Addressing epistasis in the design of protein function

Area of Science:

• Biochemistry
• Molecular Biology
• Protein Engineering

Background:

• Protein active sites are crucial for function but sensitive to mutations.
• Epistasis, the interaction between mutations, complicates protein evolution and variant design.
• Understanding epistasis is key to overcoming limitations in natural and laboratory evolution.

Purpose of the Study:

• To investigate the role of epistasis in protein active site evolution.
• To explore sequence- and AI-based strategies for predicting epistatic relationships.
• To design and create novel, highly functional protein variants using computational approaches.

Main Methods:

• Analyzing mutational patterns in natural and experimental evolution data.
• Employing artificial intelligence (AI) and sequence-based strategies to infer epistasis.
• Utilizing atomistic design calculations to predict combinatorial mutations.
• Experimentally validating designed protein variants.

Main Results:

• Successfully predicted and designed thousands of functional active-site variants.
• Demonstrated that critical design determinants for epistasis are now sufficiently understood.
• Showcased the power of combining computational and AI approaches for protein design.

Conclusions:

• Epistasis, while a challenge, can be overcome with advanced computational and AI tools.
• Protein design strategies can significantly expand the explored sequence and structure space.
• This work deepens the understanding and control of protein activity and function.