Converting .qza file to .fasta to upload to MG Rast for publication

I am in the process of uploading 16S microbiome data for a publication and I made a filtered feature table including only the samples included for this manuscript. That file is in .qza format. I need to upload this file to MG Rast and they need it in .fasta format. I was able to convert .qza to .biome, but MG Rast doesn’t recognize that. Is there a way in qiime2 to get the fast file for this table?



If you need to upload a fasta file, then your FeatureTable[Frequency] is the wrong artifact to be looking at — sounds like you need to export your FeatureData[Sequence] artifact.

1 Like

I have somehow the same question. Please explain more about FeatureData artifact.

Best Wishes

So I assume I would need to filter the rep-seqs file based on the metadata and then export that filtered seq set? Is there are filter seqs command like there was a filter feature table command?

For example, for the table I filtered I included only samples related to 2 poultry flocks, using the following:

qiime feature-table filter-samples \

--i-table Desktop/RawSeqData/PasturedPoultry/A2A3B1C1D1E1I1/qiime2/table1.qza \
--m-metadata-file Desktop/RawSeqData/PasturedPoultry/A2A3B1C1D1E1I1/qiime2/PPMappingFileNoYear.txt \
--p-where "FarmFlock IN ('B1', 'D1')" \
--o-filtered-table Desktop/RawSeqData/PasturedPoultry/A2A3B1C1D1E1I1/qiime2/B1D1.qza

I want to do the same for the rep seqs so I can export a fasta file containing only the samples from those 2 flocks.

Yes, it is called filter-seqs. Type qiime feature-table filter-seqs --help into your terminal for more details.

I do not know exactly what mgrast expects, but I doubt that is what you really want to upload. The rep seqs are the unique ASV sequences themselves, but contain no abundance information. I assume mgrast needs the raw data, e.g., before denoising and with some abundance information (i.e., replicate sequences in each sample). If so, I recommend just exporting your data after demultiplexing to get a directory containing a fastq file for each sample. You can convert that to fasta (outside of QIIME 2) and upload to mgrast.

Thanks! I followed your advice at the end about exporting demux, and I think that is the route to go for MGRast.

Thanks again



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