APPLES: Scalable Distance-Based Phylogenetic Placement with or without Alignments.

Area of Science:

• Computational Biology
• Phylogenetics
• Bioinformatics

Background:

• Phylogenetic placement is crucial for updating phylogenies and identifying unknown samples via DNA sequencing.
• Existing Maximum Likelihood (ML) methods are computationally intensive and not scalable for large reference trees.
• ML methods require assembled reference and aligned query sequences, limiting their use with unassembled data like in genome skimming.

Purpose of the Study:

• To introduce APPLES, a novel distance-based method for phylogenetic placement.
• To address the scalability and data requirements limitations of current ML-based phylogenetic placement methods.
• To enable phylogenetic placement on large datasets and with unassembled sequencing reads.

Main Methods:

• Developed APPLES, a distance-based phylogenetic placement algorithm.
• Utilized k-mer based distances for placement on unassembled sequence data.
• Benchmarked APPLES against ML methods on varying tree sizes and data types.

Main Results:

• APPLES is an order of magnitude faster and more memory-efficient than ML methods.
• APPLES successfully placed species on reference trees up to 200,000 leaves.
• APPLES demonstrated higher accuracy than ML on dense reference trees and accurately handled unassembled query data.

Conclusions:

• APPLES offers a scalable and efficient solution for phylogenetic placement, especially for large datasets.
• The method's ability to use unassembled data and k-mer distances expands phylogenetic placement applications.
• APPLES enhances the utility of dense taxon sampling in phylogenetic analyses and sample identification.