Laurynas Karpus, Vykintas Jauniškis and Irmantas Rokaitis

Enzymes are used in a wide range of industries, however, many commercially available enzymes are not optimised for these applications, limiting their efficiency and scalability. Traditional enzyme engineering follows a top-down approach in which existing natural enzymes are modified. This method limits the types of molecules that can be built, as well as the reactions and properties they can support. As a result, it narrows the range of potential products and technologies. Lithuanian entrepreneurs Karpus, Jauniškis, Rokaitis and their team at Biomatter have introduced Intelligent Architecture™, an AI-based enzyme engineering platform that enables enzymes to be designed from first principles, rather than adapted from nature. This method has the potential to accelerate progress in biotechnology by creating enzymes that are tailored for specific industrial applications. 

The platform incorporates AI and physics-based modelling to design enzymes at the molecular level. AI models, trained on large datasets of both natural and synthetic enzymes, predict optimal enzyme structures and refine them for industrial use. This method differs from traditional approaches in that it begins by defining the desired chemical reaction, allowing for the creation of enzymes that do not exist in nature. The platform continuously improves as it integrates experimental data, enhancing both precision and efficiency. 

 Advancing the bioeconomy 

The inventors met at Vilnius University’s Institute of Biotechnology, where inspired by developments in AI-generated imagery, they applied similar machine-learning techniques to enzyme design. They built a generative adversarial network algorithm for new functional enzyme generation, named ProteinGAN, and published this work in Nature Machine Intelligence journal. Initially, the inventors collaborated with researchers at Chalmers University of Technology, who tested their AI-generated enzymes and confirmed their functionality. Encouraged by positive experimental results, they co-founded Biomatter in 2018 together with Donatas Repečka and Professor Rolandas Meškys. 

 Since its founding, the company has successfully applied its method in multiple industry collaborations. For example, Biomatter partnered with Kirin to develop enzymes for the production of Human Milk Oligosaccharides (HMOs), essential nutrients for infant health that are difficult to produce at scale. Another collaboration, with ArcticZymes Technologies, focuses on optimising enzymes for molecular diagnostics and pharmaceutical manufacturing, such as those used in gene therapy and vaccine production. 

Reflecting on their work and success, Karpus notes, “I’ve been incredibly fortunate with my team, starting from the co-founders to the immensely passionate team of 30 we are today.” 

Towards the UN Sustainable Development Goals (SDGs) 

The company’s AI-driven enzyme engineering supports SDG 3 (good health and well-being) by enabling the design of novel proteins and enzymes that could lead to new drugs, therapeutics and improved diagnostics. It also advances SDG 9 (industry, innovation, and infrastructure) by providing a platform with applications across multiple industries.