deCYPher is a new European Horizon project that seeks to merge synthetic biology with artificial intelligence and machine learning. Launched with a kick off meeting in Gent, Belgium, representatives of 10 partners from 6 European countries came together to discuss deCYPher’s ambitious goal: overcoming hurdles to produce valuable compounds (usually extracted in small amounts from plants) in industrial biotechnology – specifically terpenoids and flavonoids, natural compounds that encompass about 30.000 different chemical substances.
Terpenoids and flavonoids are traditionally extracted from plants and have been given much attention due to their importance for a variety of applications and markets, including pharmaceuticals, fragrances, flavours, food preservatives, and insecticides. However, only a tiny fraction of these large and versatile groups of compounds can be commercially extracted from plants, and due to their low contents within the plants such commercial processes require unsustainable large-scale field cultivation. The diversity of terpenoids and flavonoids stems from multiple variations on common molecule structures, which for a majority comes from oxygenation carried out by a vast family of enzymes called Cytochromes P450 (often dubbed CYPs). Terpenoids and flavonoids illustrate once more the complexity involved in biological processes and the difficulty of decoding and re-engineering them in a lab, not to mention at industrial and commercial scales.
“Artificial intelligence and machine learning advances reduce the design possibilities to the most probable ones and can discover new-to-nature solutions,” noted project coordinator Marjan De Mey.
On top of its contribution to decode nature’s building blocks and processes, deCYPher will apply an innovative and holistic approach that seeks to innovate how bioprocesses are developed. deCYPher will develop a standardised artificial intelligence and machine learning platform for biotech applications as well as guiding bioprocesses across all steps of the development chain. Drawing on specific targets for case studies (and decoding CYPs), deCYPher plans to methodically work on taking bioprocess design and control further from an art form to a reproducible and sustainable system.
ML6 is the commercial ML/AI partner of the consortium and will focus on building the ML toolbox that brings protein discovery / design (AlphaFold, ESMfold...) and active learning to the DBTL cycle of wet lab work. Considering both functional CYP expression for pathway & host optimisation as well as for upscaling.
Details and figures at a glance:
For more information contact:
Jan Gerlo, Business Lead Life Science at ML6
Pieter Coussement, Squad Lead Bio at ML6