Classify Whole Metagenome Data

Classify Whole Metagenome Data performs taxonomic classification of whole metagenome sequencing data using the same principles as the open-source tools Kraken2 [Wood et al., 2019] and Bracken [Lu et al., 2017]. This tool uses a fast and memory-efficient algorithm that facilitates analysis with large, comprehensive reference indexes, enabling accurate species-level resolution on laptops.

Each read is classified to the most specific taxonomic level. If a read match equally well to multiple references, it may be assigned to a higher taxonomic level. Reads that do not match the database are marked as unclassified.

The number of reads assigned to each taxonomic level is used to create an abundance table, excluding any taxonomy with fewer than 10 reads. Some reads may have matched higher taxonomic levels. These reads are probabilistically reassigned to species level using Bayes' theorem, considering the likelihood that a read from a certain species may be redundant and therefore assigned to a higher taxonomic level.


Subsections