After publication, we realized that an error had occurred in the alignment of the 16S rDNA sequences that were produced for McFrederick et al. (2012). Specifically, after the sequences had been aligned, we overlooked a screening step that would have removed sequences that were poorly aligned and had been truncated by a filtering step. This alignment (Supporting information Data S2 in McFrederick et al. 2012) was used in the analysis that resulted in Fig. 3 (McFrederick et al. 2012). To determine if this oversight affected downstream analyses, we screened out truncated sequences, and additionally trimmed the 3 prime end of the alignment. We then reran the downstream analyses, and found that while there were minor differences between the original (Fig. 1, corrigendum) and corrected figures (Fig. 2, corrigendum), the differences were small and did not affect our interpretation of the results. In addition to the alignment that was used in the analysis for Fig. 3, we also generated an alignment with Lactobacillus sequences removed, and further scrutiny of this alignment confirmed that we had correctly performed the post-alignment screening steps for this file. Downstream analyses of the alignment lacking the Lactobacillus sequences were therefore not affected (Fig. 4 in McFrederick et al. 2012). We apologize for this oversight, and refer readers to the corrected FASTA file, which is available as Data S1 (Supporting information) in the online version of this corrigendum.