DendroBLAST: approximate phylogenetic trees in the absence of multiple sequence alignments

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Genome data is rapidly expanding, increasing the need for efficient phylogenetic inference algorithms.
• Existing methods for phylogenetic tree reconstruction have limitations in speed and accuracy.
• Accurate phylogenetic trees are crucial for understanding evolutionary relationships and informing biological research.

Purpose of the Study:

• To introduce DendroBLAST, a novel method for reconstructing phylogenetic trees from protein sequences.
• To evaluate the accuracy and efficiency of DendroBLAST compared to existing phylogenetic methods.
• To demonstrate the utility of DendroBLAST in reconstructing archaeal phylogeny and its potential for downstream bioinformatics.

Main Methods:

• DendroBLAST utilizes the Basic Local Alignment Search Tool (BLAST) for sequence comparison.
• The method incorporates a simple model of sequence evolution to assess the reliability of inferred tree bipartitions.
• Phylogenetic trees were reconstructed using simulated sequence data and applied to archaeal phylogeny.

Main Results:

• DendroBLAST produces more accurate phylogenetic trees than common distance-based methods.
• The method's accuracy is comparable to maximum likelihood methods when using high-quality multiple sequence alignments.
• Application to archaeal phylogeny yielded results with high precision and recall compared to conventional methods.

Conclusions:

• DendroBLAST provides a viable approach for approximate phylogenetic tree reconstruction without requiring multiple sequence alignments.
• The method offers a valuable platform for enhancing and guiding subsequent bioinformatics analyses.
• A web implementation of DendroBLAST is available, promoting accessibility and further research.