Schematic representation of protocol for analyses
An average of >50,000 sequences of about 441 bp per sequence are typically obtained in a MiSeq run of about 90 samples. Bad reads are removed from analyses. Species-specific, 16S rRNA-based oligonucleotide “probes”, many of which were originally designed for HOMIM, are used in a BLAST program (called ProbeSeq for HOMINGS) written by Sean Cotton to identify the frequency of oral bacterial targets. At present, 598 oligonucleotide probes of 17 to 40 bases target individual oral bacterial species or, in some cases, a few closely-related species. In order to get nearly complete coverage, an additional panel of 94 genus-specific probes is used.
Outputs are expressed in excel spreadsheets as % frequencies of target taxa (partial output shown below). If desired, arrangements can be made to download raw data, e.g., FASTQ files. QIIME analysis of raw data can be provided at an additional charge, which is dependent upon the size and extent of the project. These data will provide microbial profiles at the genus or higher taxon level.
Partial output of ProbeSeq analysis. Relative levels are indicated by color gradients shown in green.
Charts of the % frequencies are also provided in the form of stacked bar columns (shown below). Sequence identity and % frequency for each bar is revealed using the mouse rollover.
Stacked bar graphs of ProbeSeq analysis. Relative proportions and comparisons of bacterial taxa can be seen at a glance.
For final analyses, % frequency data can be incorporated in any number of statistical packages, such as the R project, MeV, SAS, and QIIME. In a simple example shown below, 16S rRNA profiling data from clinical samples compared 5 periodontally healthy subjects with 5 periodontitis. Data were incorporated into QIIME to produce bubble graphs (PCA plots). These graphs differentiated the healthy profiles from the disease profiles. In addition, the relative abundance of specific species can be indicated by the size of the spot.