assigning taxonomy at 97 or 98% similarity


I'm using Qiime 2021.11 version and want to assign taxonomy to some individual sequences files. I performed 16S v3/v4 sequencing through the company GENEWIZ for some bacterial isolates, and besides providing me with the raw forward and reverse reads, they also provided me with 1 consensus sequence filed (already merged and trimmed) for each sample. So starting with this sample I did the following:

qiime tools import --input-path 132_1B.fasta --output-path venom.qza --type 'FeatureData[Sequence]'

qiime feature-classifier classify-consensus-blast --i-query venom.qza --i-reference-reads silva-138-99-seqs.qza --i-reference-taxonomy silva-138-99-tax.qza --o-classification blast-venom.qza

qiime metadata tabulate --m-input-file venom_dada2_taxonomy.qza --o-visualization venom_dada2_taxonomy.qzv

However, I believe that by assigning taxonomy this way I am doing it to 99% similarity, is that correct? If so, how could I do to do to 98 or 97% similarity?
I can get down to genera for most samples but not species and I'm wondering if I could get closer by loosening my similarity restrictions.

Thank you!

1 Like


Unfortunately this is likely due to the limits in the resolution of V3V4 sequencing, where taxonomic annotation is often limited to the genus rather than species level. There are a few techniques that you could try to improve your results a bit though! The simplest is to simply use the classify-sklearn tool in feature classifier, which generally can give slightly better results. This video rom our Youtube channel discusses the different methods in a bit more detail.

The other thing that you can do is use a weighted taxonomic classifier which has been shown to increase accuracy of taxonomic annotation. Here is the paper and a tutorial for the plugin(q2-clawback).

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