Introduction
Requirements
1. Taxonomic assignment of sequences using Kraken 2 and complete database
We recommend Kraken 2 + all NCBI genomes if possible. Kraken non-redundant database is next-best.
Our in-lab database uses an 803 GB database built using RefSeq V214 genomes (803GB; updated version of the best database described in our paper):
- all bacteria + archaea + viruses
- only complete eukaryotic genomes
- other adapter + plasmid sequences
1.1. Build or acquire Kraken 2 database
1.2. Run Kraken 2
Confidence thresholds
Some more of a note to go here!
2. Separation of sequences into different domains
2.1. Get bacterial, archaeal, and unclassified sequences
2.2. Get eukaryotic sequences
2.2.1. Get all eukaryotic sequences
2.2.2 Get non-metazoan sequences
2.3. Get viral sequences
3. Classify bacteria, archaeal, and unclassified with Kraken 2 + GTDB database
3.1. Run Kraken 2
3.2. Run GeCoCheck
3.3. Filter results based on GeCoCheck output
4. Classify non-metazoan eukaryotes with Kraken 2 + microEuk database
4.1. Run Kraken 2
4.2. Run GeCoCheck
4.3. Filter results based on GeCoCheck output
5. Collate all output
Authors
Authors: Robyn Wright, Dhwani Desai, Gina Filloramo & André Comeau
First created: 2 June 2021
Last updated: 16 October 2026
Based on initial versions by: André Comeau (Metagenomics Standard Operating Procedure v3) and Robyn Wright (Microbiome Helper 2 Annotation of reads with Kraken2 (and taxonomic profiling with Bracken))
