Cambridge Consultants brainpower sparks protein engineering AI revolution

Capgemini, a multinational IT services and consulting company, has unveiled a new generative AI-driven methodology for protein engineering that uses a specialised protein large language model (pLLM) to predict the most effective protein variants.

With a patent pending, this novel approach will help accelerate the advancement of the global bioeconomy and drive critical scientific breakthroughs across industries including healthcare, agriculture, and environmental science.

By reducing the datapoints required to design protein sequences by over 99 per cent, the new methodology harnesses the power of gen AI to drastically reduce the time and resources needed for R & D. 

Using this approach, Capgemini can help clients reduce the development cost of biosolutions and unlock business cases that were not previously viable.

The methodology was created in the bespoke gen AI-driven biotechnology lab of Cambridge Consultants. It was applied to several critical use cases to demonstrate how it could drive a step-change in innovation.

Advancements in engineering biology are expected to disrupt all industries, with half of business leaders predicting this transformation will happen within the next five years.

However, data can be a critical bottleneck in research timelines. This new approach makes scientific breakthroughs possible with significantly smaller data sets, enabling organisations to innovate even in resource-constrained environments.

Using this novel approach, Capgemini says it is exceptionally positioned to help clients find and develop innovative solutions to global challenges such as disease, food security, and climate concerns.

A key advantage is a 60 per cent increase in plastic degradation efficiency; by making it easier to degrade plastic, this breakthrough can support sustainability objectives and help lower operational costs associated with waste management.

Using gen AI predictions, Capgemini also reduced the number of experiments needed to identify an improved variant of the commonly cited Green Fluorescent Protein benchmark, from thousands to just 43 data points, achieving a brightness level seven times greater than that of the natural jellyfish protein. 

This significantly cuts down on the time and resources typically required for experimental testing, enabling quicker deployment across a range of fields, from accelerating drug discovery and enhancing diagnostic tools to advancing bioengineering applications.

Roshan Gya, CEO of Capgemini Invent said: “Capgemini’s proprietary generative AI-driven approach means we are uniquely placed to enable clients to significantly accelerate their bio-journey in previously untapped areas and, crucially, contribute to helping solve many of humanity's most pressing challenges.

“Our new methodology is faster, more cost-effective, and opens the door to new opportunities for clients to develop innovative bio-based solutions. The Capgemini Group delivers end-to-end engineering biology and scale-up capabilities so that our clients can derive significant business value and develop proprietary IP, moving away from traditional carbon-based approaches and fueling growth in the bioeconomy.”

The bespoke DeepTech facility at Cambridge Consultants is home to an unrivalled combination of multidisciplinary experts in biology, chemistry, gen AI, digital twins, electronics, software, sustainability and more.