SuperTriplets: a triplet-based supertree approach to phylogenomics

Area of Science:

• Computational Biology
• Evolutionary Biology
• Bioinformatics

Background:

• Phylogenetic tree-building methods reconstruct evolutionary history using molecular data.
• Reconciling multiple gene phylogenies into a single species phylogeny is a common challenge.
• Existing methods often use binary representations (clades, triplets, quartets) followed by maximum parsimony (MP) analysis, but are limited to small datasets.

Purpose of the Study:

• To introduce SuperTriplets, a novel algorithm for phylogenetic tree reconstruction.
• To optimize an alternative formulation of the maximum parsimony criterion using triplet-based representations.
• To overcome limitations of existing triplet-based binary matrix representations for large datasets.

Main Methods:

• Focuses on the triplet-based representation of source phylogenetic trees.
• Relates parsimony analysis to the median tree notion using this representation.
• Introduces the SuperTriplets algorithm to optimize this formulation.

Main Results:

• SuperTriplets avoids practical limitations of binary matrix representations, enabling analysis of large datasets.
• The algorithm reconstructs the correct phylogeny when triplet resolutions are consistent.
• Simulations and a mammalian phylogenomics case study demonstrate its advantages.
• SuperTriplets produced less resolved but more reliable supertrees than Matrix Representation with Parsimony.

Conclusions:

• SuperTriplets provides a more reliable method for inferring phylogenies from large molecular datasets.
• The algorithm is particularly effective when source trees contain consistent triplet resolutions.
• It offers an improvement over existing maximum parsimony-based supertree methods.