Sample identification with Kraken

To identify a sample from sequencing reads, we can use the tool “Kraken”. This tool can also be used to identify members in a mixed set of reads, for metagenomics.

e.g. reads from one sample → Kraken → 95% Staphylococcus aureus.
e.g. mixed reads → Kraken → 50% Staphylococcus aureus, 40% Campylobacter concisus, 10% unclassified.

In this tutorial we will use Kraken to confirm the identify of reads from a bacterial isolate.

New to Galaxy? First try the introduction and then learn some key tasks

Get data

Log in to your Galaxy instance (for example, Galaxy Australia, usegalaxy.org.au).

Use shared data

If you are using Galaxy Australia, you can import the data from a shared data library.

In the top menu bar, go to Shared Data.

Click on Data Libraries.
Click on Galaxy Australia Training Material: Sample Identification.
Tick the box next to the four files.
Click the To History button, select As Datasets.
Name a new history and click Import.
In the top menu bar, click Analyze Data.
You should now have four files in your current history.

Or, import from the web

Only follow this step if unable to load the data files from shared data, as described above.

In a new browser tab, go to this webpage:

Right click on a name: select “copy link address”
In Galaxy, go to Get Data and then Upload File
Click Paste/Fetch data
A box will appear: paste in link address
Click Start
Click Close
The file will now appear in the top of your history panel.
Repeat this for all the files on Zenodo.

Shorten file names

Click on the pencil icon next to the file name.
In the centre Galaxy panel, click in the box under Name
Shorten the file name.
Then click Save

You should now have four files in your current history:

galaxy history

Run Kraken

We have a sample that should be Staphylococcus aureus. The paired-end FASTQ read files are:

staph_R1.fq and staph_R2.fq.

(We will look at the other set of files later on in the tutorial).

Go to Tools search bar and type in “Kraken”. Click on the Kraken tool.
Set the following parameters:
- Single or paired reads: Paired
- Forward strand: staph_R1.fq
- Reverse strand: staph_R2.fq
- Select a Kraken database: Minikraken
- leave other settings as they are
- Click Execute

Examine the output

The output is a file called Kraken on data x and x: Classification. This will be at the top of your history pane.

Click Refresh if the file hasn’t yet turned green.

refresh

When the file is green, click on the eye icon to view.

We will turn this output into something easier to read in the next step.
Column 2 is the sequence ID.
Column 3 is the taxon ID (from NCBI).
Column 5 is a summary of all the taxon IDs that each k-mer in the sequence matched to (taxon ID:number of k-mers).

classification

Kraken report

Go to Tools search bar and type in “Kraken-report”. Click on the Kraken-report tool.
Set the following parameters:
- Kraken output: Kraken on data x and x: Classification
- Select a Kraken database: Minikraken
- Click Execute

The output file is called Kraken-report on data x.

Click on the eye icon to view.
Column 1: percentage of reads in the clade/taxon in Column 6
Column 2: number of reads in the clade.
Column 3: number of reads in the clade but not further classified.
Column 4: code indicating the rank of the classification: (U)nclassified, (D)omain, (K)ingdom, (P)hylum, (C)lass, (O)rder, (F)amily, (G)enus, (S)pecies).
Column 5: NCBI taxonomy ID.

kraken report

Approximately 95% of reads were classified as Staphylococcus aureus, confirming the correct identity of our bacterial sample.

Of these reads, roughly half were uniquely present in S. aureus subsp. aureus, and most of those were uniquely present in strain HO 5096 0412.
The sample strain is therefore most related to the HO 5096 0412 strain.

The remaining reads within the S. aureus clade were classified into various taxa.