Proteins are the most advanced nanotechnology we know of. At Adaptyv Bio, our mission is to make proteins easier to engineer. We are building technologies that allow protein engineers to design and test new proteins in order to develop new medicines, better enzymes for research and industrial applications or functional materials with novel properties.
We believe that designed proteins will be at the core of the bio-revolution and radically change how we manufacture goods, treat aging & diseases and even how we think about applications such as computation and data storage. We want to unlock that potential and help advance humanity to a post-scarcity future.
Closing The Loop
As the field of generative AI for proteins continues to advance, our ability to create new proteins that do not yet exist in nature is opening up exciting new possibilities. However, this also means that with each new design we are now venturing into uncharted territory, and by definition we cannot know beforehand if this novel protein designs will work as intended.
In order to gain knowledge about the function of our designs, we must rely on feedback from reality. When using AI for other applications such as generating new code, that feedback can be rapid. Just copy-paste the code and run it or have the AI run it for you via an API. But in the case of proteins, we need to make a physical thing and see how it interacts in its environment or application in order to get that feedback from reality.
As everyone knows, the world of atoms is fundamentally more difficult to work with than the world of bits. So imagine every time you use ChatGPT to generate some code for you, you then had to wait 10 weeks for it to execute or to tell you that it had a bug. And now imagine each execution costs 1000 USD. That’s pretty much the situation for protein designers today.
We’re here to change that.
We’re building a foundry - an integrated lab that allows protein engineers to rapidly iterate their designs and generate data about their proteins. To enable that tight feedback loop from computational design to experimental validation, we’re innovating on all levels of the technology stack: From custom synthetic biology techniques for designing DNA, synthesizing proteins and assaying them, to novel lab-on-chip modular workcells that can replace multiple commercial instruments at once, to advanced lab robotics that allow us to run experiments autonomously.
This way we can reduce reagent consumption, time per experiment and thus cost per data point generated by multiple orders of magnitude compared to today’s approaches. To make proper use of these new capabilities, we’re building our foundry from the ground up to integrate with AI at every step of the protein engineering cycle - from sequence design, to planning experimental protocols, to running those protocols and tracking all data and metadata along the way.
This is how biology should be done.
Future: Roadmap toward molecular nanotechnology
Take a look at the molecular machinery found in every single one of the trillions of cells in our body. Imagine the kind of technological progress we could make by re-purposing and re-designing some of those proteins for new applications.
Over the past couple of years we’ve seen significant progress in the field, starting with predicting protein structure (Alphafold) and now generating entire proteins (RFDiffusion & co). Soon, we'll be designing arbitrary binding surfaces. However, we want to be able to create more complicated proteins than just simple monomeric ones or symmetric assemblies. Most of the interesting molecular machinery we know of today is inherently composed of many different subunits. And we don't just want to design stationary protein structures, but we want to use protein technology to rearrange atoms and actively transform molecules. Solving the most common enzyme classes and designing new ones will enable us to rapidly transition many of today's industrial processes to bio-based methods.
However, these new design challenges will inherently require an increasing and more complex amount of data. A lot of our excitement about this space comes from the fact that new protein engineering capabilities have a very interesting compounding effect. The more tools we're developing to engineer proteins, the easier we can also create new proteins that help us generate new data about protein behavior which in turn makes it easier to engineer more complex proteins. For instance, nanopores are such engineered proteins that now offer a fundamentally new way of generating more knowledge about biology. Or think about enzymes that can function as biological logic gates to regulate gene expression, thereby controlling the protein design space that can be explored.
With this ever expanding toolkit, we'll be able to tackle more complicated protein engineering challenges, such as motor proteins or complicated molecular machines for energy generation or computation. Today we can merely imagine a tiny fraction of all the amazing new applications that will be possible. Proteins really are the most advanced nanotechnology we know of and we want to unlock that potential and help advance humanity to a post-scarcity future.
Enamoured by the potential of synthetic biology but frustrated by the challenges of actually running biological experiments in present day labs, we teamed up in early 2021 to build better infrastructure & tools for protein designers to engineer proteins.
In the summer of 2021 we did YCombinator and then raised our pre-seed round led by Wingman Ventures. Since then we’ve been laser-focused on building our tech at the Biopole Life Sciences Campus.
Today we’re starting to share some of the things we’ve been working on - and we’re also opening up Early Access to our platform!
If you are a protein designer or want to become one, come talk to us - we want to get to know you.
If you align with our vision and want to make protein engineering easier, come work with us - we are always looking for extraordinary people to join our team.
— Julian for Team Adaptyv Bio