phyBWT2

phyBWT2 reconstructs phylogenetic trees directly from raw sequencing data without alignment, assembly, or reference genomes, enabling molecular phylogenetic inference from short reads, contigs, or whole genomes.

Key Features:

• Alignment-, Assembly-, and Reference-Free Methodology: Operates directly on raw sequencing data such as short reads, contigs, or whole genomes without requiring de novo assembly or reference mapping.
• Extended Burrows-Wheeler Transform (eBWT) and Positional Clustering: Leverages the combinatorial properties of the eBWT together with a positional clustering framework to identify significant blocks of longest shared substrings across sequences without fixing a predetermined length.
• Partition Tree Construction: Detects relevant sequence blocks and constructs partition trees to infer phylogenetic relationships while bypassing pairwise sequence comparisons and distance matrices.
• Improved Performance Over phyBWT: Reconstructs phylogenetic trees step-by-step by considering multiple partitions simultaneously, reducing running times while maintaining high-quality tree reconstruction compared to its predecessor.

Scientific Applications:

• Population-level phylogenetics: Infers evolutionary relationships among individuals within a population from raw sequencing data.
• Viral evolution and outbreak analysis: Provides insights into the origins and evolutionary trajectories of viral diseases using sequence data.
• Complex evolutionary trajectory reconstruction: Elucidates complex patterns of sequence evolution by detecting shared substrings and partitioning genomes.
• Cross-data-type sequence analysis: Supports phylogenetic inference from diverse sequencing outputs including short reads, contigs, and whole genomes.

Methodology:

Processes raw sequencing data without alignment, assembly, or reference mapping; uses the eBWT positional clustering framework to detect shared substrings across sequences and constructs partition trees from those detected sequence blocks, enabling step-by-step phylogenetic tree reconstruction by considering multiple partitions simultaneously.