A new parameter-rich structure-aware mechanistic model for amino acid substitution during evolution

Area of Science:

• Computational Biology
• Molecular Evolution
• Biophysics

Background:

• Markov models for amino acid substitution are crucial for phylogenetic studies but often neglect molecular structure.
• Existing models poorly describe amino acid substitution over long evolutionary periods and lack site-specific detail.

Purpose of the Study:

• To develop an improved pseudo-energy-based model for protein structure-aware phylogenetic studies.
• To better characterize amino acid substitution probabilities by incorporating site-specific physical constraints.

Main Methods:

• A new model was developed using site-specific parameterization of pseudo-energy terms within a coarse-grained force field.
• The model evaluates the importance of parameters like contact number, solvent accessibility, and secondary structure elements.

Main Results:

• The new model better describes local heterogeneity in physical constraints on amino acid substitution compared to previous models.
• Model parameters showed biologically reasonable relationships, and proposed substitutions aligned with observed site-specific frequencies in gene family alignments.

Conclusions:

• The developed model enhances the accuracy of simulating protein sequences over phylogenies.
• This advancement supports downstream applications such as positive selection detection, ancestral sequence reconstruction, and protein engineering.