StatAlign

StatAlign integrates structural information with Bayesian inference to co-estimate multiple alignments, phylogenetic trees, and evolutionary parameters for protein, DNA, and RNA sequence and structure evolution.

Key Features:

• Integration of Structural Information: Incorporates structural information into sequence analysis to improve alignment and phylogenetic inference for highly divergent sequences.
• Stochastic Model of Pairwise Structural Evolution: Extends a stochastic model to multiple structures on a phylogenetic tree and analytically integrates over ancestral structures for joint sequence-structure likelihood computation.
• Reduction in Errors: Reduces topology and alignment errors and can alter consensus topologies by providing structural signals beyond sequence data alone.
• Phylogenetic Stabilization: Stabilizes phylogenetic inference by favoring specific topologies with higher posterior probability compared to sequence-only inferences.
• Branch-Specific Evolutionary Rates: Infers branch-specific rates of structural evolution to reveal rate variation across the tree.
• Improved Fit to Empirical Data: Allows structural divergence rates to vary over the phylogeny to achieve a better fit to observed pairwise RMSD values.

Scientific Applications:

• Phylogenetic Analysis: Produces more reliable phylogenetic trees by integrating structural data, particularly for highly divergent sequences.
• Evolutionary Studies: Supports analysis of evolutionary relationships and divergence patterns among proteins such as cytoglobins, myoglobins, and hemoglobins.
• Homology Identification: Identifies homologous regions in highly divergent sequences by leveraging conserved protein structures.

Methodology:

StatAlign uses a Bayesian framework with Markov Chain Monte Carlo to co-estimate multiple sequence alignments, phylogenetic trees, and evolutionary parameters; structural information is incorporated via the StructAlign plugin implementing a stochastic model of structural evolution on a phylogenetic tree with analytical integration over ancestral structures to enable efficient joint sequence-structure likelihood computation and inference of branch-specific structural evolution rates.