Low-abudance ASVs filtering

Good morning to all,
I have submitted a paper regarding the microbiota analyzed with qiime2, and the referee ask:

did the authors filter out low-abundance ASVs?

I think he ask me to do a filter on the relative frequency.
Can someone help me to understand how I can do it in qiime2?

Thanks in advance

Qiime2 developers have a nice tutorial on this topic, check it out!

Hi @rparadiso,
The tutorial @timanix pointed you is certainly the best way to see the various ways you can filter your table in QIIME 2 based on frequency but just to be aware that currently these tools operate on absolute abundance and not relative abundance, so in order to do filtering based on relative frequency (which you asked in your original inquiry) you’ll have to do some manual calculation. The summary visualization of your feature-table should help with this.

Thank you for your answer…
In the next step now I have an error (clustering of features):

I used this script:
qiime vsearch cluster-features-closed-reference
–i-table table.qza
–i-sequences rep-seqs.qza
–i-reference-sequences 99_otus.qza
–p-perc-identity 0.99
–o-clustered-table table_cr_99.qza
–o-clustered-sequences rep_seqs_cr_99.qza
–o-unmatched-sequences unmatched_cr_99.qza

This is the error:
**Some feature ids are present in sequences, but not in table. The set of features in sequences must be identical to the set of features in table. Feature ids present in sequences but not table are: a1c0e21e3acf5751064c5fb4d26ff1af, 26d9292bedf672200b9b6ef952611310, 011f23cd40aa5e38947c1735e7623b46, 0582638ecb2294334724854cb004da7b, 9cf8a91bf3ad65904f03aa10105701a1, efb793327f6305c0b27a1957be737994, 94f0bc533378bb2219e2f3ad9071b012, bbb096d1ccef0c287633d1243f170ccc, 79700f04454fa8d5aae1ad78e84aa298…)

So I need to remove from the rep_seqs.qza the features.

Can you help me?
thank you so much


Hi @rparadiso,

Could you describe a little bit more what you trying to do here?
The connection with the topic above is not very explicit to me and I wonder if this this would be better a separate topic :thinking:!

Actually, you did apply a filtering step before the ‘qiime vsearch cluster-fuetures-closed-reference’ plug in. But what is not clear to me is at which stage you did it and which command you use.

Going back to the reviewer’s question, I would have interpreted as filtering out low abundance features (as total frequency or relative abundance) from the OTUs table (OTUs because it seems to me you performing a closed-reference approach) before performing alpha- and beta- diversity analysis, without the need to repeat the analysis from scratch (that would be my lazy approach, lacking of more information :grin:)

If you really want go on and filter the rep-seqs.qza to match the table.qza, please have a look at: https://docs.qiime2.org/2020.6/plugins/available/feature-table/filter-seqs/?highlight=filter
and use your rep-seq.qza as input for ‘–i-data’ and your table.qza at ‘–i-table’



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