We use a variety of data analysis tools such as OneCodex and Megan6 for quantifying and visualizing metagenomic sequencing datasets.
Kraken2 is a popular taxonomic classification tool for metagenomic and microbiome sequencing results. The original Kraken tool developed by Wood and Salzberg has been improved for classification speed and accuracy.
Fig. 1 Kraken2 Krona chart
Fig. 1 shows a Krona chart view of Kraken2 classification of the human oral microbiome. Krona charts display hierarchical and taxonomic data in pie chart format. In Fig. 1 the highest level Linneaen taxonomic level is shown in the center of the chart, with progressively lower taxonomic categories radiating outwards. Kraken2 can show taxonomic categories down to the subspecies and strain level. The relative abundance of each taxonomic category is identified by the relative size of each pie chart sector and listed as a percentage of assigned reads.
Which Algorithms Does Kraken Use?
Kraken2 uses an internal k-mer to LCA (lowest common ancestor) mapping algorithm to classify sequence reads against user-defined genome databases. The LCA algorithm generates a taxonomic classification tree, assigning sequence reads to an appropriate taxonomic level.
Which Databases Does Kraken Use?
By default Kraken2 includes several prebuilt genomic databases:
Additional databases can be added to this list and a unique set of custom databases can be designed for use with Kraken2. The prebuilt default databases will classify bacterial, archaeal, and viral sequence reads. With database customization, Kraken2 can be used for more targeted classification tasks, say, for identifying only viral sequences. Or its classification scope can be expanded by adding more databases, for example, plasmid sequences, eukaryotic sequences, etc.