Picking deepest taxonomic node in closed ref otu picking

8 views
Skip to first unread message

Rittik Deb

unread,
May 5, 2016, 7:11:23 AM5/5/16
to Qiime 1 Forum
Hi,

I am running closed or open ref otu picking in qiime on gut microbiome data.

For both the algorithm, it matches my data to a reference dataset (atleast initially for open ref as well). Now what I figured from closed ref otu that in uclust algorithm it picks top 3 best matches with a similarity cut off of 90% (default 0.9 in the code) and assigns a taxonomy using LCA method. This often gives me upto genus level classification without known species ID. Is it possible to change certain parameters to get the deepest assignment for each data (upto species level) and the similarity value for that. When I checked the uclust .uc file it shows the similarity value and the top three hits. Is it possible to only get the deepest node.

One more related query, what happens if among the top three hits two are of one species and the third is of another (or all of them are different species). Does it give g__ i.e. upto genus level taxonomic id.

Thanks in advance for the reply or suggestion.

With warm regards

Rittik

Kyle Bittinger

unread,
May 6, 2016, 4:50:56 PM5/6/16
to Qiime 1 Forum
You probably want to try using the blast method of assign_taxonomy.  You can lower the e-value threshold to filter for more precise matches.
Reply all
Reply to author
Forward
0 new messages