Bayesian Detection of Convergent Rate Changes of Conserved Noncoding Elements on Phylogenetic Trees

Area of Science:

• Evolutionary biology
• Genomics
• Bioinformatics

Background:

• DNA sequence conservation indicates function, and changes in conservation suggest functional shifts.
• Evolutionary rate variation can signal shared functional shifts, potentially revealing genomic underpinnings of phenotypic convergence.
• Current methods struggle to detect complex patterns of rate variation, hindering the identification of convergent evolutionary scenarios.

Purpose of the Study:

• To introduce PhyloAcc, a novel Bayesian method for modeling substitution rate changes in conserved genomic elements.
• To enable accurate detection of complex patterns of rate variation and convergent rate shifts across phylogenetic lineages.
• To provide a tool for identifying genomic correlates of phenotypic convergence.

Main Methods:

• PhyloAcc models substitution rate changes by assigning multiple rate categories to phylogenetic branches.
• It estimates substitution rates per category and identifies rate changes based on the posterior probability of category switches.
• The method was evaluated using simulations and applied to analyze conserved nonexonic elements in cases of convergent evolution.

Main Results:

• Simulations demonstrate PhyloAcc's superior ability to detect rate shifts in multiple species compared to existing methods.
• PhyloAcc shows higher accuracy in reconstructing complex substitution rate change patterns than Bayesian relaxed clock models.
• Analysis of flight loss in birds and marine transition in mammals identified numerous conserved nonexonic elements with lineage-specific accelerations.

Conclusions:

• PhyloAcc is an effective tool for detecting evolutionary rate shifts in conserved elements across phylogenies.
• The method successfully identifies genomic regions associated with convergent phenotypes.
• PhyloAcc is broadly applicable to diverse conserved elements where multiple rate changes are anticipated.