This essay is heavily inspired by Andreessen Horowitz’s bio fund, and their stance on biology and computer science. You can read more about them here.
The world is built off of evolution. Over millions of years, species evolved from single-cell organisms to organisms that each consist of nearly 32.7 trillion cells! Over thousands of years, wolves evolved to become domesticated, and house cats evolved from an ancient breed of wildcat in Egypt.
Evolution has brought forward amazing features, and with them, bugs. For example, humans fight cancer, alzheimers, and diabetes, but many species never get them. We are full of imperfections in a complex system of proteins, cells, membranes, and DNA. These minute imperfections of evolution are what cause some of the world’s toughest diseases.
The way you and I learned biology is that cells are the fundamental unit life was built from- trillions of them exist in all animals and humans. But in reality the cell is one of the most complex mechanisms, made up of billions of biochemical molecules acting in incredible motion to create proteins and replicate cells. In reality, the structure of a single skin cell is probably far more complex than an entire space shuttle or nuclear submarine.
Because of its complexity, biology was a very scattered research area with very little progress towards eradicating disease. We formed few conclusions and far more questions. But now, things are starting to change, as biology is starting to be approached from an engineering perspective. This change, in my opinion, is likely to help us advance research in biology at a far more rapid pace than it ever has before, almost like the computer revolution of the 70s. Biology is entering its very own industrial revolution.
Engineering is creating the modularization of biology. It is breaking down biology into modules, where each module combines to create complex solutions, kind of like a physics equation with several variables. This approach to biology enables us to do incredible things like re-engineer gene circuits and cure cancer in as many as 94% of patients using programmable immune therapies. The idea of being able to create, edit, and improve on treatments towards disease, almost like writing a computer program, is something we could never do before.
But the toughest problem with trying to understand biology is that it’s full of data across thousands of different samples and sources. The human genome is over 3 billion nucleotides long, and virtually unreadable by the human eye. This makes it incredibly hard for scientists to draw patterns from all this data and find correlations between some of the toughest diseases in the world, such as cancer.
This is where self-learning computers, or machine learning comes in. Machine learning exceeds human capabilities since these machines are faster and more accurate. They can identify patterns by learning over millions of iterations within short periods of time that would take humans years to learn. To put this into perspective, discoveries that might take scientists 2 years to find can be found by computers within 20 hours, if we can teach them using the right data. This past week, Google Health released a paper highlighting their image detection algorithm for diagnosing diabetic retinopathy, which beat the accuracy of medical professionals with years of training and experience by nearly 20%. It did so by one reason: it practiced on thousands of images, simulating the same experience the same doctor had developed.
AI-powered platforms have the potential to connect the dots in ways which humans never could; to generate new discoveries; even to change the nature of discovery itself. This will drive new therapies and next-generation tools that give us the ability to detect diseases like cancer earlier, perhaps even stopping disease before it begins.
When you look at industrial revolutions, there tends to be a key shift between the artisan, bespoke (custom-fit) approach to making things towards an industrial, standardized approach. With tools like gene-editing CRISPR-cas9 and immune system therapies like CAR-T, biology is evolving from a field that was formerly laborious, time-consuming and inefficient to an industry where easily reprogrammable therapies and tools like machine learning are saving lives.
In the industrial revolution of biology, we’re taking control of the future survival and longevity of our species from evolution, and putting it in our hands.
Thanks for taking the time to read through my essay on how biology is entering its stage of industrialization.! If you have any questions or comments, or just want to chat about biotech, feel free to chat with me on Twitter. Or if you would like to learn about any new content I put out, subscribe to my monthly newsletter to see new projects, conferences I go to, and articles I put out!
I’m currently working on building Project De Novo, where I’m working on building models to understand how mutations affect gene expression, and how we can use machine learning for computer-aided drug design.