search

DerrickWood / kraken2

The second version of the Kraken taxonomic sequence classification system
MIT License
687 stars 266 forks source link

Does Kraken2 deal with chimeras #761

Open alanxelena opened 9 months ago

alanxelena commented 9 months ago

Hi! Sorry if this is a rather silly question but I've been reading and couldn't find an answer on this one. I've been using Kraken2-Bracken on metagenomes and it runs great. I wanted to test the pipeline on a set of 16S short reads and compare them to my current workflow (dada2), in both cases I'm using the Silva database to assign taxonomies. I noted that there are differences in both outputs but the most abundant genera remain roughly the same, I guess this has to do with the different approaches. If I now check the least abundant ASVs there are far more genera assigned with Kraken than with dada2, I thought this could be related to dada2's chimera removal step, and now I'm not sure wether Kraken2 deals with chimeras at all. By the way I also see a big difference in the assigned ASVs (299 for dada2 vs. 1353 for Kraken2)

Many many thanks for any insight.

EDIT: I'm also using the default --confidence as I'm a little confused as to what it really means. Should I set it to 0.5?

AlexandreThibodeauUdM commented 8 months ago

Hi! I am just starting to explore 16s classification using Kraken2, but my feeling is that this is a classification tool, so you need to run the normal 16S pipeline and somehow get a fastafile containing all reads (not just unique) for each sample seperately, and then run kraken2 on these, run bracken and then start your analysis. I do not think you can run Kraken2 out of the box on 16s data. Just my 2 cents.