Page History

Select "Build" to create a new database from a genomic library (--build).
Select "Shrink" to shrink an existing database to haveonly specified number of k-mers (--shrink).

Name of the output Kraken database (corresponds to --dbthat is used with --build, and to --new-dbthat is used with --shrink).

Genomes that should be used to build the database.
The genomes should be specified in FASTA format. The sequence IDs must contain either a GI number or a taxonomy ID.

The minimizers serve to keep k-mersthat are adjacent in query sequences close to each other in the database, which allows Kraken to exploit the CPU cache.
Changing the value of the parameter can significantly affect the speed of Kraken, and neither increasing nor decreasing of the value will guarantee faster or slower speed.

To shrink the database before the full build, input the size of the database in Mb (this corresponds to the --max-db-size parameter, but Mb is used instead of Gb).

The size is specified together for the database and the index.

Supply a smaller hash size to Jellyfish, if you encounter problems with allocating enough memory during the build process (--jellyfish-hash-size).
By default, the parameter is not used.

Use multiple threads (--threads). 

To classify single-end (SE) reads or scaffolds, received by reads de novo assembly, set this parameter to "SE reads or scaffolds".
To classify paired-end (PE) reads, set the value to "PE reads".

Use CLARK-l on workstations with limited memory (i.e., "l" for light), this software tool provides precise classification on small metagenomes.

It works with a sparse or ''light'' database (up to 4 GB of RAM) while still performing ultra accurate and fast results.

A path to the folder with the CLARK database files (-D).
It is assumed that "targets.txt" file is located in this folder (the file is passed to the "classify_metagenome.sh" script from the CLARK package via parameter -T).

For example, for 1 (or, 2), the program will discard any discriminative k-mer that appear only once (or, less than twice).

Set the mode of the execution (-m):

• "Full" to get detailed results, confidence scores and other statistics.
• "Default" to get results summary and perform best trade-off between classification speed, accuracy and RAM usage.
• "Express" to get results summary with the highest speed possible.

"Gap" or number of non-overlapping k-mers to pass when creating the database (-п).

Increase the value if it is required to reduce the RAM usage. Note that this will degrade the sensitivity.

Request the loading of database file by memory mapped-file (--ldm).

This option accelerates the loading time but it will require an additional amount of RAM significant.

This option also allows to load the database in multithreaded-task (see also the "Number of threads" parameter).

Use multiple threads for the classification and, with the "Load database into memory" option enabled, for the loading of the database into RAM (-n).

Specify the output file name.

mode

sequencing-reads

Input data

string

tool-variant

Classification tool

number

database

Database

string

k-min-freq

Minimum k-mer

frequency

number

mode

Mode

bool

gap

Gap

number

preload

Load database into memory

bool

threads

Number of threads

number

output-url

Output file

string

tax-data

tax-classification