Kraken 2 Databases

As of September 2020, all available Kraken2 indices will be hosted on the following Amazon Web Services site: https://github.com/BenLangmead/aws-indexes. This website will be frequently updated with the most widely used Kraken2 indices.

As of January 2026, a new Kraken 2 database is available, tentatively called PrackenDB. This contains all NCBI reference^* assemblies (Genbank and RefSeq) of bacteria, archaea, protists, and fungi as of October 7, 2025. Also includes the human genome, RefSeq viral genomes, and UniVec Core. A key difference from other Kraken DBs is that it has only a single reference genome per species (with a couple of exceptions such as normal and pathogenic E. coli), which is useful for methods that count k-mers per species. This database is on the Amazon cloud here: https://genome-idx.s3.amazonaws.com/kraken/k2_NCBI_reference_20251007.tar.gz

^*NCBI uses the term "reference" genome to mean "representative," which they define as "the best available genome for the species."

Kraken 2 Code Download

• Source (99 KB): Kraken2's source code and README. The current official version of Kraken 2 is v2.0.8-beta (released April 25th, 2019). However, additional fixes/changes are in the newest github version.
• The source code for the current release and the newest version are both available at the Kraken 2 GitHub repository.

Pre-built Kraken 2 Databases

These additional databases have been provided by non-CCB labs:

• GTDB_r89_54k linkA collection of database files for use with Centrifuge, Kraken 1, or Kraken 2 that can be used to classify metagenomes using the GTDB_389_54k index. More information and details at: https://github.com/rrwick/Metagenomics-Index-Correction
• Maxikraken2 and Kraken2-microbial databases. These databases are maintained by LomanLab. More information at the link provided.