Innovation by homologous recombination

Area of Science:

• Protein Engineering and Computational Biology
• Molecular Biology and Biochemistry

Background:

• Protein homologs can be recombined to create chimeric proteins with novel properties not found in parent proteins.
• Existing computational methods utilize structural and sequence data to design functional protein chimeras.

Purpose of the Study:

• To explore the design and engineering of functional protein chimeras through fragment swapping.
• To leverage computational approaches for predicting and elucidating the structure-function relationships of engineered proteins.

Main Methods:

• Utilized fragment swapping among protein homologs to generate chimeric proteins.
• Employed computational methods incorporating structural and sequence alignment data.
• Applied machine learning models, including linear regression, Gaussian processes, and support vector machines, to model sequence-function relationships.

Main Results:

• Successfully generated chimeric proteins exhibiting diverse properties, such as altered thermostability, mechanical stability, enzyme substrate specificity, and optogenetic functions.
• Demonstrated the predictive power of machine learning models in identifying potentially useful chimeric proteins.
• Provided insights into the structural basis underlying the novel functions of engineered chimeras.

Conclusions:

• Fragment swapping is a viable strategy for engineering protein chimeras with desired and novel functions.
• Computational and machine learning approaches are powerful tools for designing and understanding protein chimeras.
• This work facilitates the rational design of proteins with tailored properties and deepens the understanding of protein structure-function dynamics.