An empirical test for branch-specific positive selection.

Area of Science:

• Evolutionary biology
• Genomics
• Bioinformatics

Background:

• Phylogenetic analysis is crucial for understanding gene evolution.
• Close evolutionary relationships, like between humans and chimpanzees, complicate detecting human-specific positive selection.
• Maximum-likelihood (ML) analysis of codon substitution patterns is a common method.

Purpose of the Study:

• To evaluate the effectiveness and limitations of ML analysis for detecting human-specific positive selection in recently diverged species.
• To assess the impact of evolutionary model parameters (branch length, d(N)/d(S) ratio) on ML analysis.
• To develop a more sensitive method for detecting lineage-specific positive selection.

Main Methods:

• Simulations were performed using parameters derived from primate gene alignments.
• ML analysis was applied to codon substitution patterns.
• An alternative empirical null model was developed using simulated neutral evolution data.

Main Results:

• The standard ML method showed limited power to detect positive selection since the human-chimpanzee common ancestor.
• Varying branch length and d(N)/d(S) ratios impacted ML analysis performance.
• The developed empirical test demonstrated increased sensitivity for detecting lineage-specific positive selection.

Conclusions:

• Standard ML methods are insufficient for reliably detecting recent human-specific positive selection.
• An empirical, gene-specific null model enhances the detection of positive selection in closely related species.
• This improved approach is vital for understanding recent human evolution and gene adaptation.