AI tools like ChatGPT are dramatically changing the way text, images and code are generated. Similarly, machine learning algorithms and generative AI are disrupting traditional methods in life sciences and accelerating drug discovery and materials development timelines.
DeepMind’s AlphaFold is probably the most famous machine learning model in the field. It predicts the three-dimensional structure of proteins from amino acid sequences, and has been used by more than one million researchers in the 18 months since its release. Since then, many other AI tools have emerged, including the recently open-sourced RFDiffusion. This allows researchers to generate computational protein designs using only a laptop.
However, translating these computational designs into concrete functional proteins remains a challenge. Adaptyv Bio aims to address this issue with the next generation protein foundry. By integrating technologies from advanced robotics, microfluidics and synthetic biology, Adaptyv Bio has built a full-stack platform to enable protein engineers to validate his AI-generated protein designs.
Julian Englert, CEO and co-founder of Adaptyv Bio, said: “Proteins are at the heart of the biorevolution as new drugs, improved enzymes for research and industrial applications, or materials with unique properties. Protein designers have access to amazing new AI tools such as AlphaFold and RFDiffusion. But validating protein designs in the laboratory to see if they work remains a major challenge.”
AI models leverage data for training and better predictions. By simplifying the process of generating data on the efficacy of engineered proteins, Adaptyv Bio enables protein engineers and her AI models to receive more feedback on their designs, resulting in better performance. of protein.
Engratt added, “Think of the AI in a self-driving car. To keep the car on the road and reach its destination, the AI model needs a tight fit by capturing large amounts of high-quality data from the car’s camera sensors. It requires a feedback loop.The same principle applies to AI models that design new proteins, and the feedback mechanism involves actually making the protein in the lab and testing its performance.”
Adaptyv Bio was founded by a group of engineers from the Swiss Federal Institute of Technology EPFL in Lausanne. Motivated by the time-consuming process of conducting biological experiments in the lab. In 2022, he secured $2.5 million in pre-seed funding from Wingman Venture after joining Y Combinator, the world’s most selective startup his accelerator. The team has since expanded to his 12 engineers with diverse backgrounds in synthetic biology, microengineering, software development and machine learning. The company develops its technology in a state-of-the-art laboratory facility located at the newly constructed Biopol His Life His Science His campus in Lausanne, Switzerland, with beautiful views of Lake Geneva and the Swiss-French Alps.
Adaptyv Bio’s foundry centers around protein engineering workcells, custom automated setups that miniaturize processes that would normally require multiple experimental devices and run them in parallel on tiny microfluidic chips. The user can create (or let AI create) an experiment protocol, and the workcell autonomously runs the experiment while closely controlling and monitoring the parameters of the experiment. All measurement data is automatically processed and uploaded, allowing users to refine their machine learning models with each experiment.
Englert says: “Our workcell is fully automated, uses 1,000x fewer reagents than commercial alternatives, and can run thousands of different proteins per day for each individual setup. Streamline your experimental workflow. We have developed a number of custom synthetic biology and automation techniques to support our customers, and over the next 12 months we plan to further scale up our lab to increase the number of protein design applications we can support, and also allow users to develop their own protein design projects. We have just entered early access for submissions and are working hard to onboard new projects as soon as possible.”
To further accelerate the field of protein engineering, Adaptyv Bio has open sourced two internal tools that are already gaining traction among researchers and engineers in the field. ProteinFlow is a Python library that enables protein designers to easily create high-quality datasets for better AI models. Automancer is an extensible software platform for running automated experiments, allowing researchers to build their own experimental protocols and integrate various experimental setups.
“Our mission is to make protein engineering easier and enable more researchers to design new proteins. If we could start designing new proteins for personalized medicines, industrial applications such as new enzymes, or better, more sustainable materials, what kind of technological change would humanity make? Imagine if you could make progress. Added Julian Englert.
