Dusting off my QIIME knowledge......trying to get to work in Galaxy

Greetings QIIME2 community!

It's been a while. I previously did a lot of QIIME analyses circa 2011-2013ish with things setup on my desktop at work during my postdoc. However, I ended up taking a position at a school where there was little time to do the biocomputing I enjoyed previously. Any research I was able to do was limited, and we relied on summary output from sequencing facilities to summarize results. However, now I am at a different institution in a different position that involves teaching and allows for more research, and I am trying to get back into the swing of things (although feeling WAY behind and overwhelmed in terms of the pipeline).

Currently, I have some sequencing data (both raw and some analysis output files) from a sequencing facility and was hoping to analyze it in QIIME2. The sequencing facility has demultiplexed, removed primers, filtered, dereplicated, and denoised the data and generated a .fa file with the zOTUs. They also did a taxonomic analysis and produced files showing relative abundance data and counts at different taxonomic levels. If possible, I would like to use any of these output files rather than start from scratch (unless I really need to) and go a bit further in the analysis.

Specifically, I would like to do a few things:

  1. Generate some diversity metrics to use to compare among samples (I sequenced 4 replicate samples from each of 4 species)
  2. Create a phylogenetic tree or do PCoA to visualize differences/similarities among the samples
  3. Do an ANOSIM to test for significant differences in community composition among the different samples

I have created an account through the Galaxy server and was hoping to use it to run my analyses, since my current institution does not allow remote access. I found the Alpha and Beta Diversity tutorial and was going to follow it, but I am having a hard time getting myself oriented. It mentions using a table.qza file and a rooted-tree.qza file that was generated in a "Moving Pictures" tutorial. But, I was not sure how this fit in with the files I have or how to generate them with the files I have. None of my files are .qza (I wasn't really sure what program generates that file type). I tried to find out how to generate the feature table (I think this is the table.qza file mentioned), but it looks like that is generated during denoising, and although mine are denoised, I am not sure I have that.

This is a listing of the pipeline files provided by the sequencer with their description:
Within the pipeline folder are 6 different files, which are generated through various stages of our analysis pipeline.

  • -pr.fastq: Demultiplexed dataset stripped of forward and reverse primer sequences in fastq format.
  • -filtered.fa: Reads are filtered based on Q score and expected error probability and any read with a number of expected errors greater than 1.0 are discarded.
  • -uniques.fa: Dereplicated quality filtered reads
  • -zmap.txt: zOTU read map
  • -zotus.fa: Denoised unique sequences; reads with sequencing or PCR errors are removed followed by the removal of chimeras
  • -zotutab.txt: A zOTU table created with the number of reads assigned to each zOTU in each sample. All reads, pre and post filtering are considered for zOTU table construction.

Can someone please point me in the right direction as to how to take the pipeline files I have and begin doing the diversity analyses I am interested in? I believe once I can get my mind around just integrating my files into the pipeline that I will be okay.

Apologies for my deficiencies. I have since initially using QIIME had 2 kids (one of which is only 2yo) and survived a stressful enrollment decline and ultimately job loss at the teaching institution and am trying my best to pick myself up with whatever little energy I have left after a 3hr round trip commute each day and get back in the groove of things.

Thank you QIIME2 community in advance!


Hello and welcome back to the Qiime2 community!

At that time I had no idea about Qiime at all and even could not yet speak any English...

I think you will catch up very fast!

If you trust the company that performed the analyses and data processing, then you can import the necessary files to proceed with other analyses that you are planning to do. In my experience, sequencing companies do not always follow the best practices and implement their own workflows that are... too common and generalized and not always fit for the data.

Yes, there are plugins in Qiime2 that can perform tasks you described.

If you have not so many samples then many of the tasks also can be performed locally. You can install qiime2 on Linux, Mac or Windows (WSL).

So, feature table is a table with your features (ASVs or OTUs) as rows and samples as columns with features counts as values. You can recreate it from raw data in Qiime2 or import your "zotutab.txt" file (tab separated, convert first to the biom and then import biom to the qiime2 format (qza), check out the importing tutorial).
Representative sequences, or rep-seqs.qza artifact contains a fasta file with sequence IDs and sequences themselves. It can be used for tree construction and taxonomy assignment. Looks like "zotus.fa" file can be imported to qiime2 as rep-seqs.qza.
Tree can be constructed in Qiime2 using either imported files or files recreated in qiime2 from raw data.

That can be done in Qiime2 after importing raw reads.

So, you can import files with zOTU counts as feature table and fasta file with zOTU IDs and sequences as representative sequences. You can also use their taxonomy annotations and import it to qiime2.

Another approach (my personal choice!) is to import raw data (fastq.gz) files to Qiime2, double-check that primers are removed, and denoise reads with Dada2. This action will create a feature table and rep-seqs.qza files that can be used for all analyses you plan to do (tree construction, core--metrics for alpha and beta diversity, taxonomy annotation and so on).

Wow, perfect time to teach your kids some qiime2 analyses!
I wish you good luck with your works (raising kids is the most important one!).



Thank you so much, Timur!

I am going to take your comments and try to go from there. I sincerely appreciate your feedback and taking the time to explain to me!

1 Like

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.