Smoothed Bootstrap Aggregation for Assessing Selection Pressure at Amino Acid Sites

Area of Science:

• Molecular Evolution
• Bioinformatics
• Genomics

Background:

• Detecting positive selection at amino acid sites is crucial for understanding protein evolution.
• Current methods, like empirical Bayes (EB), rely on parameter estimates from codon evolution models.
• Inaccurate parameter estimates can lead to unreliable positive selection inference.

Purpose of the Study:

• To introduce and evaluate a novel method, smoothed bootstrap aggregation (SBA), for detecting positive selection.
• To address limitations of the Bayes Empirical Bayes (BEB) method, particularly concerning parameter estimate uncertainty.
• To improve the accuracy and power of site-specific positive selection inference.

Main Methods:

• Smoothed bootstrap aggregation (SBA) was developed, utilizing bootstrapping of site patterns from protein-coding DNA alignments.
• Kernel smoothing techniques were employed to refine parameter uncertainty corrections.
• The performance of SBA was compared against the Bayes Empirical Bayes (BEB) method through simulations and real data analysis.

Main Results:

• SBA effectively accommodates uncertainty in evolutionary parameter estimates.
• SBA demonstrates comparable or superior performance to BEB in balancing accuracy and power.
• SBA's advantages become more pronounced when parameter estimates are unstable.

Conclusions:

• SBA provides a robust and improved approach for detecting positive selection at individual amino acid sites.
• The method enhances site-specific inference by directly addressing parameter estimation uncertainty.
• SBA's applicability extends to various inference problems within molecular evolution research.