Discovering design principles of collagen molecular stability using a genetic algorithm, deep learning, and...

Area of Science:

• Biomaterials Science
• Protein Engineering
• Computational Biology

Background:

• Collagen is a vital structural protein in human tissues, essential for mechanical properties.
• Collagen-based biomaterials are crucial for tissue repair and regeneration.
• Controlling collagen's melting temperature (Tm) is key for stability during processing and in vivo function.

Purpose of the Study:

• To develop a robust framework for designing collagen sequences with specific melting temperatures (Tm).
• To enable precise control over collagen stability for biomaterial applications.
• To advance a 'materials by design' paradigm for collagen-based therapeutics.

Main Methods:

• Developed a general model using a genetic algorithm within a deep learning framework.
• Generated 1,000 de novo collagen sequences.
• Verified Tm values using experimental and computational methods, including molecular dynamics (MD).

Main Results:

• The model accurately predicts Tm values within a few degrees Celsius.
• Identified frequently occurring collagen triplets for direct incorporation.
• Discovered a direct correlation between hydrogen bonds (MD) and triple-helical quality.

Conclusions:

• The developed deep learning model facilitates the design of collagen sequences with targeted Tm values.
• This work is a critical step towards engineering collagen for specific manufacturing methods and biomedical applications.
• Enables a mechanistic approach to materials design for collagen-based biomaterials.