How to build a custom Kraken database?

Prepare the Environment

• Download and install Kraken software from the official GitHub repository or your package manager. Ensure it's the latest version.

 

• Install necessary dependencies, including `jellyfish` and `kraken-build`. Verify all installations by checking their versions with command-line tools.

 

• Create a dedicated directory for the database; this ensures organized storage structure and easy access.

Collect Reference Sequences

• Determine which genomes or sequences you need based on your research or application needs. The NCBI or EMBL-EBI databases are comprehensive sources.

 

• Download the selected genomic data in FASTA format. Consider using command-line tools like `rsync` or `wget` for bulk downloads.

 

• Organize downloaded files into appropriate subdirectories within your custom directory.

Prepare and Format Sequences

• Ensure all downloaded FASTA files have consistent naming conventions without spaces or special characters.

 

• Use quality-check tools to validate the integrity and format of your sequence files.

 

• Optionally, use scripts to preprocess sequences, such as trimming or removing low-quality reads if necessary.

Build the Kraken Database

• Navigate to your working directory in the command-line interface.

 

• Execute `kraken2-build --download-taxonomy --db /path/to/database` to download the latest taxonomy data.

 

• Incorporate your downloaded sequence files using `kraken2-build --add-to-library /path/to/sequences --db /path/to/database` for each organism or set.

 

• Construct the database by running `kraken2-build --build --db /path/to/database`. Monitor the output for errors or warnings.

Verify the Database

• Check the log files created in your database directory for any issues during the build process.

 

• Test the new database with a few sample sequences to confirm accurate assignments using the `kraken2 --db /path/to/database /path/to/sample` command.

 

• Ensure consistency and accuracy by comparing test outputs against known reference datasets.

Maintain the Database

• Regularly update your taxonomy and sequence files using the `kraken2-build --update --db /path/to/database` command to incorporate new data.

 

• Back up your database and related files periodically to avoid data loss from system failures or corruption.

 

• Document all changes or updates to the database structure for future reference and reproducibility of results.