seed-Kraken

seed-Kraken is a modification of Kraken software, a taxonomic sequence classifier that assigns a taxonomic label to short DNA reads. Original Kraken does this by examining the k-mers withing a read and querying a database with those k-mers.

seed-Kraken works with spaced seeds instead of contiguous k-mers, and this turns out to give it some advantage in sensitivity/precision trade off [1].

A spaced seed is a local match triggering an alignment, encoded with a pattern of matches interleaved with spaces – jokers (“don’t care” positions). Below is an example of a spaced seed of weight 28 (non-spaces) and span 40 (total length):

SEED="######-##-#-#-##-###-#-##---###--#######"

You can find more about spaced seeds and why is it worth working with them in section about theory of 'Spaced seeds'.

Information on the original Kraken software can be found at http://ccb.jhu.edu/software/kraken/.

If you use seed-Kraken, please cite our paper:

  1. Spaced seeds improve k-mer based metagenomic classification Karel Brinda, Maciej Sykulski, Gregory Kucherov http://bioinformatics.oxfordjournals.org/content/early/2015/08/10/bioinformatics.btv419