deSAMBA

deSAMBA classifies noisy long reads from metagenomic samples using a de Bruijn graph-based sparse approximate match block (SAMB) pseudo-alignment approach to assign taxonomic origins.

Key Features:

• Sparse Approximate Match Block (SAMB) pseudo-alignment: Identifies sparse approximate match blocks within reads to enable pseudo-alignment without requiring full sequence alignment.
• De Bruijn graph-based k-mer representation: Leverages de Bruijn graph principles by breaking reads into k-mers to manage sequencing complexity and support mapping and classification.
• Optimization for noisy long reads: Specifically handles long read sequencing technologies and their high error rates to improve classification of metagenomic reads.
• Performance on real datasets: Benchmarked on real datasets showing improved processing speed and classification yield relative to existing state-of-the-art tools.

Scientific Applications:

• Metagenomic taxonomic classification: Assigns taxonomic origins to long reads in metagenomic studies to profile microbial community composition.
• Environmental and clinical microbiome analysis: Supports environmental monitoring and clinical diagnostics by enabling rapid and precise classification of microbial communities from long-read data.

Methodology:

Breaks reads into k-mers to leverage de Bruijn graph structures and applies the sparse approximate match block (SAMB) algorithm to identify approximate matching regions for pseudo-alignment-based classification.