UNFOLDING THE FUTURE OF PROTEIN DESIGN

Protein Engineering

through Directed Evolution

PHAGEVO - PANCE X EVOLUTION.T7

Natural protein evolution has been a driving force behind the diversity of life on Earth for more than 500 million years. However, nature hasn’t had the time to evolve new proteins with desired, novel, and enhanced functionalities. Directed evolution can. This protein engineering method mimics and accelerates this process by generating random mutations and selecting desired variants.

This project introduces PHAGEVO, a novel tool that combines Phage-Assisted Non-Continuous Evolution (PANCE) and the Evolution.T7 system, providing a powerful method for in vivo protein evolution. While PANCE enables efficient selection of functional variants, it lacks targeted gene specificity. We decided to combine this technology with a system we developed in 2021: Evolution.T7. This technology enables ‘targeted and modular double-stranded in vivo mutagenesis using RNA polymerase-guided cytosine and adenine deaminases’ that introduce mutations in specific gene sequences. By combining these two technologies, PHAGEVO increases mutation precision and enhances efficient variant selection.

As a proof of concept, the project initially focuses on evolving the XylS transcription factor, a protein engineered to detect plastic degradation by-products such as Phthalic Acid (PA) and Terephthalic Acid (TPA) in water. Enhancing the specificity of XylS variants for these non-natural compounds would lead to more sensitive and effective biosensors for detecting trace amounts of small molecule pollutants and toxins in air, water and soil - improving environmental monitoring and biosensors design.

A NEW COMPUTATIONAL GENERATION MODEL

Artificial Intelligence

versus Nature : Who wins ?

FLINT - POCKETGEN EVOLUTION

We are entering the AI era, where new advanced language models could reshape in silico protein by compiling data from millions of years of Evolution on earth. Our project, PHAGEVO, combines latest directed evolution approaches and a AI model, nature against AI in order to optimise the XylS transcription factor for plastic degradation metabolites recognition and specificity. We aim to compare outputs/results from our tool to an AI-based predictive protein model we designed. By comparing nature’s evolution with advanced computational tools we explore whether our tool or AI is more effective/efficient in finding optimal protein variants.