Reclassify my data with a custom database

[jspence]

I received 16S sequence data from a sequencer provider who ran the raw sequences through their quality control, and provided me with the resulting files:

align.txt = UTAX classification alignments

tax.tsv = UTAX classification table

otu.fasta = final OTU centroid sequences (representative OTUs)

otu.fasta.obs = final OTU centroid observations per sample table

contam.otu.fasta = filtered contaminant centroid sequences

contam.otu.fasta.obs = filtered contaminant centroid obs table

unk.otu.fasta = unclassified OTU centroid sequences

unk.otu.fasta.obs = unclassified OTU obs table

log.txt = summary report

mapping.txt = QIIME mapping file

otu.tsv = OTU table of abundances, including taxonomy

otu.biom = OTU abundance+tax table in BIOM format

msa.fasta = multiple sequence alignment of final centroids

otu.tre = phylogenetic tree of final centroids

core_diversity_analyses/ = folder containing QIIME core diversity analyses output

*I have not been provided with the raw .fasta file containing all my OTUs, or the OTU map generated by pick_otus.py.  

My workflow is as follows:

#Reclassify:

assign_taxonomy.py -i otu.fasta -r ROOT/SILVA_128_QIIME_release/rep_set/rep_set_all/97/97_otus.fasta -t /ROOT/SILVA_128_QIIME_release/taxonomy/taxonomy_all/97/taxonomy_all_levels.txt

#Make a biom file with output from assign_taxonomy.py:

make_otu_table.py -i otu_map.tsv -t /ROOT/uclust_assigned_taxonomy/otu_tax_assignments.txt -o reclassified.biom

#convert biom to .tsv

biom convert -i reclassified.biom -o reclassified.txt --table-type "OTU table" --to-tsv --header-key taxonomy --output-metadata-id "ConsensusLineage"

I am able to successfully reclassify my otu.fasta file using SILVA, however, I have not been provided with the otu_map.tsv file required for make_otu_table.py. I can't use the otu.tsv file my sequencer provided, as it contains the old taxonomic classifications, thus I need to generate a new OTU map file using pick_otus.py. However, I am uncertain how to do this as I am unsure if I can use the otu.fasta file as input for pick_otus.py since it only contains the representative sequences for my data.

Thank you in advance for your help, please let me know if I can provide any more info.

[Colin Brislawn]

Good evening!

Thanks for posting to the Qiime 1 forums. Qiime 2 is being build right now and it's coming along nicely. You might find better support over there:

https://forum.qiime2.org/

To answer your core question, yes, you are correct: you can't pick a new OTU table unless you have the full demultiplexed sequences. But your sequencing center should be able to provide them to you. With all your sequences, you can plug them into Qiime 1 (or Qiime 2) and make a new .biom table with new taxonomy. Both Qiime 1 and 2 provide tutorials on how do this this process. If you are starting now, it's a great time to learn Qiime 2.

Let me know if that helps answer your question.

Colin