In silico characterization of protein chimeras: relating sequence and function within the same fold.

Area of Science:

• Protein engineering
• Computational biology
• Biotechnology

Background:

• Recombination of protein fragments offers a pathway to novel protein functions.
• Existing methods for creating chimeric proteins are limited in scope and predictability.

Purpose of the Study:

• To develop a predictive model for the structural viability of chimeric proteins.
• To guide the construction of new protein libraries using sequence-function models.

Main Methods:

• A novel framework was developed to predict chimeric protein structural viability using sequence and parental structure.
• The model accommodates sequences from parents with similar folds, overcoming previous limitations.
• Training and testing data were integrated from various protein engineering experiments, including site-directed recombination and DNA shuffling.

Main Results:

• The model achieved state-of-the-art prediction accuracy on hold-out data from site-directed recombination experiments.
• The framework demonstrated convincing performance on chimeric proteins from different experimental origins.
• Successfully assessed the structural viability of P450 chimeras with low sequence similarity to training data.

Conclusions:

• The developed framework effectively predicts chimeric protein structural viability.
• This approach facilitates the design of novel protein libraries with enhanced functionalities for diverse applications.
• Advances protein engineering by enabling semi-random searches for desired protein activities and properties.