The potential of single molecule detection as an analysis tool in biological and medical fields is well recognized today. This fast evolving technique will provide fundamental sensitivity to pick up individual pathogen molecules, and therefore contribute to a more accurate diagnosis and a better chance for a complete cure. Many studies are being carried out to successfully apply this technique in real screening fields.

In this dissertation, several attempts are shown that have been made to test and refine the application of the single molecule technique as a clinical screening method. A basic applicability was tested with a 100% target content sample, using electrophoretic mobility and multiple colors as identification tools. Both electrophorestic and spectral information of individual molecule were collected within a second, while the molecule travels along the flow in a capillary. Insertion of a transmission grating made the recording of the whole spectrum of a dye-stained molecule possible without adding complicated instrumental components. Collecting two kinds of information simultaneously and combining them allowed more thorough identification, up to 98.8% accuracy.

Probing mRNA molecules with fluorescently labeled cDNA via hybridization was also carried out. The spectral differences among target, probe, and hybrid were interpreted in terms of dispersion distances after transmission grating, and used for the identification of each molecule. The probes were designed to have the least background when they are free, but have strong fluorescence after hybridization via fluorescence resonance energy transfer.

The mRNA-cDNA hybrids were further imaged in whole blood, plasma, and saliva, to test how far a crude preparation can be tolerated. Imaging was possible with up to 50% of clear bio-matrix contents, suggesting a simple lysis and dilution would be sufficient for imaging for some cells.

Real pathogen DNA of human papillomavirus (HPV) type-16 in human genomic DNA was probed with fluorescently-labeled probe molecules and imaged. When only the probes were stained and hybridized in a vial, it had 6 orders of magnitude dynamic range with a detection limit of ∼0.7 copy/cell. A second dye was added to lower the false positive levels. Although there was a sacrifice of two orders of magnitude in detection limit, the number of false positives was reduced to zero.

HPV-16 DNA was also hybridized and detected on surface-tethered probes. When the entire human genomic DNA and HPV was labeled and hybridized, the detection limit was similar to that of one-color assay detected in capillary. However, non-specific adsorption was high, and the dynamic range was narrow because of saturation of the surface and electrostatic repulsion between hybridized targets on the surface. The second probe was introduced to lower non-specific adsorption, and the strategy succeeded in 4 orders of magnitude linear dynamic range in a log-log plot, along with 2.4 copies/cell detection limit.

DNA extracts of cell lines that contained a known copy number of HPV-16 DNA were tested with the four strategies described above. The calculated numbers from observed molecule counts matched the known values. Results from the Pap test sample with added HPV DNA were similar to those of purified DNA, suggesting our method is compatible with the conventional Pap test sample collection method.

Further optimization will be needed before this single molecule level detection and identification can actually be used in a real clinical lab, but it has good potential and applicability. Improvement such as automated imaging and scanning, more accurate data processing software as well as sensitive camera, should help increase the efficiency and throughput.