kraken2

kraken2 classifies metagenomic sequence reads using exact k-mer matches to assign taxonomic labels for microbial community profiling.

Key Features:

• Taxonomic classification: Assigns taxonomic labels to metagenomic sequence data at read level.
• Exact k-mer matching: Uses exact k-mer matches from query sequences to identify candidate taxa.
• Lowest Common Ancestor (LCA) assignment: Matches each k-mer to the lowest common ancestor of all genomes containing that k-mer to inform classification.
• Memory efficiency: Reduces memory usage by 85% compared to its predecessor, enabling larger reference genomic databases.
• Speed: Increases classification speed approximately five-fold relative to its predecessor.
• Translated search mode: Provides a translated search that considers nucleotide-to-protein translations to increase sensitivity, especially for viral metagenomics.
• Bracken compatibility: Supports downstream use of Bracken (Bayesian Reestimation of Abundance after Classification with KrakEN) to probabilistically reestimate species abundances from read assignments.

Scientific Applications:

• Metagenomic taxonomic profiling: Classification of reads from metagenomic sequence data to characterize microbial community composition.
• Viral metagenomics: Enhanced detection and classification of viral sequences via translated search mode.
• Large-scale reference analyses: Enables use of larger reference genomic datasets for comprehensive metagenomic surveys due to reduced memory footprint and increased speed.
• Species/strain abundance estimation (with Bracken): Enables workflows for estimating species-level abundances when paired with Bracken's Bayesian redistribution of reads.

Methodology:

Performs exact k-mer matching from query sequences, assigns each k-mer to the lowest common ancestor of genomes containing that k-mer, offers a translated search mode that translates nucleotides into protein space, and relies on Bracken to probabilistically redistribute read assignments within the taxonomic tree for abundance estimation.