Sampling Modification Effects in the Subgingival Microbiome Profile of Healthy Children.

Oral microbiota are considered major players in the development of periodontal diseases. Thorough knowledge of intact subgingival microbiomes is required to elucidate microbial shifts from health to disease. This comparative study investigated the subgingival microbiome of healthy children, possible inter- and intra-individual effects of modified sampling, and basic comparability of subgingival microprints. In five 10-year-old children, biofilm was collected from the upper first premolars and first molars using sterilized, UV-treated paper-points inserted into the subgingival sulcus at eight sites. After supragingival cleaning using an electric toothbrush and water, sampling was performed, firstly, excluding (Mode A) and, secondly, including (Mode B) cleansing with sterile cotton pellets. DNA was extracted from the pooled samples, and primers targeting 16S rRNA hypervariable regions V5 and V6 were used for 454-pyrosequencing. Wilcoxon signed rank test and -test were applied to compare sampling modes. Principal coordinate analysis (PCoA) and average agglomerative hierarchical clustering were calculated with unweighted UniFrac distance matrices. Sample grouping was tested with permutational MANOVA (Adonis). Data filtering and quality control yielded 67,218 sequences with an average sequence length of 243bp (SD 6.52; range 231-255). (2.8-24.6%), (9.2-25.1%), (4.9-50.6%), (16.5-57.4%), and (2.2-17.1%) were the five major phyla found in all samples. Differences in microbial abundances between sampling modes were not evident. High sampling numbers are needed to achieve significance for rare bacterial phyla. Samples taken from one individual using different sampling modes were more similar to each other than to other individuals' samples. PCoA and hierarchical clustering showed a grouping of the paired samples. Permutational MANOVA did not reveal sample grouping by sampling modes ( = 0.914 by = 0.09). A slight modification of sampling mode has minor effects corresponding to a natural variability in the microbiome profiles of healthy children. The inter-individual variability in subgingival microprints is greater than intra-individual differences. Statistical analyses of microbial populations should consider this baseline variability and move beyond mere quantification with input from visual analytics. Comparative results are difficult to summarize as methods for studying huge datasets are still evolving. Advanced approaches are needed for sample size calculations in clinical settings.