The overall goal of our program is to develop a robust, high throughput, fully automated DNA sequencing instrument based on replaceable polymer solutions using a multicapillary array. Significant effort has already been devoted to column and polymer chemistry in order to obtain long read lengths per run in fast analysis time. In this paper we report on progress in instrument considerations and data processing software. A simple instrument design, based on no moving parts for continuous illumination of the capillaries and detection of the fluorescent light was used for this study. Our polymer solution replacement system with the permanent connection between the buffer/chamber manifold and capillary columns on the detector side is designed to prevent the trapping of air bubbles during matrix solution replacement. A special construction of a column-electrode couple on the injection side precludes air trapping during sample injection from small sample volumes. Our in-house software now features the significant reduction of the crosstalk signal from neighbor columns, which may be a potential problem in densely packed large capillary array sequencers.

A laser-induced fluorescence detection system compatible with a capillary electrophoresis array was developed. The design incorporates fiber-optic excitation and a detection system including a diffraction grating and a CCD camera. The system employs no moving parts and is capable of producing data comparable to commercially available systems. It is based on a spectrally-resolved four-dye sequencing scheme. The conceptual design was proven, however, refinements must be made to optimize performance for high-throughput capillary-array DNA sequencing. Automated sample preparation and loading in combination with a refillable separation- matrix capillary-array system could prove to be an invaluable tool for completion of the Human Genome Project.

So far we have used OSS strategy to sequence over 2 megabases DNA in large-insert clones from regions of human X chromosomes with different characteristic levels of GC content. The method starts by randomly fragmenting a BAC, YAC or PAC to 8-12 kb pieces and subcloning those into lambda phage. Insert-ends of these clones are sequenced and overlapped to create a partial map. Complete sequencing is then done on a minimal tiling path of selected subclones, recursively focusing on those at the edges of contigs to facilitate mergers of clones across the entire target. To reduce manual labor, PCR processes have been adapted to prepare sequencing templates throughout the entire operation. The streamlined process can thus lend itself to further automation. The OSS approach is suitable for large- scale genomic sequencing, providing considerable flexibility in the choice of subclones or regions for more or less intensive sequencing. For example, subclones containing contaminating host cell DNA or cloning vector can be recognized and ignored with minimal sequencing effort; regions overlapping a neighboring clone already sequenced need not be redone; and segments containing tandem repeats or long repetitive sequences can be spotted early on and targeted for additional attention.

This work describes the development of micro-devices for high-throughput DNA sequencing applications. Basically, two research efforts will be discussed; (1) fabrication and characterization of micro-reactors to prepare Sanger chain terminated DNA sequencing fragments on a nanoliter scale and; (2) x-ray photolithography of PMMA substrates for the high aspect ratio preparation of electrophoresis devices. The micro-reactor consisted of a 5'-biotinylated catfish olfactory gene, which was amplified by PCR, and attached to the interior wall of an aminoalkylisilane derivatized fused- silica capillary tube via a streptavidin/biotin linkage. Coverage of the interior capillary wall with biotinylated DNA averaged 77 percent. Stability of the anchored template under pressure and electroosmotic rinsing was favorable, requiring approximately 150 h of continuous rinsing to reduce the coverage by only 50 percent. The capillary micro- reactor was placed inside an air thermocycler to control temperature during Sanger ddNTP chain extension and directly coupled to a capillary separation column filled with a LPA solution via low dead volume capillary interlocks. The complimentary DNA fragments generated in the reactor were heat denatured from the immobilized template and directly injected onto a gel-filled capillary using electropumping for size fractionation and detection using NIR-LIF analysis. The total amount of termination fragments in the 31 nL reactor volume was estimated to be 5.2 X 10¹³ moles and sequencing was shown to produce read lengths on the order to 400 bases. Work will also be described concerning the development of micro-electrophoresis devices in x-ray sensitive photoresists using LIGA techniques. An electrophoresis device with an integrated fluorescence detector was constructed for the high resolution separation of DNA oligonucleotides. The choice of substrate for the electrophoresis was PMMA, due to its intrinsic low electroosmotic flow. Using x-ray lithography in PMMA substrates, the aspect ratios associated with the micromachining was estimated to be > 10,000:1.

Previously, we described a quantitative measure of electrophoretic resolution called resolving power that can be computed for individual bands of DNA separated by electrophoresis. An alternate approach is to determined analytical functions, based on a few experimentally determined parameters, that describe the resolving power of a particular electrophoretic system for al of the length classes of resolving power of a particular electrophoretic system for all of the length classes of molecules that are separated. Such analytical functions have been obtained for single-stranded DNA separated in a polyacrylamide gel and detected at a fixed distance from the origin of electrophoresis. Six experimentally determined constants are required to describe the analytical function for resolving power in this system: four constants describe the mobility of DNA in the gel as a function of molecular length, and two describe the width of the bands as a function of the length of molecular length. One of our goals is to develop systematic methods of improving the resolving power of longer molecules, and hence extending the number of bases that can be determined in a single sequencing experiment. Our approach is to determine how the small number of parameters that describe resolving power depend on experimentally controllably conditions. Knowing such relationships should permit systematic selection of combinations of experimental conditions that improve resolving power. Here, we show the relationship between the four parameters describing DNA mobility as a function of molecular length for systematic variations of gel composition and electric field strength. The data set employed is for double-stranded DNA separated in agarose gels, but the principles are similar to those encountered in DNA sequencing studies.

We applied a short-pulse diode laser emitting at 637 nm with a repetition rate of 30 MHz in combination with a confocal microscope to study bursts of fluorescence photons from individual labeled mononucleotide molecules in water. A newly synthesized oxazine dye and the commercially available carbocyanine dye Cy5 were used as fluorescent labels. Multichannel scalar traces, the fluorescence autocorrelation function and fluorescence decay times determined by time- correlated single-photon counting have been measured simultaneously. The time-resolved signals of the two mononucleotides were analyzed and identified by a maximum likelihood estimator. The results showed out that 60 detected photons per transit of a single molecule are sufficient to distinguish two labeled mononucleotides in water with a misclassification of less than 10 percent via their characteristic fluorescence lifetimes of 1.07 +/- 0.27 ns and 1.89 +/- 0.34 ns.

Since laser mass spectrometry has the potential for achieving very fast DNA analysis, we recently applied it to DNA sequencing, DNA typing for fingerprinting, and DNA screening for disease diagnosis. Two different approaches for sequencing DNA have been successfully demonstrated. One is to sequence DNA with DNA ladders produced from Sanger's enzymatic method. The other is to do direct sequencing without DNA ladders. The need for quick DNA typing for identification purposes is critical for forensic application. Our preliminary results indicate laser mass spectrometry can possible be used for rapid DNA fingerprinting applications at a much lower cost than gel electrophoresis. Population screening for certain genetic disease can be a very efficient step to reducing medical costs through prevention. Since laser mass spectrometry can provide very fast DNA analysis, we applied laser mass spectrometry to disease diagnosis. Clinical samples with both base deletion and point mutation have been tested with complete success.

The identify and location of base pair mismatches in non- covalent DNA:RNA duplexes are established using MS and MS-MS on a quadruple ion trap with electrospray ionization (ESI). MS-MS experiments on a 14mer duplex (D) with a single C:A base pair mismatch using lower activation energy results in selective cleavage of the mismatched A nucleobase, even in the presence of the wild-type duplex. The location of the mismatch base pair can be discerned via presence of the wild-type duplex. The location of the mismatch base pair can be discerned via selection of the (D-5H)^5- ion and fragmentation of the backbone at that location in a n additional MS-MS experiment. Selective fragmentation is observed for C in a C-C mismatched base pair, which is very difficult to detect using chemical cleavage or E. coli mismatch binding protein. In an RNA:DNA duplex with a single base pair mismatch, the DNA base is removed without fragmentation of the RNA strand, greatly simplifying the interpretation of the resulting MS spectrum. A method is presented for detecting two DNA strands, for example a point mutation which generates an oncogenic phenotype, and the wild-type message. The results suggest that ESI-MS-MS may provide a rapid and selective method to identify and locate genetic mutations without the need for chemical degradation or protein binding followed by gel electrophoresis.

The mass spectrometric analysis of oligoneucleotides is now performed using electrospray ionization (ESI) and matrix- assisted laser desorption-assisted laser desorption/ionization (MALDI). However, the processes involved in the production of gas-phase ions by these ionization methods are not well characterized. In this paper, we report investigations into the role of oligonucleotide solution-phase behavior on the subsequent gas-phase behavior of the oligonucleotide. Characterization of mononucleotide ion intensity as a function of phase behavior on the subsequent gas-phase behavior of the oligonucleotide. Characterization of mononucleotide ion intensity as a function of solution pH, determination of the gas-phase acidity of mononucleotides in order to identify the site(s) of deprotonation, and the use of organic bases to reduce cation adducts without the formation of the ammonium slat of the oligonucleotide were studied. The results of these studies show that the ionization process is likely influenced to a great extent by the gas-phase properties of oligonucleotides, although there are several important solution phase factors to be considered when analyzing oligonucleotides via MALDI-MS.

We have developed a nucleic acid detection assay using an upconverting phosphor DNA reagent as a reporter probe for the detection of target single-stranded DNA molecules. Upconverting phosphors allow the detection of DNA against zero background signal because low energy IR light is absorbed by a two-photon process and remitted as a visible photon, eliminating the presence of autofluorescence. The spectral properties of upconverting phosphors, which emit in the blue, red, or green region of the spectrum, are not sensitive to solvent conditions or to temperature and can therefore be used in a variety of buffers for the qualitative and quantitative detection of biological molecules. Short DNA oligonucleotides were covalently attached to the surface of submicron size phosphor particles and used to detect M13mp18(+) DNA targets captured by magnetic beads labeled with a capture probe specific to the target. This system allowed the detection on the order of femtomoles of M13mp18(+) strand DNA. Buffer conditions were optimized for the use of the phosphor reporters. The phosphor-labeled probes are stable in aqueous solution for 2 weeks and have a lyophilized shelf life of at least one month.

The interaction of DNA and lanthanide ions using Matrix assisted laser desorption time of flight mass spectrometry. 2', 4', 6' hydroxy acetophenone was the matrix used, and Europium chloride was the lanthanide salt used to interact with Adenine and Guanine containing synthetic oligonucleotides. The MALDI analysis permitted the detection of DNA/Europium adducts and also provided information on the stoichiometry of the interaction.

While molecular weight determination of proteins by matrix- assisted laser desorption/ionization (MALDI) has been quite successful, the measurement of deoxyribonucleic acid polymers in a size range useful for fine-structure genome analysis has proven more difficult, due to decreasing sensitivity and resolution at the higher molecular weights. Tailored cleanup techniques and the use of delayed extraction time-of-flight mass spectrometry are approaches that offer considerable promise for improved MALDI analysis of DNA.

Recent advances in molecular biology have enabled the deposition of nucleic acids on solid phases to form arrays of oligonucleotides. Such arrays are being applied in milli to nanoscale molecular and biochemical analysis such as genetic mutation detection, gene expression quantitation and DNA sequencing using so-called DNA chip arrays. A major obstacle to continue use of such arrays is detection and analysis of the arrayed nucleic acids, nucleic acids, nucleic acids' targets that bind via hybridization to the arrays and the products of array-based biochemical reactions. We report here on the design and utility of our experimental set-up for analysis of surface bound, arrayed oligonucleotides, and our experience in detection and quantitation of multiple fluorescent labels bound to the surface through attachment, hybridization, and arrayed primer extension.

A number of algorithms have been developed for three-dimensional (3D) deconvolution of fluorescence microscopical images. These algorithms use a mathematical-physics model for the process of image formation and try estimate the specimen function, i.e. the distribution of fluorescent dye in the specimen. To keep the algorithms tractable and computational load practical, the algorithms rely on simplifying assumptions, and the extent to which these assumptions approximate the actual process of image formation and recording has a strong effect on the capabilities of the algorithms. The process of image formation is a continuous-space process, but the algorithms must be implemented using a discrete-space approximation to this process and render a sampled specimen function. A commonly-used assumption is that there is one pixel in the specimen for each pixel in the recorded image and that the pixel size in the recorded image is small compared to the size of the diffraction limited spot or Airy disk, a condition necessary to satisfy Nyquist sampling criterion. Modern CCD cameras, however, have large wells that integrate into a single pixel an area of the image that is significantly larger that the Airy disk. We derived a maximum-likelihood-based algorithm to accommodate for these large CCD pixel sizes. In this algorithm we assume that each pixel in the recorded image integrates several pixels that satisfy Nyquist criterion. The algorithm then attempts to estimate the specimen function at a resolution better than that allowed by the CCD camera. Preliminary results of this sub-pixel resolution algorithm are encouraging. Keywords: Sub-pixel resolution, missing information, maximum-likelihood, expectation-maximization, super-resolution

While sedimentation equilibrium is most commonly used to characterize the molecular weight and state of association of single proteins, this technique is also a very powerful tool for probing the interactions between two or more different proteins, and can characterize both the binding stoichiometry and the equilibrium constants. To resolve the complex binding interactions that can occur in such systems, it is crucial to globally fit data from many experiments to a common binding model, including samples made with different mixing ratios and a wide range of total concentration. It is often also essential to constrain the parameters during fitting so that the fits correctly reproduce the molar ratio of proteins used in making each sample. We have applied this methodology to probe mechanisms of receptor activation for a number of hematopoietic receptors and their cognate ligands, using receptor extracellular domains expressed as soluble proteins. Such data can potentially help in the design of improved or new protein therapeutics, as well as in efforts to create small- molecular mimetics of protein hormones through structure- based drug design. Sedimentation equilibrium has shown that stem cell factor, erythropoietin, and granulocyte-colony stimulating factor can each dimerize their respective receptors in solution, but the mechanism of ligand-induced receptor dimerization for these three systems are strikingly different.

Vinca alkaloids induce tubulin to self-associate into coiled spirals that further align into sheets and macrotubes. The energetics of spiral formation has been studied by sedimentation velocity in a Beckman Optima XLA analytical ultracentrifuge. The analysis process involves conversion of an absorption based sedimentation pattern into a sedimentation distribution, g(s), and determination of mechanism involving isodesmic ligand-medicated or ligand- mediated plus ligand-facilitated self-association. We have compared the vinca alkaloid-induced self-association of porcine brain tubulin in the presence of 50 (mu) M GTP or GDP. For each drug investigated the affinity is shown to be enhanced by GDP and allosterically linked to GDP/GTP, NaCl and divalent cation binding. These allosteric effectors differentially interact with one another. Thus, GDP enhances self-association over GTP by 0.90 kcal/mol, but the enhancement is reduced to 0.35 kcal/mol by increasing NaCl concentration to 150 mM. High salt stimulates spiral formation but it affects GTP-tubulin preferentially over GDP-tubulin. Divalent cations stimulate spiral formation but GDP-tubulin differentiates between Mg⁺², Ca⁺², and Mn⁺², while GTP-tubulin does not. Divalent cations, Ca⁺² and Mn⁺², induce spiral condensation and macrotube formation, but this is also inhibited by high salt. The differential action of these effectors suggests an interpretation of the energetics in terms of a structural model where charge localization, binding domains and conformational interactions within and between the tubulin heterodimers are responsible for the observed allosteric effects.

The determination of the biophysical properties of polysaccharides is becoming increasingly important in commercial and pharmaceutical applications. molecular weights distributions, conformations, and size distributions are obtainable through experiments with analytical ultracentrifugation. These determinations can be difficult, however, due to the high thermodynamic nonideality and polydispersity with respect to both molecular weight and composition, as well as potential self-association commonly observed with these molecules in solution. Usually, a combination of complementary techniques are needed for the full understanding of the solution characteristic of the macromolecules. Although carbohydrates do not have a chromophore absorbing in either the ultraviolet or visible spectrum, optical systems based on refraction of light by the solute can be used for the study of the macromolecular characteristics. (1 yields 3)-(beta) -Glucans are known to exhibit immunomodulatory and biological activities. These activities are dependent on the conformational structure and molecular weight of the glucan. A soluble (1 yields 6)-(beta) - branched (1 yields 3)-(beta) -glucose homopolymer was isolated from the cell wall of Saccharomyces cerevisiae. In this paper we describe the preliminary characterization of a purified fraction from this soluble glucan measuring molecular weights using analytical ultracentrifugation, light scattering, and size-exclusion chromatography - and we explore the effects of temperature using these techniques.

A fluorescence detector has been developed for the XLA/XLI analytical ultracentrifuge. The excitation source consists of an Argon ion laser coupled to a single-mode, 3.5 micron diameter fiber. The remaining optics are in the rotor chamber and attached to a 500 steps/second, 2 microns/step motorized stage mounted above the centrifuge rotor. The excitation and emission paths are coaxial. The excitation spot is about 7 microns in diameter and is focused about 1 mm below the top window of a centrifuge cell. Radial resolution is achieved by scanning this spot along a radial axis. Emitted light is gathered by the condensing lens and redirected by the beam splitter to a condensing lens focused on a pinhole. A long-pass filter sits between the pinhole and a photomultiplier tube. Operation of the stepping motor and photomultiplier electronics is under computer control, allowing data to be acquired from a radial stripe around the rotor after each step. This data acquisition system will permit the monitoring of specifically labeled compounds in the presence of a large background of other molecules. Both the sensitivity and the selectivity of this optical system will be useful for a range of biological problems.

The search for increased sensitivity of bio-analytical techniques has recently shifted from signal generation to detection. While enzyme amplifiers and chemiluminescent reporters developed by chemists over the last two decades gradually moved detection limits to the attomol level, it has taken engineers only a few years to reach single- molecule sensitivity with the development of new instrumentation. A number of different approaches have successfully achieved single-molecule fluorescence detection including confocal and near-field scanning optical microscopy, photon-counting cameras, fluorescence- correlation and time-gated spectroscopy. They detect labels immobilized on substrates, diffusing in solution and flowing in electro-osmotic and hydrodynamically focused streams. Biotechnology has created numerous application s for single- molecule detection. In research labs, it can dramatically increase the rate of DNA sequencing, screen libraries for products of directed evolution, and characterize compounds in drug discovery programs. In medical diagnostics, ultra- sensitive detection technologies can be used for genetic screening, detection of infectious diseases, or multi- analyte profiles. It can be applied to immunoassays as well as DNA or RNA hybridization assays.

We describe the application of single molecule detection (SMD) technologies for the analysis of natural and synthetic transport systems. The need for advanced analytical procedures of these complex and important systems is presented with the specific enhancements afforded by SMD with flowing sample streams. In contrast to bulk measurements which yield only average values, measurement of individual species allows creation of population histograms form heterogeneous samples. The data are acquired in minutes and the analysis requires relatively small sample quantities. Preliminary data are presented from the analysis of low density lipoprotein, and multilamellar and unilamellar vesicles.

Analytical instrument systems present assemblies of components which each introduces its won contribution to the overall system precision. Combining realistic models of detector response with a Beer's Law model for absorbance (A) provides a prediction for the overall system uncertainty. For the constant detector uncertainty model, reasonable values of sample parameters lead to substantial increases in the absorbance at minimum relative concentration error (RCE), the value of the RCE and the limits of the range of A values for best RCE. Similar effects follow from the square root detector model. In both detector models these effects are increased by concentration uncertainty. Assay chemistry factors dominate the system RCE at low A, making high precision difficult at low A. For fluorescence measurements, since analyte concentration is nearly proportional to detected signal, the same detector models indicate no minimum in RCE.

Biological time-series data pertaining to human circadian and ultradian hormonal rhythms are often short, sparse, irregularly spaced, and noisy. In addition, they often have missing data points and have variable experimental uncertainties. The objective of collecting and analyzing such data is to find the amplitude, phase, and period of the primary rhythmic component contained within the data. Often the question is simply: Does a rhythm exist. The theoretical aspects of some Fourier techniques are discussed, including methods for detrending non-stationary time-series and the evaluation of confidence intervals. Analysis of typical biological data are also presented.

A new class of reporter label, consisting of rare earth elements embedded in a crystalline particle, has been developed for in vitro diagnostic applications. These unique labels upconvert low energy (IR) radiation to high energy light by a multiphoton absorption process and subsequent phosphorescence emission. As a result, upconverting phosphors can be visualized with no biological background or autofluorescence signal. In addition, phosphors have narrow absorption and emission bands, making them ideal for simultaneous multianalyte test. The crystalline nature of the phosphors makes them insensitive to environmental conditions, with essentially infinite shelf life and no photobleaching at the irradiances used for excitation. We have covalently coupled (Y_0.86Yb_0.08Er_0.06)₆O₂S phosphor labels to antibody probes to create a reporter reagent that can be excited by 980 nm radiation from a diode laser and detected by a modified spectrofluorimeter. Target analyte sensitivities of approximately 10 ng/mL to Staphylococcal enterotoxin B have been demonstrated using a sandwich assay in a magnetic bead or capillary wick formats in a non-optimized assay system. These results are directly applicable to the development of assays that can be performed on microfabricated biochips or in microflow channels.

Using upconverting phosphor reporter probes to detect antigenic and nucleic acid analytes in environmental samples required the design and manufacture of a novel sampling and assay device that exploits the unique characteristics of the upconverting phosphors. The absence of natural materials that upconvert energy make single particle detection possible with these reporters. Ultrasensitive detection and quantitation of analytes is governed by the area of the capture surface, the density of the capture probe,s as well as the specific upconversion efficiency of the phosphor label. The optimization of a flow channel and capture surface for single phosphor particle detection is a complex function of the volume of sample that is drawn through the channel, the probability that this sample volume contains an analyte particle, the fraction of analytes contained in the sampled volume that bind to the capture site, and the optical dimensions of the capture site. The result of this optimization using theoretical diffusion models for submicron upconverting phosphor particles was a rectangular flow channel with a 200 X 300 micrometers capture surface. Capture surfaces of this size have been prepared using photo-directed synthesis methods inside glass capillary wicks that have been subsequently used for the phosphor- based assays.

We describe an integrated detection system based on upconverting phosphor particles bound to capture sites on the inside surfaces of rectangular wick capillaries. This deice can be used with either antibody or nucleic acid to detect specific micro-organisms. The system uses a high- power, 980 nm, semiconductor diode laser to illuminate 200 X 300 X 20 micrometers capture surfaces. The rectangular capillary wicks are held in a tray that is inserted into the detection system, positioning the capture surface at the object plane of the optical system. Phosphorescent light emitted from the capture surface is collected by a high numerical aperture microscope objective and directed through a series of filters onto either a CCD camera or a photomultiplier. A combination of band-reject filters attenuates the 980 nm laser excitation light and its harmonic at 490 nm, and a tunable liquid crystal filter provides for rapid scanning from 400 to 750 nm. The data acquisition and control is controlled by a laptop PC with a custom GUI interface developed using LabWindows/CVI. The system can detect a single phosphor particle bound to a capture surface.

The fine structure of the various morphological stages of development in vitro and in vivo of Yersinia spp. culture has been studied with the aid of light and electron microscopy and computerized television morphodensitometry. An analysis is made of the ultrastructural location of cell wall elements with regard to the cell covers and other forms of exogenous material. The results of cytochemical investigation and then morphodensitometrical estimation on various forms of Yersinia L-transforming are discussed.

Using the changes of light absorption coefficient for measuring the ratio of haemoglobin and oxyhaemoglobin, the oxygen saturation can be calculated and has got successful applications. According to the spectroscopic characters of haemoglobin and oxyhaemoglobin, this paper presented a new system for monitoring the arterial oxygen saturation.

Addition of Luminol to nondiluted blood of healthy donors results in a short and weak increase of chemiluminescence (CL) from it. Contrary to that in 25 cases of stable angina pectoris the intensity of CL from blood of patients sharply increased upon addition of luminol exceeding that form healthy donors' blood 10-100-fold. 24 hours after the 3D intravenous low-level treatment CL burst in patients' blood in the presence of Luminol was in general significantly lower than before the beginning of the treatment. After the 7th treatment the pattern of CL kinetics was in most cases similar to that of healthy donors' blood. However, after the 10th treatment intensity of Luminol-enhanced CL usually increased and for blood of some patients even exceeded its values obtained before the treatment. Some correlation CL from nondiluted blood with neutrophil activity studied by NTB-test and plasma viscosity of same blood was noted. Using highly sensitive single photon counters it is possible to reveal abnormal levels of CL from no more than 0.1-0.2 ml of blood within 3-5 min.

We have constructed a compact, all solid state, time correlated single photon counting device for sensitive measurement of the time resolved behavior of DNA fragments passing through a capillary tube. The instrument consists of a pulsed diode laser PDL, a single photon avalanche diode SPCM-211 and a PC board containing all TCPC electronics. We achieved an overall instrument response function of the system of less than 300 ps. This is fast enough for measurement of fluorescence lifetimes. We demonstrated the utility and the sensitivity of the instrument; dynamic measurements of fluorescence decays for NIR dye-labelled nucleotide bases were measured during capillary electrophoresis. The results indicate that in a two-dye experiment the characteristic lifetime of the probe could be used to identify the terminal nucleotide base. The lifetimes of the tow dyes bound to DNA fragments were determined to be 670 ps and 530 ps.

We have developed and are currently using a first generation fully automated hybridizer and sequence imager for multiplex sequencing. This instrument processes four multiplex membranes concurrently with a five hour cycle time, yielding a projected throughput of more than 50 million bases per year. We have also begun development of a second generation instrument which will be capable of outputs in excess of 1.5 billion raw bases per year by processing up to one hundred multiplex membranes concurrently. After direct transfer electrophoresis, each DNA containing membrane is placed in a flat tray for processing. Solutions are automatically added to and removed from the independently processed trays/membranes. In order to achieve short cycle times, a non-radioactive detection method is utilized in which oligonucleotide probes are conjugated to an enzyme and then hybridized to each of the multiplexed sequencing reaction products on the membrane. Chemiluminescent substrate is added and converted by the enzyme into an unstable product which emits light upon decay, thus revealing the sequence pattern. In the first generation instrument, the emitted light is collected at an imaging station consisting a single scientific CCD camera mounted on linear slides. The second generation instrument will increase throughput by utilizing a stationary array of cameras. By capturing segments and then tiling to form an image of the entire membrane, high collection efficiency and high resolution optics may be used to detect the extremely low light levels associated with chemiluminescence.