ARPIP: Ancestral Sequence Reconstruction with Insertions and Deletions under the Poisson Indel Process

Area of Science:

• Evolutionary Biology
• Computational Biology
• Bioinformatics

Background:

• Phylogenetic methods infer ancestral molecular sequences to understand evolutionary history, gene function, and biological processes like adaptation.
• Existing methods often lack explicit models for insertions and deletions (indels), limiting biological interpretation.

Purpose of the Study:

• To propose a novel dynamic programming algorithm, ARPIP (Ancestral Reconstruction under Poisson Indel Process), for fast, joint, likelihood-based reconstruction of ancestral sequences.
• To incorporate an explicit model for indels, providing biologically interpretable events.

Main Methods:

• Developed a two-step dynamic programming algorithm: 1) identifying most probable indel points and pruning the phylogeny, and 2) inferring ancestral states on the pruned subtree.
• Utilized the Poisson Indel Process (PIP) for modeling indel events.
• Achieved linear time complexity for likelihood computation with respect to the number of sequences.

Main Results:

• ARPIP accurately reconstructs both indel events and substitutions on simulated and real Betacoronavirus data.
• The method demonstrates high accuracy and performs competitively against established tools like FastML and PAML.
• Inferred indel events possess explicit biological interpretations.

Conclusions:

• ARPIP offers a significant advancement in ancestral sequence reconstruction by integrating an explicit indel model.
• The method's accuracy, efficiency, and interpretability expand the scope of evolutionary studies.
• Potential for extensions includes exploring optimal/suboptimal reconstructions and incorporating rate heterogeneity.