how to achieve species level based on the silva database with method rdp?

798 views
Skip to first unread message

wei liu

unread,
May 28, 2014, 5:14:28 AM5/28/14
to qiime...@googlegroups.com
Dear everyone,

Now I want to assign taxonomy to my sequence based on the Silva database, my command is:

assign_taxonomy.py -o rdp_silva/ -i rep_set.fna  -r /home/liuwei/database/Silva_111_post/rep_set/97_Silva_111_rep_set.fasta  -t /home/liuwei/database/Silva_111_post/taxonomy/97_Silva_111_taxa_map_RDP_6_levels.txt -m rdp

But this can only achieve to genus level, in the /home/liuwei/database/Silva_111_post/eukaryotes_only/taxonomy_euks/ directory I can find the level 7(species) taxonomy file: 97_Silva_111_taxa_map_RDP_7_levels_euks.txt, however in the corresponding bacteria directory I only can find level 6 (genus) taxonomy file. How can I get a level 7(species) taxonomy file that can be used to annotate my sequence to species level.

If I use this file 97_Silva_111_taxa_map.txt, the command is:

assign_taxonomy.py -o rdp_silva/ -i rep_set.fna  -r /home/liuwei/database/Silva_111_post/rep_set/97_Silva_111_rep_set.fasta  -t /home/liuwei/database/Silva_111_post/taxonomy/97_Silva_111_taxa_map.txt -m rdp

I got this error: ValueError: Because the RDP Classifier operates in a bottom-up manner, each taxonomy assignment in the id-to-taxonomy file must have the same number of ranks.  Detected 6 ranks in the first item of the file, but detected 4 ranks later in the file. Offending taxonomy string: Bacteria; __Verrucomicrobia; __OPB35_soil_group; __uncultured_bacterium

So please give me some advice. How can I solve this error?  Really appreciate your help. Thanks very much.

I am sorry , I still have a problem about the methods (rdp, blast) and the database (greengenes, rdp, silva) that used to do taxonomy annotation, how much will the annotation results  be different between each other, have anyone or any paper assess this topic. If I want to increase my known species's abundance, how can I do, please give me more advice. Thanks.

Tony Walters

unread,
May 29, 2014, 2:11:50 PM5/29/14
to qiime...@googlegroups.com
Hello,

The species level RDP compatible taxonomy mapping doesn't exist, so that's not an option.

You can use uneven taxonomy mapping file 97_Silva_111_taxa_map.txt, with the method blast (-m blast), as blast is less restrictive in terms of the taxonomy levels having even depths.

-Tony


--

---
You received this message because you are subscribed to the Google Groups "Qiime Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qiime-forum...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Wei Liu

unread,
Jun 23, 2014, 4:25:19 AM6/23/14
to qiime...@googlegroups.com
Dear Tony,

Thanks very much. But the species level annotation was not very good,
most of them are unkown. How can we improve this situation, can you
give me some advice.

Really appreciate your help.

Best Regards,
Wei Liu.
> You received this message because you are subscribed to a topic in the
> Google Groups "Qiime Forum" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/qiime-forum/SX44UFbEe2g/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to

Tony Walters

unread,
Jun 23, 2014, 12:42:15 PM6/23/14
to qiime...@googlegroups.com
Hello Wei,

The uncultured/characterized taxa are going to be a problem with any database (our ability to sequence free-living microbes is a bit ahead of the work to characterize all of these organisms).

You could try a newer database from Silva, such as the 115 database (119 isn't out yet, but sometime soon): http://www.arb-silva.de/no_cache/download/archive/release_115/Exports/

You would need to create your own taxonomy mapping files for the newer (than 111) releases though. Mike Robeson has put together some scripts here that could help:

Wei Liu

unread,
Jun 23, 2014, 8:03:45 PM6/23/14
to qiime...@googlegroups.com
Thanks Tony, I will follow your advice, and try the newer Silva
database. Really appreciate your help.
Reply all
Reply to author
Forward
0 new messages