Modeling the site-specific variation of selection patterns along lineages

Area of Science:

• Evolutionary Biology
• Molecular Evolution
• Bioinformatics

Background:

• Detecting positive Darwinian selection in protein-coding DNA relies on comparing nonsynonymous and synonymous substitution rates.
• Existing models often assume constant selection regimes across lineages for each site, limiting their accuracy.
• A need exists for models that accommodate dynamic changes in selection processes over evolutionary time.

Purpose of the Study:

• To develop a novel statistical framework for analyzing site-specific selection processes that vary along phylogenetic lineages.
• To model switches between different selection regimes at individual amino acid sites during evolution.
• To improve the accuracy of detecting and quantifying positive Darwinian selection in DNA sequences.

Main Methods:

• Developed a codon substitution model incorporating site-specific selection process variation along lineages.
• Employed maximum likelihood estimation to parameterize the switching model.
• Applied the model to homologous sequence data sets from HIV-1 env.

Main Results:

• The proposed switching model significantly outperforms models that assume constant selection regimes.
• The model provides a substantially better fit to empirical sequence data, including HIV-1 env.
• Ignoring site-specific variation in selection regimes can lead to inaccurate estimations of selection strength and frequency.

Conclusions:

• Site-specific selection processes can vary dynamically along phylogenetic lineages.
• Accounting for these switches is crucial for accurate evolutionary inference and detection of positive selection.
• The developed model offers a robust statistical framework for studying evolutionary dynamics at the molecular level.