2.2.3. Taxonomic assignment of OTUs

The most commonly used software programs for taxonomic assignment of OTUs are the Ribosomal Database Project’s Naïve Bayesian Classifier (NBC, (Wang et al., 2007)) and the Basic Local Alignment Search Tool (BLAST, (Altschul et al., 1990)).  The database against which sequences are compared, however, appears to be more important than the tool used to assign sequences to taxonomy (Newton and Roeselers, 2012; Sabree et al., 2012). For example, many of the 16S rRNA gene sequences from uncultured and undescribed honey bee gut symbionts have not been incorporated into highly curated databases such as SILVA and greengenes, and use of these databases may result in inconsistent assignment of A. mellifera gut symbionts (Newton and Roeselers, 2012). NBC classifications using custom training sets that explicitly include representatives of A. mellifera gut symbionts will obviate such inconsistencies (Newton and Roeselers, 2012) as will searches against the complete nucleotide collection at NCBI (Sabree et al., 2012). Two of the A. mellifera core gut symbionts, Gillamella apicola and Snodgrasella alvi, have recently been cultured and formally described (Kwong and Moran, 2013), and are presently represented in the greengenes, SILVA, and ribosomal database project alignments.  As formal description of the A. mellifera microbiota expands, the highly curated alignments will more accurately classify these sequences.  Searches against NCBI’s nucleotide collection, however, will continue to provide insight into uncultured sequences that may be closely related to the query sequence and are recommended.