.png)
"Enzidia has signed an exclusive license with DTU for MillionFull, a technology generating millions of enzyme sequence-to-performance datapoints per run — now powering a collaboration with Nobel laureate Frances Arnold's lab and eXoZymes on ML-driven enzyme design.
ENZIDIA and DTU - Technical University of Denmark have signed an exclusive licensing agreement for the commercialization of the MillionFull technology!
MillionFull is a technology for obtaining massive full-length enzyme sequence-to-performance datasets, capable of gathering millions of datapoints each run.
The development was motivated by the data drought holding back the exciting potential of AI for enzyme acceleration. The solution came together while I was working as a postdoc at DTU Biosustain. The dataset from the demo run, with >100,000 data points, was open-sourced half a year ago.
With the licensing agreement, we can now apply it to solve real-world problems in biomanufacturing.
Great thanks to Morten Birkeland, our IP advisor, for representing Enzidia in the licensing term discussions, so that we could get to a fair deal structure quickly with clear rationales. I got to learn so much by being the student in the process!
Many thanks to Jens Kindtler and other colleagues on the DTU side for a smooth agreement forming process. We got to see first-hand how DTU executes its mission of "technology for people & society" by supporting research translation.
Great appreciation to Dr. Lei Yang, Prof. Alex Toftgaard Nielsen for their support with scientific input and resources. This work would not have been possible without your recognition and support of new ideas.
Thanks to Bjarke Erichsen and Simon Richter Krarup for helping make the idea a reality.
Thanks to DTU Biosustain and Novo Nordisk Foundation for providing the breeding ground for innovation.
More acknowledgments can be found in our preprint.
----
There's a second part to this post:
MillionFull is a foundational technology that serves as the crystallization core for much to build upon.
Our collaboration with the group of Nobel laureate Frances Arnold, a leading lab for integrating machine learning into enzyme engineering workflows, and eXoZymes (Nasdaq: EXOZ), a company with a highly efficient cell-free enzyme prototyping platform, is the prime example.
Based on the massive data enabled by MillionFull, and the cutting-edge ML algorithms developed by the Arnold group (Jason Yang, Sonia Yuan), we've been making exciting progress in the learn and design steps of the Design-Build-Test-(Data)-Learn cycle. ExoZymes' platform (Bastian Vögeli, Tyler Korman, Michael Heltzen) has allowed fast testing of the designed enzyme variants.
Special thanks to Sonia for leading this chapter of the development! Attached is her presenting our progress at the Generative and Experimental Perspectives for Biomolecular Design workshop (https://www.gembio.ai/), part of the ICLR conference at Rio de Janeiro in April.
Meanwhile, our open-sourced dataset has been downloaded ~750 times by the ML protein/enzyme engineering community! We don't know most of the developers, but we look forward with excitement to the fruits of the synergy between big datasets and the right ML approaches built for them.
