Hi @Muslih,
First of, we are working in Qiita to generate a fragment-insertion tree from meta-analyses so you don’t have to do this. We hope to have this ready for the next release (mid to end of March). In the meanwhile, you can run this command:
biom summarize-table --observations -i your_biom.biom | tail -n +16 | awk -F ':' '{print ">"$1"\n"$1}' > rep_seqs.fna
This will be the fna with the rep seqs, which I would suggest using to import and then use with the fragment insertion plugin in Qiime2 to create a tree. Just to be clear, I’m suggesting to use fragment-insertion vs. building a de-novo tree.
Hope this helps.
Cheers,