Mutational Paths with Sequence-Based Models of Proteins: From Sampling to Mean-Field Characterization.

Area of Science:

• Evolutionary biology
• Bioengineering
• Computational biology

Background:

• Identifying and characterizing mutational paths is crucial for understanding evolutionary processes.
• These paths have significant implications for bioengineering and protein design.

Purpose of the Study:

• To propose and validate an algorithm for sampling and characterizing mutational paths.
• To apply the algorithm to both theoretical protein models and natural protein sequence data.
• To extend evolutionary distance estimation methods to complex epistatic models.

Main Methods:

• Development of a novel algorithm for sampling mutational paths.
• In silico benchmarking using exactly solvable protein models.
• Application to data-driven protein models derived from sequence data using restricted Boltzmann machines.
• Utilizing mean-field theory for path characterization and evolutionary distance extension.

Main Results:

• The proposed algorithm effectively samples and characterizes mutational paths.
• Successful application to both simplified and complex, data-driven protein models.
• Extension of evolutionary distance estimation to epistatic models of selection.

Conclusions:

• The developed algorithm provides a robust tool for studying protein evolutionary paths.
• This work bridges theoretical evolutionary biology with practical bioengineering applications.
• The findings offer new insights into sequence-based epistasis and evolutionary dynamics.