Microbial Detection Arrays That Can Be Applied for Virus and Bacterial Detection

Microbial detection arrays provide an assessment of known pathogens, complementary to those provided by functional gene arrays, such as those used to detect the virulence and antibiotic resistance gene families described in. Detection arrays can tell which organisms are present, and functional arrays can tell what abilities those organisms may have. Together, these tools can be used to detect antibiotic-resistant or virulent pathogen variants, natural chimeras, or genetically engineered organisms with unusual gene content.

Detection/discovery microarrays have been shown to identify or discover viruses with homology to known species. Therefore, they can be used to guide the selection of subsets of samples for further analysis by sequencing. Arrays can also be used to study clinical samples for which PCR diagnostics are not informative. Usually, the cause of a clinically serious infection is unknown, complicating decisions about whether to treat with antibiotics, antiviral drugs, or other therapies. Additionally, arrays can aid in the discovery of co-infections with more than one organism. Microbial detection arrays can also be used to examine isolates and vaccines for foreign contaminants.

Finally, arrays can be used to assess the complexity of metagenomic samples to determine the required sequencing depth, potentially saving the cost of low-complexity samples. Compared to sequencing of typical-sized clone libraries, microarrays can reveal greater diversity in complex environmental samples. Microarrays will continue to be a valuable tool until the processing time and cost of high-throughput sequencing, including data analysis, are reduced enough to process large numbers of samples at sufficient depth.

The MDA array design and accompanying analysis algorithms have been found to perform well in identifying mixtures of known pathogens. If the sample contains an organism that has not been sequenced (or whose sequence is not in our analysis database) but is sufficiently similar to other sequenced microorganisms, the analysis will identify multiple related organisms that are most similar to the sample. Similar results are seen when sample DNA is degraded or at low concentrations, so analysis cannot determine the presence of new or unsequenced organisms.

Therefore, users of MDA will need to interpret the data in the context of other known samples to determine whether the predicted organism is an exact match to a known species, or is new but shares some similarities with other sequenced microbes. Highly novel targets that are not similar to probes on the genome or array in the database will not be detected. The failure of MDA to detect VSV NJs in pure culture illustrates this well and highlights the shortcomings of our approach to designing arrays using only the complete genome. Future versions of this array will include all available sequence data, including partial sequences and gene fragments, for species that lack complete genomes. When parts of the genome are included in the database, modified statistical algorithms will be required to handle sequence length bias. Most importantly, as newly discovered or newly sequenced organisms acquire new sequence data, the MDA must be updated with probes to detect them.