Capgemini has unveiled a GenAI-driven method that drastically cuts the data needed for protein engineering, accelerating bio innovation.
Capgemini’s new methodology leverages a specialized protein large language model (pLLM) to predict optimized protein variants with 99% fewer data points. This GenAI removes a major research bottleneck, allowing organizations in healthcare, agriculture, and environmental science to innovate even in resource-constrained environments. By reducing experimental requirements, the approach lowers R&D costs. Making previously unfeasible bio solutions commercially viable.
The method, developed at Capgemini’s AI-driven biotechnology lab in Cambridge Consultants, has demonstrated remarkable results. It enhanced the efficiency of a plastic-degrading enzyme by 60%, offering a scalable solution for tackling plastic waste. Additionally, it reduced the number of required experiments for improving Green Fluorescent Protein from thousands to just 43, increasing brightness sevenfold. These applications illustrate how GenAI accelerates scientific discovery by minimizing time-consuming trial-and-error testing.
Capgemini sees this innovation as a enhancement and improvement for the bioeconomy side. This aid enables businesses to transition from traditional carbon-based solutions to sustainable biotechnologies. With a dedicated AI lab and expertise in engineering biology, Capgemini aims to help clients develop proprietary IP and scale up bio-based innovations faster. Experts, including Professor Stephen Wallace from the University of Edinburgh, recognize the potential of this approach in reshaping industrial bioprocesses. Making bioengineering more efficient and scalable.
Capgemini’s GenAI-driven protein engineering could transform industries by significantly reducing costs, speeding up discovery, and enabling more sustainable solutions. This methodology is expected to open new opportunities for bio-based innovations in the coming years.