De novo protein design-From new structures to programmable functions.

Area of Science:

• Biochemistry
• Computational Biology
• Synthetic Biology

Background:

• Artificial intelligence (AI) methods trained on extensive sequence and structure data enable the design of novel proteins.
• Traditional protein design relies on naturally occurring proteins, limiting innovation.
• Emerging techniques integrate engineering principles for enhanced control.

Purpose of the Study:

• To review the current state of de novo protein design.
• To highlight the convergence of AI and physics-based modeling.
• To discuss future directions and challenges in the field.

Main Methods:

• AI-driven sequence and structure prediction.
• Physics-based computational modeling.
• Experimental validation of designed proteins.

Main Results:

• Successful design of novel protein folds and assemblies.
• Achieving tunable control over protein conformations.
• Enabling precise molecular recognition through shape complementarity.

Conclusions:

• De novo protein design is rapidly advancing, merging AI and physical sciences.
• The field shows promise for deconstructing cellular functions and building synthetic systems.
• Further research is needed to address remaining challenges and unlock full potential.