I am very new to qiime 2 and bioinformatics. In short I am working on a pipeline to cluster OTUs in a sample and find out their relative frequency in that sample. To fit in with current protocol for now I will have to cluster to OTUs, as it needs to be comparable to an old method.
I don’t have any actual data at the moment as we want to get the pipeline hypothetically working, and I am using the data from the moving pictures tutorial to make sure.
I am wondering if it is possible to combine the data from feature tables to show this?
I will tackle your main question first, then circle back to the other questions/comments in your post:
It sounds like what you are looking for is feature-table merge. That allows you to merge multiple feature tables into one, so long as those feature tables are already imported into QIIME as FeatureTable[Frequency] artifacts. Check out the other merge-* methods in q2-feature-table to learn how to merge feature data (sequence or taxonomy) files.
I trust you have found this tutorial on using vsearch to cluster to OTUs (I realize that's not your question, just trying to point out relevant docs in case you haven't spotted them)
Perhaps I misunderstand — but you would still be able to compare sequence variants, e.g,. the output of dada2 and deblur, with OTUs. Sequence variants are in a sense OTUs clustered at 100% similarity. (you can also use q2-vsearch to cluster into OTUs following dereplicating/denoising with deblur/dada2, thereby getting the denoising benefits of these methods while still clustering into OTUs).
Just to get a working pipeline any data would do — so the moving pictures data is fine. But for comparing your pipeline to older methods, you may want to consider using mock communities or other test datasets with known composition to assess how well your pipeline performs relative to another "gold standard" method.
Okay, it seems I may have asked the wrong question, however I will definitely be using this method!
Yes it had! Thanks, I managed to execute the method and visualise the table-dn-99.qza file.
I would like to be able to have some sort of table showing each sample, which OTUs it contains and at what relative quantity is this possible?
Unfortunately the relative 'gold standard' is light microscopy... I am hoping to be able to group diatoms sequences within a sample so they can be quantified and 'scored' as per the light microscopy method. Hope this makes a bit ore sense?
Ah now I understand. Yes, absolutely! It sounds like you are just interested in OTUs (not their taxonomic assignments), in which case you may like feature-table heatmap, which will generate a heatmap of OTU abundances in each sample (this can also be used for tables collapsed into taxonomic abundances, as described below).
If you really just want a table listing the abundance (or relative abundance) of each OTU in each sample, you can just export your feature table. This will export your feature table to biom format and you may need to convert to tsv, e.g., to make human readable (some R packages can handle biom data if that's your goal).
If you have taxonomy classifications for those OTUs, you may want to check out taxa barplot which will allow you to plot barplots of the relative abundance of each taxon in each sample, and also download tables of taxa abundance per sample as a tsv. If you want to view taxonomic relative abundance at a specific taxonomic level, convert to relative frequency, collapse at the desired taxonomic level, and proceed with exporting (or the plotting commands) as described above.
Sounds awesome! I had a feeling you were working on something "unusual" (in the QIIME 2 sense). If you have any more difficulty or questions, please let us know.