PHYRN: a robust method for phylogenetic analysis of highly divergent sequences

Area of Science:

• Computational Biology
• Evolutionary Biology
• Bioinformatics

Background:

• Phylogenetic analysis is challenging for proteins with low sequence similarity (≤ 25% identity), often termed the "twilight zone."
• Existing multiple sequence alignment (MSA) and tree estimation methods struggle with extreme sequence divergence.

Purpose of the Study:

• To evaluate the accuracy of phylogenetic inference methods at extreme sequence divergence.
• To introduce and assess a novel MSA-independent method, PHYRN, for phylogenetic analysis.

Main Methods:

• Simulated data sets were generated to represent varying degrees of sequence divergence.
• Four leading MSA methods (MAFFT, T-COFFEE, CLUSTAL, MUSCLE) and six tree estimation programs were tested.
• The novel PHYRN method was developed and compared against traditional approaches.

Main Results:

• PHYRN demonstrated superior performance in reconstructing high-resolution phylogenies at extreme genetic distances (e.g., ~7% pairwise identity).
• PHYRN outperformed traditional MSA-based methods, particularly in the "midnight zone" of divergence.
• The PHYRN algorithm successfully inferred deep evolutionary relationships within the DANGER protein superfamily.

Conclusions:

• PHYRN is a powerful tool for analyzing highly divergent protein sequence data, overcoming limitations of traditional phylogenetic methods.
• The PHYRN method enables exploration of evolutionary relationships in previously intractable datasets.
• This work advances the field of phylogenetics for extreme sequence divergence.