How to get feature abundance per sample when taxa_barplot specifies 100% unclassified?


I would like to get a table with each feature id and the frequency it occurs in each sample. in qiime I was unable to classify any of the features (they come back almost 100% unassigned in my taxa barplot)

index Unassigned;;;;;; d__Bacteria;p__Cyanobacteria;c__Cyanobacteriia;o__Chloroplast;f__Chloroplast;g__Chloroplast;s__uncultured_bacterium d__Bacteria;p__Proteobacteria;c__Alphaproteobacteria;o__Rickettsiales;f__Mitochondria;g__Mitochondria;s__Saccharina_japonica d__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Oceanospirillales;f__Marinomonadaceae;g__Marinomonas;s__uncultured_marine Sequences Strain Culture_ID Date Site culture barplot_ID
1-18s 64672 0 0 0 1-18S_R1_001.fastq.gz,1-18S_R2_001.fastq.gz QT003 2_L_Flask_stage 1/24/22 Common garden 2_L_Flask_stage 2_L_Flask_stage 1/24/2022
10-18s 70964 0 0 0 10-18S_R1_001.fastq.gz,10-18S_R2_001.fastq.gz QT005 Pre-split_pond_C3 3/1/22 Common garden CommonGarden_Pre-split_pond_QT005 CommonGarden_Pre-split_pond_QT005 3/1/2022
11-18s 60600 0 0 0 11-18S_R1_001.fastq.gz,11-18S_R2_001.fastq.gz QT002 Pre-split_pond_C4 3/1/22 Common garden CommonGarden_Pre-split_pond_QT002 CommonGarden_Pre-split_pond_QT002 3/1/2022

but I have blasted each sequence and gotten results, so I would not like to tie those results to abundance. when I run

qiime feature-table summarize
--i-table uchime-dn-out-18s-trunc/table-nonchimeric.qza
--o-visualization uchime-dn-out-18s-trunc/table-nonchimeric-18s.qzv

I just get a visualization that shows me frequency per sample, but looks like this:


not quite what I'm looking for. I can also get:

Frequency # of Samples Observed In
f8bd3eed5b215a0bd7d09fd5f5ce26e6 168,828 1
dbe8f0eac8b7970b6b082145eb7167d3 124,751 1
bcf19f8c7828812fd03aa7fd575770ec 113,814 2

which has the feature id and frequency in all the samples, but I need frequency PER sample.

I basically want the exact csv output I would get from taxa barplot but with just feature id and abundance. It doesn't help me to just see 'unassigned' and have all those features lumped together.

I hope this makes sense.


Hi @sabitondo,
Probably the easiest way to do this is to transform your table into a relative frequency table using qiime feature-table relative-frequency relative-frequency: Convert to relative frequencies — QIIME 2 2020.2.0 documentation.

Then you will want to use qiime tools extract Exporting data — QIIME 2 2020.2.0 documentation. and go into the data folder. In the data folder there will be a biom table that has the information you want. I would use biom convert, Converting between file formats —, to convert this to a tsv.

This should give you relative abunances for your feature-ids that you can map to your BLAST results.

I hope that helps!


Wow yes this is exactly what I needed! Thank you so much :slight_smile: