Aligning Clustered OTU Sequences to an Outside Reference Database

[ThatGuySam (Sam Flett)]

Hi everyone, and thanks in advance for your help.

I am currently attempting to compare representative sequences from in-house experiments with an outside, publicly available, reference database for coal seam microbiomes. The goal is to identify whether or not bacteria in our samples are potentially common to other coal seam environments. In the example paper I followed, samples were clustered at 97% similarity using UPARSE, and the clustered OTU sequences were aligned to the reference dataset using BlastN. I mimic’d this as closely as possible on QIIME2 2021.4, but am now not even sure that it’s possible.

1. Step one - import .fa file containing rep seqs from ref database
   1. qiime tools import --input-path CSMB_V1-0_sequences.fa \\ --output-path CSMBV1_Seqs.qza \\ --type 'FeatureData[Sequence]'
      1. Successful importing to CSMBV1_Seqs.qza
2. Step two - cluster OTU sequences against reference database at 97% similarity
   1. qiime vsearch cluster-features-closed-reference \\ --i-table table_dada2.qza \\ --i-sequences rep_seqs_Dada2.qza \\ --i-reference-sequences CSMBV1_Seqs.qza \\ --p-perc-identity 0.97 \\ --i-clustered-table table-cr-97.qza \\ --o-clustered-sequences rep-seqs-cr-97.qza \\ --o-unmatched-sequences unmatched-cr-97.qza
   2. This gave me what I assume to be my clustered OTUs aligned to the reference database, condensed into three folders listed as the –output above
3. Here’s where I get stuck. From this, I’ve been trying to follow different BlastN tutorials, namely from More on Taxonomy | Getting Started with Qiime2
   1. qiime feature-classifier classify-consensus-blast \
      --i-query
      --i-reference-reads
      --i-reference-taxonomy
      --p-perc-identity
      --o-classification
      --verbose

      1. I understand everything up until what I’m supposed to be putting in for –i-reference-taxonomy. Originally, I tested the following: qiime feature-classifier classify-consensus-blast \\ --i-query rep_seqs_Dada2.qza \\ --i-reference-reads rep-seqs-cr-97.qza \\ --i-reference-taxonomy table-cr-97.qza (obviously wrong) \\ --p-perc-identity 0.97 \\ --o-classification blast-taxonomy.qza \\ --verbose and received the following error: Invalid value for '--i-reference-taxonomy': Expected an artifact of at

         least type FeatureData[Taxonomy]. An artifact of type TaxonomicClassifier (or FeatureData[Sequence]

         was provided.

      2. Because the database I’m pulling from does not have a file with the reference taxonomy, it makes me feel like I am going down the wrong path.

I am inexperienced with bioinformatics and am hanging on by a thread. Thanks again.

[Nicholas_Bokulich]

Hi @ThatGuySam ,
So you want to see which reference sequences also appear in your dataset?

In that case the closed-reference OTU clustering seems like enough, and classify-consensus-blast is not the right action (as this is specifically for taxonomy classification, which you don't want/cannot do as you do not know the taxonomic annotations of the references)

There is also a blast action that you could use to align against the reference seqs, see qiime feature-classifier blast --help

Good luck!

[ThatGuySam (Sam Flett)]

Running that command gives me the following error:

qiime feature-classifier blast --help

Error: QIIME 2 plugin 'feature-classifier' has no action 'blast'.

Any thoughts? Thanks for the redirection!

[Nicholas_Bokulich]

ah the issue is that you are using QIIME 2 version 2021.4. This action was added in the past few years, so please upgrade to a later version of QIIME 2.