I recently tried running mothur on CyVerse (Ubuntu command line). I downloaded both mothur (and all its buddy tools) as well as the silva database directly to the site using wget. I.e.:
wget https://github.com/mothur/mothur/releases/download/v1.46.1/Mothur.Ubuntu_20.zip unzip Mothur.Ubuntu_20.zip rm Mothur.Ubuntu_20.zip mv mothur/ Mothur/ mv Mothur/* ./
#Uncompress database file
tar xzvf silva.nr_v138_1.tgz
Everything runs hunky-dory until I get to the classify step, where I try and classify the reads using teh unaltered, untouched silva files:
#Next we need to identify the seqs
classify.seqs(fasta=ABC.trim.contigs.good.unique.pick.fasta, count=ABC.trim.contigs.good.pick.count_table,reference=silva.nr_v138_1.align, taxonomy=silva.nr_v138_1.tax, cutoff=80)
This resulted in several iterations of " ‘XXX’ is in your template file and is not in your taxonomy file. Please correct."
I rootled around online and found some people had had a similar issue and tried the following work around that had been suggested 18S Analysis: Trimmed align file used in classify.seqs is missing taxons · Issue #505 · mothur/mothur · GitHub :
#Fix the broken databases
list.seqs(fasta=silva.nr_v138_1.align) get.seqs(taxonomy=silva.nr_v138_1.tax, accnos=silva.nr_v138_1.accnos)
I also looked for any naughty spaces that may be in the tax file, as suggested by Dr. Schloss in this post classify.seqs errors in reading taxonomy file.
However, upon running the classify.seqs step after that (with my newly made silva.nr_v138_1.pick.tax file), I had the same error that may have actually had more “missing” taxa.
The most puzzling part is, is that I ran this exact set of sequences through mothur (with the same 138 silva release) on both my Mac and on one of our Linux servers without issue. I similarly tried using the databases (tax and align) that I had downloaded to each of those in the Cyverse environment, and alas, the same error. I’m very flummoxed.
Any help on getting this working would be really appreciated!