Kraken is a taxanomic sequence classifier that assigns taxonomic labels to short DNA sequences, typically from microbiome or metagenomic studies (1). Kraken classifies reads to a best-match location in a taxonomic tree (the lowest common ancestor), so not all sequences will be classified to a particular level such as species. Kraken matches k-mers (nucleotide sequences of k bases in length) within a read to a database of k-mer sequences from known genomes with established taxonomic relationships to perform its classifications. Partek Flow currently wraps Kraken version 2.0.8-beta (2), but Kraken 0.10.5-beta is also available (to roll back to the older version, see Task Management).
Kraken takes FASTQ files as input. Reads can be single- or paired-end.
Kraken generates a Taxonomic data node. This data can be used as input for the Alpha & Beta diversity and Choose taxonomic level tasks. If you want to obtain the Kraken output files, select the new data node and choose Download data from the toolbox on the right.
Partek distributes the Kraken databases Bacteria, Human, Plasmids, Viruses, and MiniKraken. MiniKraken includes sequences from bacterial, archaeal, and viral genomes in RefSeq; however, it contains only 2.7% of the k-mers from the original database. Running Kraken using the MiniKraken database is significantly faster and less resource-intensive than using a full database, but will not give as complete a result.
If enabled, an Unaligned reads data node including reads that could not be classified by Kraken is produced. Default is Disabled.
If enabled, uses the first hit or hits (--quick). Default is Disabled.
The Kraken task report presents a table and graph summarizing the results (Figure 2)
The table lists each sample and gives the following values:
The number of reads that were classified by Kraken for each sample.
The number of reads that were not classified by Kraken for each sample.
Lists the number of different taxa that were detected within each taxonomic level among the classified reads for each sample. For example, in Figure 2, there were 139 different species detected in Sample 1.
The stacked bar chart shows the abundance or relative abundance of the different phyla. The legend lists the color of each phyla. Mouse over a bar to view the breakdown of families within the phyla (Figure 3). Use the radio button to switch between Absolute abundance and Relative abundance.
Click the in the View column of the table opens an interactive Hierarchical pie chart that can be used to explore the taxonomies present in each sample (Figure 4). Mousing over a section of the pie chart gives the number of reads classified to this level, the number of reads in its children, the percentage of total reads represented by this group, and the percentage of the reads in the root that are represented by this group.
Clicking a section of the pie chart zooms in to that section by setting the selected level as the root (Figure 5).
Click the green arrowto move up one taxonomic level (set the root to one level higher). Click reset to show the entire pie chart.
The mini-map on the upper left is shaded in green to indicate which section of the original pie chart is currently shown (Figure 5).
1. Wood DE, Salzberg SL: Kraken: ultrafast metagenomic sequence classification using exact alignments. Genome Biology 2014, 15:R46.
2. Wood DE, Lu J & Langmead B: Improved metagenomic analysis with Kraken 2. Genome Biol 2019, 20:257