enable design of these much more complex genetic functions, especially in light of the astounding complexity of evolved biology, right? Evolution has been hacking together biology for the past 4 billion years. There's an insane amount of technical debt associated with it. Well, We and others in the field have been inspired by electronics design and the trajectory that it's taken over the past several decades. So what I'm showing you here is Intel in the 1960s, and what these engineers are doing are carving, by hand, a semiconductor circuit using a scalpel. So that red sheet is a vinyl sheet called Rubylith. They're going to use that as a stencil, a photomask, to pattern the silicon. So literally designing a semiconductor circuit by hand. If you're going to pack 8.5 billion transistors into an iPhone, you need better tools than a scalpel. And so fast forward to 2019. This is the user interface for a state-of-the-art electronic design automation software tool called Altium Designer. You can see the complexity and the sophistication of electronic circuits that can now be grappled with. And in fact, this transition from manual design to computer-aided design isn't unique to electronics. This happens for every engineering discipline as it matures. Think about how we now build bridges and skyscrapers. We didn't need computer-aided design for the first flight with the Wright brothers in— at Kitty Hawk, but we sure need computer-aided design for Boeing 787s and building spacecraft, rockets like the Falcon Heavy. Inevitably, this transition is going to occur for genetic engineering as well. And so at my company, Asimov, that's what we're building. We're building an integrated platform, a compiler and a debugger, to program living cells. We do this by close integration of biology and software. About half of our team works on synthetic biology in the lab, and half of the company works on computational tools. Our initial applications are for mammalian cells, including human cells, specifically for therapeutics applications like protein therapeutics and some of these next-generation modalities like cell and gene therapy. We want to build tools that radically advance our ability to program complex functions inside living cells. But we're not the only ones that are leveraging computation, uh, to advance our ability to program biology. There are many such groups. One really cool area is cancer vaccines. So everyone's tumor is unique. What if we could develop a vaccine that's personalized to a patient and we administer it to them, we supercharge their immune system to go after that tumor? Well, the way that tumors are unique include what are called neoantigens. These are short protein sequences that are displayed on the surface of a tumor cell. If we knew what those neoantigens were, then that would guide the molecular makeup of the vaccine. And so groups are using cutting-edge machine learning techniques to predict what those neoantigens are directly from a tumor's mutated DNA sequence. Another cool application is what's called the Human Cell Atlas. This is an international effort between many groups to build a map of every single human cell type. And there's, you know, about 40 trillion cells in the human body, give or take. And so you're probably familiar with neurons and skin cells, but each of those cells actually has many cell subtypes. There are many cell types that we didn't know about until relatively recently. And this idea of building a map of all human cells requires state-of-the-art statistics. Actually, that white plot, the colorful plot, is a statistical visualization showing how we classify and categorize these cell types in a computationally tractable way. If you look at the history of genetic engineering, it's really revolved around humans finding some interesting gene that nature evolved, and then us taking it and repurposing it for something cool. And this is true kind of across the board in all biotech domains. So evolution has gifted us billions of interesting genetic parts, genes, in the genomes of diverse species. It's a treasure trove for genetic engineers. But it's very difficult for a human to look at a genome sequence and to say, "Aha, there's an interesting genetic part for me to use." But computers excel at this annotation, scanning, and retrieval of DNA sequences. For example, this short DNA sequence called an operator, which is involved in the cell's natural information processing mechanisms. And so, well, right now, We mainly do genetic engineering by repurposing things from nature. Eventually, we may be able to design these genetically encoded molecules, like proteins, directly. Okay? And so one of the cool applications of this would be development of universal flu vaccines. Right now, you go get a flu shot every year to give you some chance of coverage for that year's common flu strain. But what if you could get a vaccine that gave you coverage across all of the common flu types, past, present, and future? This is the grand challenge being pursued by David Baker's group at University of Washington. They've developed a computational protein design software called Rosetta that recently has been making incredible strides in the design of proteins, many of which have no close analog in nature. So eventually we may be able to just design things directly instead of harvesting it from the tree of life. And so where do we go from here? What do computational tools look like decades from now? Well, I like to imagine a whole team of engineers working on an ambitious project to program a single living cell. This team will have multiple sub-teams that are working on various different aspects of engineering the cell, the various subsystems, including metabolism, computation inside of the nucleus, sensing. This software will enable accurate simulation of what genetic constructs do in that cell in silico, before we even build it. These tools will enable us to grapple with very complex genetic designs, likely by using powerful abstractions. We're probably not going to be thinking about things at the level of A's, T's, G's, and C's. Then once we physically engineer that cell, we'll be generating massive datasets for all of the molecular phenomena inside of the cell, everything that we care about. This will enable us to analyze and debug those cells, but it will also provide the input to iteratively improve our computational models, most likely using machine learning. Biology is too complex to be understood by a human brain at a holistic or intuitive level for even a single cell. We need these unbiased statistical approaches that can ingest this large-scale data to learn biophysical models directly. And so the trajectory of synthetic biology is decades from now, we will be able to design entire genomes from the ground up. We will have complete control over the flow of information, matter, and energy inside living systems. And so what do these types of tools enable from a product perspective? Well, genetic engineering is going to transform all aspects of our daily lives. In the same way that the integrated circuit just utterly pervades everything about our daily life, you know, it's in our phones, in our cars, in our toys, you know, in our smart fridges, smart toasters. It's everywhere, right? Engineered biology is going to be similarly pervasive in our lives. Imagine a couple of decades from now when every single component of your house and your car is sourced using biology, which is capable of building things that are inaccessible using any other means in a sustainable manner. Imagine having materials which are themselves living. Imagine your driveway healing itself when it gets cracks in it. Imagine not just having solar panels on your roof, but panels of microorganisms that are extracting pollutants from the atmosphere, cleansing the air. There are going to be neurotechnology applications. What I'm showing you here is what's called the Brainbow. It's a slice of brain where the neurons have been genetically engineered to be slightly different colors, which enables simpler identification and labeling. The advances in this technology could allow us to one day use engineered viruses with DNA barcodes to build a full connectomic map of a human brain and potentially enable whole-brain emulation. Genetic engineering may be the key to building better brain-computer interfaces by making neurons responsive to light using tools called optogenetics. Photons have higher spatial precision than electric fields, like you would get from an electrode, and so it may be the more precise way to read and write brain states. And space, the final frontier, right? Everyone talks about rocket science being the hard problem about outer space, but biology is the really hard thing about outer space. We need to be able to manufacture food and therapeutics in the confined closed system of a space station. One day, we need to be able to terraform entire planets to be hospitable for life. We need to be able to engineer biomes. These are biology problems, not rocket problems. And so I think biology evokes many different things in people's minds. I think they think, that there's a sphere of things that biology touches. But as genetic engineering advances, that sphere of what's possible with biology will grow without bound. It's an almost infinite space. Okay? Through genetic engineering, we enable programming of matter. That's a big idea, and it's why eventually everything will be genetically engineered. Thank you.