draft tutorial for shotgun metagenomics with QIIME 2

Hi all,
Following on our announcement of the first alpha release of the QIIME 2 Shotgun Metagenomics Distribution, we are starting to develop some alpha (early draft) tutorials. At the moment, we're going to keep these on GitHub and link to specific documents that you can view there. Before too long, we'll be starting to build and host Jupyter Books that cover tutorials for the different QIIME 2 distributions that are available.

So, here's a first iteration of a bring-your-own-data tutorial covering taxonomic annotation of shotgun metagenomics reads and assembled contigs.

We are very interested in feedback on our shotgun metagenomics workflows, so please don't hesitate to reach out - especially if we're doing something that seems wrong, or if you think you have a better approach.

You can see my current mental model of this workflow. We'll be filling in pieces over the next few releases as we work toward a beta release and ultimately a full release.

Like QIIME 2 as a whole, we would love for the Shotgun Metagenomics Distribution to be a community effort. Thanks to all of the contributors so far!

Update: I put some thoughts together regarding host read filtering here.


That is really amazing!
I dreamed about completing the whole shotgun analyses within Qiime2 :person_raising_hand: :joy: :sparkler:


Still did not try metagenomic analyses in shotgun distribution, but I have a question regarding Nanopore sequences - is it possible to use qiime moshpit classify-kraken2 for Nanopore 16S reads?

Hi @timanix ,
We have not tried this yet, but in theory it should work if you build a kraken2 database with full-length 16S rRNA genes as input, as some 16S databases are supported by kraken2. If you are building a custom 16S database this might need to be done with kraken2 directly (outside of QIIME 2) and then imported. But you could try the build-kraken-db action — I think this should work for NCBI refseq data (e.g., pulled with RESCRIPt) but probably would not work with custom databases. But we have not tested this yet... right now those actions were all built for shotgun data so there will probably be issues if using for building a 16S database, but we would be keen to learn what these are and fix them down the road. So please give it a spin and let us know what you find!


Hi @Nicholas_Bokulich
Thank you for the reply.
On the resource page of Kraken2 indexes, I found Silva 132/138 and GG 13.5. I want to test it with the GG2 database to compare Nanopore classification with the GG2 plugin (similar to closed clustering), GG2 with blast and search classifiers, and now Kraken2 classification. For the last one, as I understood, I should build a database outside of qiime2. Can I import the database to qiime2, or do you mean to classify it outside as well and then import taxonomies :thinking:?

Could you please provide a link?

yes you should be able to import their pre-built 16S database as the appropriate type, and then use the kraken2 action in q2-moshpit to classify your sequences.

1 Like

I mean the page from this tutorial: Index zone by BenLangmead

1 Like

An off-topic reply has been split into a new topic: error importing Bracken database

Please keep replies on-topic in the future.

2 off-topic replies have been merged into an existing topic: Quality control for shotgun metagenomics

Please keep replies on-topic in the future.