The remaining high quality sequences were taxonomically identified using the Classifier tool at a 60% confidence level. The classifier

output was then used for analysis of similarities and difference between herds (Additional files 1, 2, 3, 4, 5). For analysis of the data at the genus level, all genera with fewer than 5 representatives were dropped from the analysis. To identify members of the family Pasteurellaceae and genus Streptococcus DNA/RNA Synthesis inhibitor to the lowest possible phylogenetic level, we obtained all the 138 near full-length type sequences from family Pasteurellaceae and genus Streptococcus from RDP release 10.22 (August 2010). We also added sequence AF486274 (“”Actinobacillus porcitonsillarum”"). Epoxomicin These 139 sequences were aligned by the Infernal aligner [16] trained by RDP [17].

The final reference set contained the region corresponding to the 454 FLX amplicon (E. coli position 578 to 784) sliced from the alignment. To determine the nearest neighbor, the 454 FLX sequences passing the RDP Pyro initial filtering were aligned by the Infernal aligner and the distance between each FLX sequence and reference sequences was calculated. The reference sequence with the closest distance was reported. In case of tie, all the reference sequences were reported. Statistical analysis For the statistical analyses of sequences, we used a 0.03% cutoff value for clustering. This is consistent with previous analyses

of 454 data [18] as well as the historical value frequently used over the past 15 years [19, 20]. Similarly we used this cutoff in evaluating members of family Pasteurellaceae and genus Streptococcus. For comparative statistical analyses, aligned sequences were clustered using the RDP Complete Linkage Clustering Tool and the resulting cluster files were used to calculate Jaccard and Sørensen indices [17]. see more For comparative statistical analyses, aligned sequences were clustered using the RDP Complete Linkage Clustering Tool and the resulting cluster files were used to calculate Jaccard and Sørensen indices [17]. Cluster files were also reformatted with the EstimateS Formatter Tool through the RDP website. Principle component analysis followed by centroid calculations with a 95% confidence limit were performed in R (version 2.10; http://​www.​r-project.​org/​) with Vegan package (http://​vegan.​r-forge.​r-project.​org) using the EstimateS formatted files. Chao 1 was calculated using the cluster files derived from each sample and from merged samples for herds using the RDP Pyrosequencing Pipeline. Simpson’s Diversity index was calculated with MOTHUR [21]. Results Community DNA was isolated from whole Pritelivir tonsil tissue (Pigs A-M) or tonsil brushings (Pigs J-M) as described in Methods. Tonsil tissue samples were collected in spring 2007 from two different herds, and again in spring 2009 from Herd 1.