Hey Aurora,
This makes sense, because
and now you also need to remove it from your DOC-table.qza. Once the sample is removed from both your mapping file and your feature table, this should work fine. Try qiime feature-table filter-samples, and let us know how you...
Oh no! Wait! 
Uh... you may have read about this in other posts, but this might cause big problems 
Don't let me tell you how to filter your table! I just wanted to mention it because I know reviewer three is going to have opinions about filtering negative controls 
wait what?
The Illumina platform doesn't always assign the right barcode to the right read, and the most abundant amplicons are the most likely get a mismatched barcode and end up in other samples. This means that a perfectly empty negative control would still get some reads assigned to it during sequencing that were never in the original sample. If all the ASVs in this negative control were removed, you would lose the most common, real ASVs in the whole data set.