Table 1

The taxonomic classification for 16S rRNA gene sequences improves with the addition of custom databases
Taxonomic Level Congruent Classifications (No. sequences) Incongruent across all three training sets Congruent Classifications with HBDB
Kingdom 4,480 0 4,480
Phylum 4,465 0 4,478
Class 4,453 4 4,479
Order 2,579 1,335 4,669
Family 1,870 2,784 4,216
Genus 595 2,552 --*

*HBDB sequences were not taxonomically assigned to genus so this level of taxonomic classification was excluded.

The number of 16S rRNA gene sequences from honey bee guts with identical or completely divergent classifications across three widely used training sets (RDP, Greengenes, SILVA) is shown. As the taxonomic levels become more fine, there is an increase in the discordance/errors in taxonomic placement across all three datasets. The addition of honey bee specific sequences greatly improves the congruence across all datasets (last column).

Newton and Roeselers

Newton and Roeselers BMC Microbiology 2012 12:221   doi:10.1186/1471-2180-12-221

Open Data