AI Tool Streamlines Protein Engineering
New method enhances precision medicine, molecular breeding applications.
A novel protein engineering technique, developed by Chinese scientists, leverages artificial intelligence to accelerate protein evolution. This advancement promises to make protein engineering more accessible and efficient for researchers.
Key Advancement: AI-informed Constraints for Protein Engineering (AiCE)
A team of Chinese researchers, spearheaded by **Prof. GAO Caixia** from the Institute of Genetics and Developmental Biology (IGDB) of the Chinese Academy of Sciences, has introduced a method called AI-informed Constraints for protein Engineering (AiCE). It integrates structural and evolutionary constraints into an inverse folding model without the necessity of training specialized AI models.
The study, which appeared in *Cell* on July 7, attempts to overcome the limitations associated with traditional protein engineering techniques.
AiCEsingle, a module predicting high-fitness single amino acid substitutions, enhances prediction accuracy by incorporating structural constraints while sampling inverse folding models. Benchmarking showed that AiCEsingle outperformed other AI-based methods by 36–90%.
The researchers also developed the AiCEmulti module to accurately predict multiple high-fitness mutations at minimal computational cost. It integrates evolutionary coupling constraints to address negative epistatic interactions in combinatorial mutations, expanding the tool’s versatility.
Applications in Precision Medicine
Utilizing the AiCE framework, the team successfully evolved eight proteins with diverse structures and functions. These engineered proteins facilitated the creation of next-generation base editors for precision medicine and molecular breeding. Those proteins include enABE8e, a cytosine base editor with a ~50% narrower editing window; enSdd6-CBE, an adenine base editor with 1.3× higher fidelity; and enDdd1-DdCBE, a mitochondrial base editor showing a 13× increase in activity.
The global market for gene editing is projected to reach $15.76 billion by 2030, driven by technological advancements and increasing applications in disease treatment (Fortune Business Insights 2023).
Impact of AiCE
AiCE is a strategy for protein engineering that is simple, efficient, and broadly applicable. By leveraging the potential of existing AI models, it enhances the interpretability of AI-driven protein redesign and presents a promising new direction for the field.
