We talk about KRAKEN the taxonomic classification software and the software suite around it and are joined by Jennifer Lu and Natalia Rincon from Johns Hopkins University Center for Computational Biology.

Dr. Jennifer Lu and Natalia Rincon from the Kraken software development team were interviewed on the MicroBinfie podcast. They discussed the various versions of Kraken and the tools developed around it. They began by explaining the original Kraken, which uses an exact camera matching process and a camera size of 31 based on jellyfish. Kraken Unique is an additional version of Kraken that includes an additional column called unique camera counting, which determines how many unique cameras are covered by each read, providing an additional way to verify microbial identification. Kraken two was developed to accommodate larger databases by using a probabilistic data structure and minimizers to map cameras to a shorter sequence size.

They then talked about how Kraken is useful for microbiome analysis, including detecting pathogens. However, the accuracy of the results depends heavily on the availability of genomic data in the database, which emphasizes bacterial and viral data. For infectious pathogen detection, Kraken one unique is combined with Bracken to approximate the abundance of species present.

The developers emphasized the importance of users being aware of available genomic data in the database because the results can only be as accurate as the data. They also talked about how Kraken is used widely in bioinformatics and can be used for various scenarios beyond metagenomics. For example, they use Kraken to treat a single genome as a metagenome as part of quality control analysis. In cases where there are conflicting taxa in the reads, Kraken results show it, making it useful in determining the presence of contamination in samples.

The Kraken team also talked about how they use Kraken for contamination work to detect contamination in pathogen genomes. They compare all eukaryotic pathogen genomes against bacteria, human genomes, and databases of vertebrates and plants to filter out any contaminants. They have found in some instances where contaminating sequences from hosts such as chicken or cow were present in eukaryotic pathogen genomes.

Moving forward, the Kraken team intends to maintain all Kraken repositories, enhance its accuracy, speed, and usefulness, and develop new scripts and downstream analysis for the Kraken Tools suite. They acknowledge the need to make the database smaller as more genomes become available and are exploring ways of indexing and sketching to achieve this.

In conclusion, Kraken has been an essential software for metagenomic analysis, and it remains a continually improving tool for pathogen detection and classification. The Kraken team advises users to keep in mind the importance of accurate data for effective pathogen detection and classification.