Problems with Taxonomic Assignment

General Discussion
1

I am working on a microbiome project with fungal ITS data and I have generated a FASTA file containing my ASV sequences. I need to perform taxonomic classification for these ASVs.

My FASTA file it contains approximately 13,000 ASV sequences.

I am facing several challenges:

  1. Limited Computing Resources: I do not have access to a powerful computer to run local BLAST or QIIME2 installations.
  2. Galaxy Project Classifier Issue: I tried using the feature-classifier classify-sklearn tool on Galaxy Project, but I could not find or upload the correct UNITE ITS classifier (.qza file). The available classifiers were for 16S data (Silva/Greengenes).
  3. DADA2 UNITE Database Issue: I also tried to use dada2::assignTaxonomy with the UNITE database, but I was unable to find the corresponding taxonomy .txt file after downloading and extracting the .tgz package from the UNITE website (it only contained .fasta files).
  4. Online Tool Issues: I've tried other online tools (like the old RDP Classifier and KASS/KEGG BLAST), but they either did not work or were too complex for my current setup/data volume.

My Goal: I urgently need to classify my fungal ITS ASVs using the UNITE database to generate a taxonomy table.

Could anyone please help me with:

  • Providing a direct and working download link for the UNITE ITS classifier (.qza file) that is compatible with QIIME2 and can be uploaded to Galaxy Project?
  • Suggesting an alternative, user-friendly online service that can perform ITS fungal taxonomic classification from a FASTA file?
  • Any other advice or solution that could help me get my ASVs classified?

I am attaching my ASVs_from_seqtab.fasta file (https://drive.google.com/file/d/15qKapduvjVwtkLPXUfpkEw0D_eISwava/view?usp=drive_link).

Thank you very much for your time and assistance.

Amalia

2

Hello Amalia,

Welcome to the forums! :qiime2:

  • Providing a direct and working download link for the UNITE ITS classifier (.qza file) that is compatible with QIIME2 and can be uploaded to Galaxy Project?

Yes! I have made these you can get here: Releases · colinbrislawn/unite-train · GitHub

I would investigate this too! Fungi can be very diverse, but also 13k features is a lot!
Perhaps a barcode is sneaking into the sequences and trimming off that barcode when running DADA2 could reduce duplicate ASVs?

1 Like