Are they the same?-"taxonomy.qza" from alignment and phylogeny building and " taxonomy_rep-seqs.qza" fromTaxonomy classification and taxonomic analyses

Hi, I am new to qiime2, currently I am trying to use qiime2 to do taxonomy analysis for my 16S sequence data.
I used two methods respectively from the instruction of qiime2 overview flow chart,
After join, quality control and de-noise with deblur for my sequences, I go through the steps in section " Sequence alignment and phylogeny building" as following,
qiime fragment-insertion sepp \
--i-representative-sequences rep-seqs_D35.qza \
--o-tree insertion-tree.qza \
--o-placements insertion-placements.qza

qiime fragment-insertion filter-features \
--i-table table_D35.qza \
--i-tree insertion-tree.qza \
--o-filtered-table filtered_table_D35.qza \
--o-removed-table removed_table_D35.qza

fragment-insertion classify-otus-experimental \
--i-representative-sequences rep-seqs_D35.qza \
--i-tree insertion-tree.qza \
--o-classification taxonomy.qza

Then I go through steps in section " Taxonomy classification and taxonomic analyses" as following:
qiime feature-classifier classify-sklearn \
--i-classifier ../reference/gg_13_8_otus/classifier.qza \
--i-reads Clean_Data_D35_joined_filtered_rep-seqs.qza \
--o-classification taxonomy_Clean_Data_D35_joined_filtered_rep-seqs.qza

metadata tabulate \
--m-input-file taxonomy_Clean_Data_D35_joined_filtered_rep-seqs.qza \
--o-visualization taxonomy_Clean_Data_D35_joined_filtered_rep-seqs.qzv

my question are the taxonomy results from both sections the same?, i.e is "taxonomy.qza" the same as "taxonomy_Clean_Data_D35_joined_filtered_rep-seqs.qza"?(looks they are very similar!)
If they are the same, that means two sections are just different methods for taxonomy assignment?
1 Like

Good morning zong,

Welcome to the forums! :qiime2:

Yes! They both include a taxonomy classification table for each of your features.

That’s correct! Note that fragment-insertion classify-otus-experimental is experimental, just like it says in the name, so you might want to use an established classification method for publication.


1 Like

@colinbrislawn Thank you very much, Colin!

Hi Colin, what does experimental mean? based on the wet lab experiment or based on some experiment census?
Another question regarding phylogenic tree building, I am doing chicken gut microbiota analysis,
which approach I should choose, "reference-based [fragment insertion] approach"? or "
de novo approach"? what is the criteria for making decision for phylogenic tree building approach?


Here, I think “experimental” means “unconventional” or “prototype” or “work in progress”. The fragment-insertion plugin includes validated methods like sepp, it also includes this new, experimental method for classifying otus.

(Once the new method has been validated, it will probably be called classify-otus)

Good question! There is evidence that reference-based [fragment insertion] approach works better than denovo tree building (especially for short reads). There is a good example of using sepp in this part of the Parkinson’s Mouse Tutorial.

:evergreen_tree: :deciduous_tree: :christmas_tree:


1 Like

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