Detecting Evolutionary Change-Points with Branch-Specific Substitution Models and Shrinkage Priors

Area of Science:

• Evolutionary biology
• Computational biology
• Genomics

Background:

• Branch-specific substitution models are key for identifying evolutionary shifts, but often require predefined change-point locations or struggle with large datasets.
• Existing methods face limitations in scalability and the need for prior knowledge, hindering broad application.

Purpose of the Study:

• To develop a novel computational framework that integrates branch-specific substitution models with shrinkage priors.
• To enable automatic detection of evolutionary change-points without prior location knowledge.
• To enhance the efficiency and scalability of evolutionary inference.

Main Methods:

• Integration of branch-specific substitution models with shrinkage priors for automatic change-point identification.
• Development of an analytical gradient algorithm for efficient estimation of branch-specific substitution parameters.
• Application to infer evolutionary dynamics in the BRCA1 gene and mpox virus sequences.

Main Results:

• The new method automatically identifies change-points and estimates distinct substitution parameters per branch.
• The analytical gradient algorithm achieves linear computation time with respect to the number of parameters.
• Significant computational speedups were observed: up to 90-fold in maximum-likelihood optimization and 360-fold in Bayesian inference.

Conclusions:

• This novel algorithm overcomes limitations of existing methods, enabling efficient and accurate evolutionary inference.
• The approach facilitates the study of selection pressure and mutational dynamics in diverse biological systems.
• The developed computational tools significantly advance the field of evolutionary genomics.