The vast majority of microscopic life on earth consists of microbes that do not grow in laboratory culture. To profile the microbial diversity in environmental and clinical samples, we have devised and employed molecular probe technology, which detects and identifies bacteria that do and do not grow in culture. The only requirement is a short sequence of contiguous bases (currently 60 bases) unique to the genome of the organism of interest. The procedure is relatively fast, inexpensive, customizable, robust, and culture independent and uses commercially available reagents and instruments. In this communication, we report improving the specificity of the molecular probes substantially and increasing the complexity of the molecular probe set by over an order of magnitude (>1,200 probes) and introduce a new final readout method based upon Illumina sequencing. In addition, we employed molecular probes to identify the bacteria from vaginal swabs and demonstrate how a deliberate selection of molecular probes can identify less abundant bacteria even in the presence of much more abundant species.
It is now widely appreciated that only a small fraction of the earth's bacteria may be grown in laboratory culture (for example, see reference 1). That statement is particularly true for the bacteria found in and on human beings (2). The Human Microbiome Project employs a metagenomic approach, principally amplifying and sequencing a small portion of the 16S rRNA gene, to identify bacteria. Like any technology, sequencing rRNA genes has strengths and weaknesses. Its principal strengths are that it overcomes the serious limitations of bacterial unculturability and genomic diversity. Its principal weaknesses are that different bacteria have different copy numbers of rRNA genes (3)(4)(5), that the application of rRNA gene "universal" primers has specificity performance limitations for some bacterial species (6-8), that chimeras are formed at a relatively high frequency when rRNA genes are amplified and cloned (for example, see reference 9), and that the presence of a highly abundant bacterium can mask the detection of less abundant species. Microarrays containing probes that hybridize to rRNA genes sequences are also effective metagenomic tools for detecting unculturable bacteria (10), but, as mentioned above, they can be limited in their ability to discern between some species.To overcome these weaknesses, we developed a genome-based strategy that targets single-copy bacterial genome sequences with high specificity and sensitivity. Originally referred to as molecular inversion probes (11), the technology was simplified for highthroughput analysis, and thus, the name was revised simply to "molecular probes." This approach is rapid, scalable, robust, and culture independent (12, 13). This assay requires only a short stretch of contiguous DNA sequence unique to the bacterial genome of interest. That means that probe technology cannot detect novel microbes, i.e., microbes for which genome sequence information is not yet available. In ...