AMF analysis, subsetting feature table


I'm analyzing AMF sequences using the 2021.8 release and the maarjAM reference database. I ran vsearch cluster-features-closed-reference to cluster sequences at 97% similarity (the standard for AMF) and got a clustered and identified qza file as well as an unmatched qza file.

I would like to re-cluster the unmatched qza file at a lower similarity (90%) to see if we can get closer to an ID on some of these sequences. However, because the unmactched qza file only contained a portion of the feature-IDs from the dada2 table, I cannot do this.

Is there a way to filter the feature-table based on sequences? It appears that the reverse is possible (filter sequences based on a large variety of inputs) but I have yet to figure out how to filter a feature table.

Thank you.


I think you are looking for feature-table filter-seqs! Just provide your .qza from DADA2 as the metadata file and it should only keep the features in your table that are also found in it.

1 Like

Hi Keegan,

Thanks for the reply, one more question. I got the sequence file to filter this way, but how do I filter the dada2 table? In order to recluster I need the dada2 table filtered as well. When I try to put in the table with the filtered sequences I get the error "Filtering with metadata and filtering with a table mutually exclusive". Is there a way to create a feature table without going through dada2 again?


@Jennifer_Bell for that you will need to use feature-table filter-features.

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