anvi-estimate-scg-taxonomy results differ from GTDB-Tk results

108 views
Skip to first unread message

Nastassia Patin

unread,
Apr 21, 2021, 7:37:25 PM4/21/21
to Anvi'o
Hey anvi'o team,

When running MAGs through anvi-estimate-scg-taxonomy (following anvi-run-scg-taxonomy of course) I often find that MAGs have no taxonomy assigned to them at all, even at the domain or phylum level. When I run these same MAGs through GTDB-Tk on KBase, they get taxonomic assignments down to the genus level. Do you know what might be causing this discrepancy? I only work with high-quality MAGs (completeness - 5 X redundancy > 40) and have updated anvi'o to the latest version.

Thanks,
Nastassia

A. Murat Eren

unread,
Apr 21, 2021, 7:56:02 PM4/21/21
to Anvi'o
Interesting. If you use anvi-split to get a contigs database specific to that MAG and send it here, perhaps we can take a look :)
--

A. Murat Eren (Meren) | he/him
http://merenlab.org :: twitter


--
Anvi'o Paper: https://peerj.com/articles/1319/
Project Page: http://merenlab.org/projects/anvio/
Code Repository: https://github.com/meren/anvio
---
You received this message because you are subscribed to the Google Groups "Anvi'o" group.
To unsubscribe from this group and stop receiving emails from it, send an email to anvio+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/anvio/3db4eeaa-b1c3-4dab-b5b2-2bba2b460b64n%40googlegroups.com.

Nastassia Patin

unread,
Apr 21, 2021, 8:23:47 PM4/21/21
to an...@googlegroups.com
Here's one example of a MAG that had this result. I didn't use anvi-split to generate it, though; I used anvi-gen-contigs-database. 

You received this message because you are subscribed to a topic in the Google Groups "Anvi'o" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/anvio/st0n1eA85-o/unsubscribe.
To unsubscribe from this group and all its topics, send an email to anvio+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/anvio/CAK5_LacDvy906rLaROTGT%3Dn1eJRNPz0agcgUYr6K6GQwCBtX-g%40mail.gmail.com.


--

Dr. Nastassia Patin

Postdoctoral Associate

Cooperative Institute for Marine and Atmospheric Studies

University of Miami/NOAA


Address:

AOML & SWFSC

8901 La Jolla Shores Drive

La Jolla, CA

92037

S2C3309-contigs.db

A. Murat Eren

unread,
Apr 21, 2021, 9:10:16 PM4/21/21
to Anvi'o
Hey Nastassia,

You get no taxonomy because none of the ribosomal proteins is similar to anything in GTDB with more than 90% amino acid sequence identity. If you lower it to 80% with,

anvi-run-scg-taxonomy -c S2C3309-contigs.db --min-perc 80

Then the closes taxon is estimated to be the following:

image.png

So it is basically too novel to be quickly associated with anything confidently using quick-and-dirty approaches, and a tree-based approach like GTDB-Tk is essential to have a better idea about its place in the tree of life :)


Best,
--

A. Murat Eren (Meren) | he/him
http://merenlab.org :: twitter

Nastassia Patin

unread,
Apr 21, 2021, 9:16:01 PM4/21/21
to an...@googlegroups.com
Got it, thank you! I didn't realize GTDB-Tk used phylogeny instead of ribosome protein % identity. Makes sense that a lot of them are Planctomycetes then, too. 

Reply all
Reply to author
Forward
0 new messages