Infectious diseases pose a growing threat to public health, due to increased rates of population growth, international trade and air travel, climate change, bacterial antibiotic resistance and a wide range of other factors. In addition, global conflicts over the past decade have raised concerns that pathogenic agents might be released deliberately by terrorist organizations or other entities. Public and private funding agencies have responded to these concerns by investing heavily in the development of new assays for microbial surveillance and discovery. The majority of these new methods involve direct detection of microbial nucleic acids. Ideally, these methods should be effective both as detection assays (for identification of known pathogens) and as discovery techniques (for revealing the presence of novel, previously uncharacterized organisms).
Most currently available methods for microbial detection and discovery using nucleic acid samples are based on three technologies. In order of increasing cost, these are the polymerase chain reaction (PCR) [1
], oligonucleotide microarrays [2
] and DNA sequencing [3
]. These platforms have different strengths and weaknesses. While sequencing provides the most in-depth, unbiased information, and is able to reveal completely novel organisms, it can be costly and time-consuming for some applications, particularly when the resources required for data processing and analysis are taken into account. Although multiplex sequencing of bar-coded samples reduces the cost per sample, it also decreases the coverage and thus the sensitivity of the analysis; this may be an issue when the organism of interest has low abundance and the sample has not been treated beforehand to remove host and/or background DNA.
At the other end of the cost spectrum, PCR assays are very fast and sensitive, but have limited capacity for multiplexing [4
]. When an assay is required to test for the presence of several organisms simultaneously, many PCR reactions may be needed, erasing any cost benefit. They are also highly specific; this is an advantage for detecting a microbe whose sequence is precisely known, but a great disadvantage for discovery of novel species, or for detecting variant strains of a known species.
Microarrays occupy a middle ground with respect to cost, processing time, sensitivity, specificity and ability to detect novel organisms. The high-density arrays available at present are able to test for the presence of thousands of different organisms simultaneously, at a cost less than US$ 100 per sample. Arrays can be designed with a combination of high-specificity probes and probes designed against conserved regions, so that they can be used in both detection and discovery modes. While most array designs select probes from fully sequenced genomes in GenBank and other databases, cross-hybridization between probes and similar but non-identical sequences allows detection of novel species, provided that they are closely related to those that were used for probe design. A limitation of microarrays is that, except for so-called universal arrays, probe designs must be updated periodically to include the ever-increasing number of microbial genome sequences being added to GenBank. Nevertheless, for many applications, microarrays offer an ideal balance of capabilities for broad-spectrum microbial surveillance.
A microarray is a miniaturized device containing short (25- to 70-mer) single-stranded DNA oligonucleotide probes (or ‘oligos’) attached to a solid substrate, as shown in . The probes are designed to have sequences complementary to segments of one or more target organism genomes. Oligos may be spotted onto the array by mechanical deposition [2
], sprayed on with a modified inkjet printer head [6
] or synthesized in situ
through a series of photocatalyzed reactions [7
]. Probes are placed on the array in a rectangular grid of ‘features’, each containing many copies of the same oligo. The density of features on the array varies between platforms, from 20
000 spots per slide for a typical spotted array, to several million for platforms such as NimbleGen and Affymetrix that use in situ
synthesized oligos. Arrays may be subdivided with a gasket into subarrays, allowing multiple samples to be tested on one slide. Replicate features, scattered randomly across the array, may be used to allow correction for scratches and spatial effects. On some arrays, negative control probes with random sequences are included, to provide a threshold level for background noise correction.
Schematic view of a microarray, showing single-stranded DNA oligo probes attached to substrate, with fluorescently labeled (green) target DNA strands bound to selected oligos.
To analyze a sample with the array, nucleic acids are extracted and converted to cDNA if necessary (e.g. if the target of interest is viral genomic RNA). The DNA is amplified if needed, fragmented and fluorescently labeled. The labeled DNA is incubated on the array surface for several hours, allowing enough time for the DNA fragments to hybridize to complementary or nearly complementary probes, if they exist on the array. The array is then washed to remove unbound DNA and scanned to produce a file of fluorescence intensities for each feature. In the resulting image, bright features will correspond to probes that are complementary to the DNA in the sample.