MXPA00001288A - A digital imaging system for assays in well plates, gels and blots - Google Patents

Abstract

A method and apparatus are disclosed for use in an area digital imaging system for assays to extract targets on a specimen containing an array of targets that may not be arranged in perfect regularity. A matrix is defined of nominal target locations including a probe template of predefined, two-dimensional size and shape at each of a plurality of fixed, predefined grid points on the specimen, and a determination is made of the most probable location of the probe template corresponding to a specific target by sensing both pixel intensity and the spatial distribution of pixel intensities in an image of the specimen at a plurality of locations in the vicinity of a nominal target location.

Description

A SYSTEM FOR FORMING DIGITAL IMAGES FOR TESTING IN PLATES OF WELLS, GELS AND SPOTS Field of the Invention The present invention relates to a system for creating digital images of fluorescent, luminescent, or bright field specimens. The system is flexible, both in its illumination mode and in that the specimens can be configured in matrices (for example, well plates) or can be randomly configured (for example, chemiluminescent colonies, gel medium).

BACKGROUND OF THE INVENTION The present invention is primarily an assay image formation system. An assay can be defined as a measurement of physical (chemical, biochemical, physiological, or other) properties within a specimen. The assays are typically used in the areas of molecular biology, genomics, and pharmaceuticals. The standard test specimen container is a plastic or glass plate that contains 96 small chambers, called wells. Detection instruments and robotic plate management technologies have evolved to make efficient use of the 96 well plate, and to provide as high performance as possible with this plate format. Current classification technologies can allow a large classification laboratory that uses 96-well plates to process a few thousand compounds per day. Recently, there has been a massive growth in the number of compounds available for testing. In part, this is due to an increased exploitation of biodiversity resources to generate natural compounds. For the most part, the proliferation of new compounds is a result of a new chemical technology, called combinatorial chemistry. In combinatorial chemistry, large numbers of related compounds are synthesized (using permutations of chemical building blocks), and then they must be tested to see their medical value. With new discoveries in natural compounds, and with the advent of combinatorial chemistry, pharmaceutical companies and biotechnology companies are generating extensive "libraries" of unproven compounds. These libraries can easily contain millions of compounds. Standard measurement technologies can not cope with volume, and new technologies are needed that will increase the rate at which compounds can be subjected to the initial test (classification) to see their medicinal use. To show a greater advantage over standard technologies, new technologies must allow performance to exceed 100,000 compounds / instrument per day. Imaging technologies are hoping to increase performance to the required levels. These have the additional advantage of flexibility, in the sense that the imaging system can be applied to well trials and other formats, and to tests that are static or change over time. Most tests are designed in such a way that changes in absorbance, transmission or light emission reflect the reactions within the specimen. Therefore, most of the test measurement instruments detect changes in luminance as their principle of operation. For detection instruments, bioluminescence or chemiluminescence provides the simplest type of assay, in the sense that there is no need to apply illumination. Absorbency assays involve the transillumination of the specimen, usually with monochromatic light. The reaction of interest affects the degree to which a fluid absorbs light, and this absorbance can be measured. Fluorescence is emitted when a fluorophore interacts with an incident photon (excitation). The absorption of the photon causes an electron in the fluorophore to rise from its ground state to a higher energy level. After, the electron returns to its original level, releasing a photon (emission of fluorescence) whose wavelength depends on the amount of energy that is hare during the regression. A given fluorophore can emit at individual or multiple wavelengths (creating an emission spectrum), as electrons fall from different orbits to their fundamental states. The emission spectrum is constant for each species of fluorophore. Fluorescence assays require the application of an intense monochromatic illumination beam, called "excitation". Fluorescence assays are used for the following types of applications: 1. A detector tuned to a specific emission spectrum can be used to locate a fluorophore. For example, wells containing cells expressing a protein fluorescently labeled from wells that do not contain them can be discriminated. 2. By measuring the fluorescence intensity, a detector system can estimate the concentration of a fluorescent molecule. 3. Changes in the fluorophore molecule (such as binding of fura-2 to Ca ++) will lead to alterations in the emission spectrum. A detector can be used to measure these spectral changes, as an indication of changes in the environment of the fluorophore.

Wells Each well contains a discrete condition of the experiment, and alterations in light emission are measured to determine if that condition produces favorable properties. "Well plate" tests are superior in performance and lower in cost than similar tests in discrete vessels. The reactions inside the wells can be of many kinds. In chemistry tests, the molecules of different compounds (for example, a drug candidate and a receptor molecule) are placed in the same well, and the interaction between these compounds is observed. In cell-based assays, each well contains a population of living cells, and the effects of the compounds on these cells are observed. Most tests are conducted by making a single measurement of each well. However, it is also possible to record changes over time, by measuring each well repeatedly. The use of repeated observations can be called a "dynamic" test. Standard well plates contain 96 or 384 wells in an area of approximately 8 x 12 centimeters. The trend is toward the miniaturization of wells. Prototypes containing 864 wells or more are under evaluation in many sites. The goal is to develop plates with high density configurations of small "microwells" (eg, thousands / plate) with small filling volumes. That is, miniaturized wells may contain 1 ul of fluid instead of the 199 ul or more used in a typical 96 or 384 well test. The miniaturized assay formats promise to achieve dramatic cost reductions, and simplify waste procedures, while allowing many more trials to be conducted.

Hybridization configurations and genetic assays Low-throughput methods of genetic analysis use different electrophoretic procedures. Methods to increase yield and lower costs of genomic assays include the configuration of DNA clones (CDNAs) or synthetic oligonucleotides on flat support membranes or treated glass slides. The configurations of cDNAs or oligonucleotides (so-called at-density grids) are then hybridized to samples of genomic material to quantitate the levels of gene expression, or to locate relevant sequences. In the past, most hybridization assays have been conducted using isotope labeling and storage phosphorus imaging systems for detection. However, non-isotopic methods (particularly fluorescence) are under investigation in many laboratories. Non-isotopic high-density grids provide the potential for very high performance at low cost, and different detection technologies have been developed for these specimens.

Free format assay Reference can be made to assays that occur within a regularly spaced configuration (wells, cDNAs within a grid) as fixed format assays. Specimens that are distributed in an irregular manner can be called free format trials. Examples of free-form assays include electrophoregrams, bacterial colonies in culture, and different combinatorial assays in which the compounds attached to a bead are distributed through a tissue culture. The common factor in these free-form tests is that areas of altered luminance can occur at any spatial location. Instruments designed for fixed-format assays (fluorescence plate readers, liquid scintillation counters, etc.) only read from defined locations in the specimen These are not useful with specimens in which the effects are found in locations that are not previously defined. In contrast, imaging systems are able to detect and quantify reactions at any position within an image, and there is an extensive history of image analysis that # are applying to free format trials.

Compendium of types of trials pharmaceutical companies face unprecedented numbers of new compounds, which must be classified by their medicinal value. The specimens are classified using luminescence, absorbance, or fluorescence technologies. Specimen formats include standard or miniaturized well plates, high-density hybridization assay configurations, and free-form assays. The present invention is designed to be useful for all these technologies and specimen formats.

Imaging Systems by Area A system of image formation by area places the entire specimen on a detector plane at a time. There is no need to move the PMTs or to scan a laser beam, because the camera forms images of the entire specimen on very small detector elements (usually CCDs), in parallel. The parallel acquisition phase is followed by an interpretation of the entire image from the detector. Interpretation is a serial process, but it is relatively fast, with speeds ranging from thousands to millions of pixels (image elements) / second. Imaging systems by area offer some very attractive potential advantages: 1. Because the image of the entire specimen is formed at the same time, the detection process can be very fast. 2. It is relatively easy to acquire a timed series of images for dynamic tests. 3. Given an appropriate lighting system, any excitation wavelength can be applied. 4. Images of luminescence reactions (biolucence, chemiluminescence) can be imaged. 5. Freeform or fixed format specimen images can be formed. In many applications (for example, film autoradiography), image formation by area has a history, instrument developers are experienced, and problems are well understood. In contrast, imaging by trial presents new and significant technical challenges. There is no established practice for shaping trial images, and very few instrument developers have practical experience with both area imaging and assay technologies.

Formation of Images by Luminescence Area We will define luminescence as light emitted from a specimen, without excitation by external light. Most luminescence reactions are completely dead, and this can make extreme demands on detection. The strategy of image formation by standard area is to use CCD cameras of scientific degree, which achieve relatively high levels of sensitivity. CCD cameras have also been used intensified. The typical prior art applies image formation by area to luminescent tests on flat membranes. The descriptions of image formation per well are defective in the sense that no correction for the parallax error is described. The telecentric lenses have not been used because the telecentric lenses of the prior art have not been able to collect enough light. The prior art also fails to describe the use of software that would allow the automated analysis of the targets that are not in a perfectly regular configuration, or software that mold the variations in the efficiency of measurement from well to well. The novel features of the present invention (efficient light collection combined with freedom from parallax, lens detection software, calibration software that minimizes variations in measurement efficiency) allow it to be used with luminescent specimens beyond of the capabilities of the luminescence imaging systems of the prior art.

Formation of Images by Fluorescence Area Image formation by area is in routine use for fluorescence microscopy, where epifluorescence is the universal method and is easily applied. In a microscope, the epifluorescence is implemented by means of a configuration that incorporates a dichroic mirror and interference filters. The prior art in fluorescence microscopy is irrelevant to the present invention, which is designed specifically for macro-specimens. Fluorescence imaging by area at a macro level is routinely applied with gels and spots. Although some specialized research systems have been reported, the prior art is dominated by low cost commercial systems for gel fluorescence / routine stain. Typically, this type of imaging has comprised suboptimal illumination delivery by excitation. The most common form of illumination by area uses gas discharge illuminators (for example, ultraviolet light box). The lamps are covered with a filter that limits the emission to the peaks that the excited gas emits inside the lamp. In this way, most prior art systems offer a limited number of excitation wavelengths.

Many prior art systems, such as those using transilluminators, pass the excitation directly through the specimen and into the collection optics. For the highest sensitivity, however, it is important that the excitation is not sent directly into the detection optics. With direct sending, both the excitation and the emission wavelength are detected, with the excitation at levels much higher than the emission. The excitation wavelength can be filtered selectively, but enough excitation remains to degrade the sensitivity. To avoid direct excitation detection, some systems of the prior art send the excitation from above, from the side, by means of a dark field, or by using refraction (for example, evanescent wave) within the specimen. In these cases, the detection optics see the emission of fluorescence from the specimen, with the minimum detection of excitation. However, these techniques have severe limitations. Side-mounted fiber optic illuminators send uneven light, and can only be used with flat specimens. When these are used with wells, the light enters the specimen at an angle and fails to penetrate deeper regions. Similarly, refractive or darkfield illuminators do not penetrate deep wells, can not be used with opaque specimens, and tend to provide rather dim levels of illumination.

We can summarize the characteristics of the prior art systems as follows: 1. Most are limited to the wavelengths emitted by gas discharge lamps. Typically some combination of UVA, UVB, UVC, and / or white light lamps is provided. Other wavelengths can not be obtained, and this is a severe inconvenience. 2. The wavelengths can not be woven during a test. If the lighting should be changed during the test (for example, as for calcium measurement with fura-2), the devices can not be adapted. 3. Insensitive to small fluorescence theories. The transillumination comes from directly under the specimen inside the optical detectors. Therefore, even very good filters fail to remove all incident lighting, and this creates a high background of non-specific lighting. Small alterations in fluorescence (typical of many trials) are lost within the non-specific background. 4. Cameras and inefficient lenses. A few systems use high performance cameras. Even these few systems use standard CCTV or photographic lenses, and can only be applied to bright specimens. 5. The parallax error makes accurate fine image formation impossible. Since they have not been •: € 7 available wide-aperture telecentric lenses, these systems exhibit parallax error when forming well images. In an example of a fluorescence system of the prior art, an optical fiber is coupled to the specimen at its input, and to a CCD camera at its output. This use of a fiber optic lens at the entrance to a camera gives freedom from parallax error and efficient light collection. The CCD camera can be operated in the photon counting mode to obtain individual photon detection capability. In achieving parallax-free images and high sensitivity to light, the camera coupled to the fiber is like the present invention. However, fiber coupling to the specimen has the following major drawbacks: 1. A fiber-coupled system is limited to specimens that are transilluminated, because there is no place to insert an epi-illumination mechanism. Therefore, the fiber-coupled system would not be as sensitive to subtle variations in fluorescence as the present invention (epi-lighting). 2. Because the optical fibers have a very shallow depth of field (typically about 1 μm), the fiber-coupled system must see the specimen through a thin, transparent film of glass, plastic, or other material. Not all tests can be conducted on plates that have thin, transparent backgrounds. If the tests should be conducted on opaque plates, the fiber coupled system becomes inefficient, because it can not be placed close enough to the fluid inside the wells to get the focus. 3. Due to its limited depth of field, an optical coupled to fiber exhibits a compromised performance with deep specimens. For example, the fluid in the well plates is typically more than 1 millimeter deep. Only a small part of the fluid is in focus with an optic coupled to fiber. This leads to the collection of inefficient light from those parts of the fluid that are not in the critical focus. 4. The fiber-coupled system could not be used for fluorescence polarization studies, due to the effects (birefringence, etc.) of the fiber in polarized light. 5. Well plates are 8 x 12 centimeters. The optical fibers that make up the image of this size are very difficult and expensive to build, and should be brought to a much smaller CCD by tapering the bundle of fibers to a minimized configuration. Minimization results in higher losses in transmission efficiency, which could fall from 70 percent to 10 percent. To avoid both the high cost of a large optical fiber, and the loss of transmission efficiency resulting from minimization, the specimen would have to be acquired as a number of sub-images. In this case, a translation table moves the specimen or the camera, and a bunch of frank or slightly minimized fibers of a reasonable size is used, after a scan, the subimages will be reassembled to show the entire specimen. This multiple acquisition is slow, affects the accuracy and makes it impossible to use the device with tests that change over time. Other examples of prior art florescence systems include the use of a cooled CCD camera, with a laterally mounted illuminator, or on the specimen. In some cases, an excitation filter wheel may be inserted before a fiber optic light delivery system, to achieve some of the functions of the present invention (selection of wavelengths). All of these prior art systems have the following major drawbacks: 1. An illuminator mounted on the side or top results in very uneven illumination with almost any specimen, and is completely unsuitable for use with wells. 2. The systems lack a telecentric lens, so that parallax would make it impossible to image well trials. 3. There is no objective detection software. There is another fluorescence system of the prior art (Plate Reader for Image Formation of Fluorescence - FLIPR, Molecular Devices) that uses laser illumination applied at an angle to the wells in a 96-well plate. A transparent dot mask is placed over the wells, so that the angled light reaches only a small part of the specimen fluid at the bottom of each well. The data is detected using a cooled CCD camera. FLIPR is inflexible, in the sense that it only works in the 96-well format, uses specific laser lines, does not offer epi-lighting, can not illuminate the entire fluid content of deep wells, and is not applicable to the formation of luminescence images, free-form image formation, or higher-density well formulations. The Science Applications International Corporation (SAIC, San Diego, CA) was developing the RTFluorimeter, a high performance fluorescence classification system for kinetic assays in the 96-well format. The RTFluorimeter includes a robot for filling the wells, a high-energy arc lamp, a filter wheel for excitation wavelength selection, and a CCD configuration detector. The system was illuminated using angled optical fibers, placed on the sides of the specimen, and did not include epi-illumination. The operating data and the specific imaging design information have not been forthcoming, and this system has been discontinued. There is much prior art in the use of imaging to detect fluorescence assays incorporated within microfabricated devices (often referred to as "genodetectors"). Some genodetectors use scan imagers (for example, Molecular Dynamics or Hewlett Packard genchip readers), and detect light with a photomultiplier. Others use area CCDs to detect alterations at test sites fabricated directly on the CCD, or on a cover slide that can be placed on the CCD. It can be seen that many other types of genodetectors will appear in the near future.

Advantages of the Present Invention for the Formation of Fluorescence Images by Macrofluorescence Area The novel features of the present invention minimize the disadvantages of the macro-fluorescence systems of the prior art. These features include the following: 1. The device can accept light from almost any source. Therefore, illumination wavelengths can be selected without taking into consideration the peak (s) of a gas discharge lamp or laser beam. 2. During a test, lighting may be altered using a computer-controlled filter wheel or other device. 3. Small alterations in the emission of fluorescence can be detected, because the fluorescence illumination comes by means of epi-lighting. The lighting by direct excitation does not enter the optician, and the non-specific background is as low as possible. 4. The camera and lenses form a highly optimized system for use with dull specimens. 5. The telecentric lens removes the parallax error, in such a way that the well plate tests are accurate. 6. Optics can be adapted to many forms of fluorescence, including time resolved and fluorescence polarization assays. 7. All fluorescence lighting and the detection system are integrated into a single mechanical assembly. This assembly is easy to integrate with the automated equipment that handles the specimen. For fluorescence, a critical advantage of the present invention is its epi-illumination lens. This is flexible in application, allows simultaneous imaging of the entire specimen, illuminates, collects data without high levels of background or parallax error, easily integrates with the automation that manages the specimen, and can be used with dull specimens. In a particular adaptability to opaque media, the lens of the present invention minimizes the disadvantages of an optical coupled fiber. The lens focuses some distance away from the specimen, uses epi-illumination, and has a large depth of field. These features allow for the focusing of fluids that are within opaque media and, more generally, the typical depth fluid approach (eg, 1 millimeter). Another advantage of the present invention is that, since there are no fiber optic lenses that act as a light demodulator, fluorescence polarization assays can be performed. Finally, the lens of the present invention forms images of the entire well plate in a single acquisition. This avoids the need for translation tables, and makes both exact and dynamic calibration tests much easier to perform.

Genomic Assays In a particular, the present invention has advantages for genomic assays. Genomic assays are in flux, and many labeling methods (fluorescence, luminescence, absorbance, isotopic) and formats are likely to appear. Unlike the scan image formers currently used for these genomic assays, the present invention allows almost any specimen format (including wells), any fluorescence wavelength, and is capable of luminescence detection. Luminescence methods may offer the highest levels of sensitivity in some genomic assays. The flexibility of the present invention is unique, and makes it suitable as a single tool for laboratories that perform the prototyping of genomic assays. In the prototyping of assays, it can be difficult to predict the specific wavelengths and specimen formats that will be used. The present invention allows a laboratory to construct an assay using almost any labeling and detection paradigm. There is no need to select from the small number of labeling and detection methods that a specific laser scanner supports. The present invention can also be used for mass sorting (using high volume manufacturing), when scanning laser excitation is inappropriate, or when luminescent detection is required. In short, it is a purpose of the invention to provide a more flexible alternative to prior art laser scanning of genomic assays.

Compendium of Fluorescence Imaging Formation by Area The present invention includes features that are novel, both individually and in combination. These are: 1. macro-epi-lighting; 2. ability of a single instrument to conduct accurate (parallax-free, calibrated) tests, both fluorescent and non-fluorescent; 3. parallax freedom; 4. It easily adapts to fluorescence polarization, to fluorescence resolved by time, and other methods; 5. easy integration of fluorescence lighting with specimen handling and dosing automation; 6. The software includes calibration methods that minimize the errors inherent in the image formation by area; 7. the software finds and quantifies the objectives. A system in accordance with the present invention includes a CCD detector, a lens, a lighting system, an imaging system, and software. Together, these components are configured as a system that can be used with fixed or free formats, with specimens from wells or planes, and with any form of illumination (fluorescence, luminescence, transmission). The system exhibits measurement performance similar to that of standard measuring instruments, not imaging, while having the additional advantages of an imaging system.

Main Components of an Invention Conformity System a) Sensitive CCD Detector Area imaging systems use CCD configurations to form images. Factors that influence the ability of CCD configurations to detect small numbers of incoming photons include quantum efficiency, interpretation noise, dark noise, and dynamic range. Preferably, the present invention uses a CCD configuration with high quantum efficiency, low interpretation and dark noise, and wide dynamic range. b) A Lens Combining Telecentricity and Large Opening The lens of the present invention is unique in that it combines the telecentricity with a large aperture. This combination allows the exact reading of the deadly tests inside the wells. c) True Epif Luorescence Optics An epifluorescence optic both illuminates and collects through the same lens. Epifluorescence optics should include a source of excitation, and a barrier filter that removes the excitation of the detected signal and leaves only the fluorescence emitted from the specimen. The present invention incorporates both an excitation source and a barrier filter inside its lens. Epifluorescence optics are beneficial for producing lower backgrounds, a wider dynamic range, and a more linear fluorescence response than the dorsal, lateral, refractive or transilluminator. The ability to send large area monochromatic epi-lumination is a unique capability of the present invention. d) Optional Components for the Formation of Light Images Work Although you can use any detection technology by area, the preferred detector is a thinned CCD, illuminated by the back and large, which is capable of forming images up to levels low fluorescence and luminescence. For more extreme low light conditions, such as with the scintillation proximity test, the present invention allows the use of a camera cooled to very low temperatures (for example, the CCD operates at -70 degrees C). The cooling may be by means of liquid nitrogen, forced probe mechanism, multi-stage thermoelectric, or any other means. As an alternative, the present invention allows the replacement of a fiber optic lens with the epi-illuminator lens. This use of a fiber optic lens sacrifices flexibility and fluorescence capacity, in exchange for higher sensitivity with luminescent assays. The present invention also allows the use of a light amplifier, inserted between the lens and the CCD camera. In a preferred configuration, this light amplifier is an image intensifier. eg Flexible Lighting Source for Fluorescence Extraction The application of up to selectable wavelength illumination over an area of 8 x 12 centimeters is a feature of the present invention. The lighting is sent by means of a standard fiber optic coupling. This allows the use of almost any lighting source (halogen, arc lamp, laser beam, etc.). In a preferred configuration incorporating a halogen or arc lamp, any wavelength of monochromatic illumination can be selected by means of precision filters (usually standard interference filters). These are easily available in the small sizes that are used. A computer-controlled optical wheel / filter coupler is preferred for the selection of excitation wavelengths. The output of this device can be taken to a fiber optic plate specially designed for transillumination, to a fiber optic ring, line, or panel light for dorsal lighting, or (preferably) to a fiber optic lighting assembly inside. of the lens, for epi-lighting. f) Software that Integrates Control, Detection and Analysis The formation of images by area is inherently more complex than either the detection of non-formation of images, or the use of image formers by exploration. In particular, non-imaging counting systems have a relatively easy task. This only needs to control the scanning process, control the internal calibration, and create a small configuration of data points that represent each well. The sequence of steps can be as follows: 1. Calibrate the detector against the internal standard. 2. Illuminate a well. 3. Place a PMT over the lit well. 4. Read the well. 5. Transfer the data to the electronic board. 6. Light the next well and repeat. A system of image formation by area has a much more difficult task. Form the image of a well plate can include the following requirements: 1. Provide adequate lighting over the entire plate. 2. Control a high performance camera. 3. Store geometric and density correction factors. 4. Form the image of the specimen. 5. Correct the geometric and density variation. 6. If necessary, calibrate the image to the standards inside the specimen. 7. Locate each well and quantify the intensity, with respect to a calibration function. 8. Transfer the data to the electronic board. These tasks can only be performed if the imaging system is equipped with software that performs functions 2-8 above. The present invention incorporates that software, which includes novel software to locate the specimens, and to calibrate the imaging system to minimize errors. g) Optional Components for Time-Resolved and Polarization Fluorescence A standard way to minimize background is to take advantage of the long-life fluorescence properties of specific fluorites. With a long-lived fluorite, the light source can be strobeed very quickly, so that excitation is applied for a short period, and then blocked. Fluorescence tends to gradually fade (eg, over a period of 10 msec after blocking the excitation), so that the measurement can be made during the excitation blockage period. This measurement after blockage of the excitation illumination must be free of reflections, and of direct contamination by the excitation wavelength. This delayed measurement is known as time resolved fluorescence (TRF). The greatest advantage of the TRF is that it results in much lower levels of background fluorescence. The major disadvantages of TRF are that it requires specialized hardware for rapid excitation and blocking cyclization, and that a limited number of special fluorites (typically using Europium, Lantanides, or Terbium chelates) are suitable for the procedure. Typically, the present invention is applied to time resolved fluorescence through the use of both a shutter and a strobe light source. The stroboscope provides brief illumination periods and the exposure CCD is regulated, while illumination is on. Different types of shutter are possible: An image intensifier is mounted between the epi-illuminator lens and the CCD detector. The intensifier serves to amplify the light coming from the fluorite and, by means of gate regulation, the intensifier serves as a shutter. When the intensifier is regulated by the gate on ignition, the light is allowed to reach the CCD. When the intensifier is regulated by the damper on off, the light is prevented from reaching the CCD. The regulation by gate can be very fast, in such a way that the on period of the intensifier coincides with the optimal period of the delayed fluorescence from the specimen. A pulse-forming wheel, liquid crystal, sheet, or other obturator can be mounted inside the optical path, after the specimen but before the detector. For example, the shutter may be located at the position of the barrier filter inside the lens. The shutter is locked by time to the excitation stroboscope, in order to achieve the required delay between the on-off cycle of the stroboscope, and the exposure period of the CCD.

Although the preferred light source is a strobe lamp, other forms of illumination may be conceived by those skilled in the art (e.g., pulse laser beam, pulse arc lamp, continuous arc lamp). If the lighting can not be regulated precisely, a second shutter mechanism is required to provide an accurate on-off cycle. A critical advantage of the present invention for the TRF is that the illumination is applied by means of the telecentric lens, epi-illuminator. This lens provides a convenient means by which fluorescence excitation, parallax-free telecentric harvesting, and efficient optics can be applied which collect as much dull light as possible, emitted in delayed fluorescence paradigms. The lens of the present invention can also be configured for fluorescence polarization studies, by the addition of a ring polarizer at the light output of the inner ring, and an analyzer within the lens barrel.

Characteristics of the Analysis Software Findings of Objectives An image contains both objectives and background. Frequently, there is some ambiguity about which parts of the image make up the objective component. The extensive prior art deals with the automated extraction of the objective data from the images. The extraction process is known as "segmentation". Typically, segmentation algorithms use density, color, texture, or other image characteristics to define a valid goal. Any pixel that meets the objective definition criteria is designated as an objective pixel. The present invention performs the segmentation of the usual classes, using any combination of image characteristics, in addition to the novel segmentation forms. The usual type of segmentation for free format specimens is preferred. Typically, targets are defined by their luminance value, and then automatically detected at any points within the image. For well imaging, hybridization configurations, and other specimens in a regularly spaced (fixed) format, the present invention uses a novel segmentation form, which allows very large numbers of targets to be analyzed (for example, example, 60,000), with full automation. The novel segmentation procedure uses a "fixed probe" strategy, in which a configuration of probes of constant size is placed on the sample. For this purpose, an analogy is made with the liquid scintillation count. In the liquid scintillation counting of isotopically labeled specimens it is common to cut (using a circular punch) fixed size membrane pieces. These pieces are placed inside fruteos of scintillation fluid for counting. As a result of the count, you get to know the total amount of radioactivity in each sample piece. Similarly, with the present software, a configuration of fixed probes is applied to the image, at intervals specified by means of the grid definition. Positioning software is required, because the configuration of the computer-generated probes is unlikely to match exactly the true locations of all the targets in the specimen. There is variation in the position of the target because the spotting robots are not perfect, the plates may vary slightly in position, and so on. Therefore, before quantification, the image is processed and analyzed to rationalize the location of each objective within the configuration. Essentially, the algorithm searches for any targets that can be recognized on the basis of different densities. This adapts these objectives that can be recognized within the most likely locations within the grid, using a confusing logical algorithm. Next, it adapts the least accurate points to its most likely locations, using the highest density points as anchor points for alignment.

For the user, the procedure seems automatic. Once the fixed probe configuration has been defined in the specimen, the computer performs automatic alignment of the configuration elements to the most appropriate positions in the array. You can manually edit any elements that remain misaligned.

Statistical Segmentation The luminance data of the aligned matrix are taken to the software, which creates a data distribution of all fixed probes. Typically, the mean and standard deviation for this distribution are calculated, and the objectives can also be defined as "hits" using a variation criterion for the distribution. For example, we can define the objectives that are more than four units of standard deviation of the middle of the distribution as hits. This use of definition of success based on variation is called "statistical segmentation". Statistical segmentation requires that every sample in the matrix be analyzed, even if it can not be detected by means of standard segmentation methods. Therefore, when used for well plate tests or other grid assays, statistical segmentation requires the use of aligned, fixed probes as described above.

The Calibration of Measurement Efficiency The present invention implements a novel means by which the measurement efficiency can be corrected. This allows the correction of variations in the efficiency of lens or camera collection. In practice, the system sees a set of reference standards that encompass all or part of the range of intensities observed in real specimens. Typically, these standards will be well plates of the same kind, and they contain a substance that responds to the same wavelengths as those to be used for the test in question. Preferably, three (or more) plates to contain low, medium and high concentrations of light emitting substance. The use of multiple reference concentrations allows non-linear variations in the measurement efficiency that will be accommodated. Once the reference plates have been seen, the system can derive discrete models for the measurement efficiency in each well. That is, reference plates allow us to create, at each spatial location in the image, a known relationship between the intensity of light observed by the system, and the concentration of light-emitting substance that is actually present. The system uses this data to model and correct subsequent measurements obtained from each well. The result is a low level of error variation between the wells. Typically, the present invention produces if- Hr. ~ 7l coefficients of variation that are very similar to those obtained from non-imaging counting systems. This use of a correction for well-to-well variation in measurement efficiency is highly desirable, if an area imaging system will achieve acceptable accuracy in quantification wells.

BRIEF DESCRIPTION OF THE DRAWINGS Other objects, features and advantages of the invention will be more fully understood by the following detailed description of a currently preferred, but nevertheless illustrative, mode, with reference to the accompanying drawings, in which: Figure 1 is a schematic illustration of a system according to a first preferred (straight) embodiment of the invention. Figure 2 is a detailed illustration of the optical and mechanical components of the lens and the emission filter holder. Figure 3 illustrates schematically the benefit of the telecentric lens. Figure 4 is a schematic illustration of how reflections that would be detected in the CCD are avoided. Figure 5 is a schematic diagram of an imaging system incorporating an intensifier of *. * 3d image.; Figure 6 is a schematic representation of the image intensifier. Figure 7 is a flow diagram illustrating the main data analysis process. Figure 8 is a flow chart showing the definition and alignment of the process grid. Figure 9 is a flow diagram illustrating a process for establishing well-for-well calibrations for measured efficiency.

Detailed Description of the Preferred Modes Turning now to the details of the drawings, Figure 1 is a schematic diagram illustrating a preferred embodiment of an imaging system 1, in accordance with the present invention. The system 1 broadly comprises a lighting subsystem 10, an imaging subsystem 12 that is provided in a housing 14, and a control subsystem 16. The image subsystem 12 comprises a CCD camera subsystem 18 hosted inside a chamber 20 of the housing chamber 14, and a lens subassembly 22 extending between the chamber 20 of the chamber and a specimen chamber 24. In operation, the lighting subsystem 10 allows the energy of the Necessary light to the specimen inside the chamber 24. The light energy emitted by the specimen is transmitted through the lens subsystem 22 to the chamber 18, where an image is formed and transmitted to the control subsystem 16 for processing. The control subsystem 16 comprises a camera control unit 26, which is a conventional unit adapted to the particular camera 18, and a computer 28 which is programmed to control the unit 26 and to receive data from the camera 18, for the purpose of to achieve unique control and processing in accordance with the present invention. The light source for the lighting subsystem 10 is preferably an arc lamp 30. An optical fiber 32 carries the illumination from the optical coupler 34 to the ring light 44 inside the lens 22. Alternatively, an liquid or fiber light guide (not shown) at the outlet of the lamp 30, and can be taken to an individual optical coupler (not shown) and from there to the optical fiber 32. The X filter wheel contains a number of interference filters one inch in diameter. It can quickly change the position of the filter wheel under the direction of the computer, giving the ability to drive procedures that require more than one excitation wavelength. The bundle of optical fibers 32 carries the illumination of the optical coupler / filter wheel 34 to the ring light inside the lens 44. The bundle 32 does not need to pierce the wall of the specimen chamber 24 light-tight. Two forms of lighting system are described, each fed by means of a discrete bundle of fibers. These are a transillumination plate (not shown), which is used for absorbency or colorimetric tests, and a ring light 44 internal to the lens (22) that performs epi-illumination for the fluorescence assays. The inner ring light 44 consists of one or more rings of optical fibers, mounted in and axially aligned with the body of the telecentric lens 22, and behind its front lens element. It is convenient to have multiple rings, in the sense that a ring can be optimally angled for even illumination over the majority of the specimen, while the second ring can be optimally angled to send additional illumination to the edges of the specimen. specimen. On the front of the illumination ring 44 is placed an input polarizer to supply the polarized excitation. By means of placing the polarizer near the fiber optic output, any effects of the polarizing fluorescence are minimized. That is, only the emitted fluorescence that is reflected on the surface of the front element of the telecentric lens assembly has the potential to reach the CCD detector. The input polarizer is ring-shaped, with a hole in its center that allows light to travel to the lens barrel unimpeded. A second polarizer works like an analyzer (not shown), and can be placed in any position inside the lens barrel. A preferred location is in the plane of the excitation polarizer. The well plate of the specimen is carried inside a holder 58 which is mounted to a focus controller 60. The holder 58 grasps the plate and wells on its edges. The bottom of the bra 58 is empty, with the purpose of not preventing the vision of the wells. By adjusting the focus controller 60, the holder 58 moves relative to the lens 22 and the specimen is focused. The lens 22 is telecentric with a wide opening. The lens contains an emission filter slot 62, which accepts interference filters for fluorescence imaging. This contains a ring light 44 of internal optical fiber, placed behind the front lens element. The lens 22 is mounted to the camera of the camera by means of a flange 64, close to its middle part. The rear part of the lens projects into the chamber 20 of the chamber, providing easy access to the slot of the emission filter 62 without disturbing the specimen. The front of the lens projects into the specimen chamber 24. The cooled CCD camera 18 is mounted with the detector positioned as close as possible to the lens. The minimum spacing between the final lens element and the CCD is important for an efficient transfer of light from the lens to the CCD. The entire system 1 is built to achieve high transfer efficiency. Three factors dominate the transfer efficiency (generated by photoelectrons / emitted by photons) of the detector system. These are the lens light collection efficiency, the efficiency of the CCD detector's quantum, and the transmittance of the lens. We can calculate the number of photoelectrons generated as follows: Npe =? * detector * -e- * Nfotoñes where? is the transmittance of the lens, approximately 85-90 percent for our lens, where f is the efficiency of the quantum of the detector CCD, typically about 35-40 percent, up to 80 percent in our case, c.e. is the efficiency of lens collection, approximately 0.3 percent in our case. In a typical scientific grade CCD camera system, using the fastest available photographic lens (F / 1.2), and with a high quality cooled detector, the CCD will generate 1 photoelectron for approximately 5,000-10,000 photons generated from a point source in the sample. The lens of the present invention offers a collection efficiency of approximately that of a photographic lens F / 1.2. The efficiency of the CCD detector is approximately twice that of other CCDs. The result is that the present invention has the theoretical ability to generate a photoelectron for approximately 2,500-5,000 photons generated from a point source within the sample. This very high transfer efficiency allows the detection of specimens from which images with the prior art systems can not be formed.

Control Subsystem The control subsystem 16 comprises the control unit 26 and the computer 28. The camera control unit is a computer controllable unit provided by the manufacturer of the camera 18, to control the camera. The computer 28 is preferably a conventional computer operating in the Windows® environment, and is programmed to achieve image acquisition and analysis in accordance with the present invention. All control, imaging, and analysis functions are resident inside the computer 28.

Illumination Subsystem The preferred configuration of the present invention uses a broadband illumination source, such as a halogen lamp or a mercury-xenon arc lamp. The advantage of an arc lamp is that its output can be done within a narrow beam, which can pass through a small and easily available interference filter, before being spread over the entire surface of the specimen. Any wavelength of monochromatic illumination can be selected by means of precision filters (usually interference filters), mounted after the lamp. Filters are preferred because variable monochromators or tunable lasers lack sufficient light output when they are diffused over large areas. Although the present invention can be configured to apply diffuse transillumination (through the specimen) or other forms of dorsal or lateral illumination, epi-illumination is preferred because it usually results in lower backgrounds, a wider dynamic range, and a response of more linear fluorescence. Figure 2 shows better the optical components of the telecentric lens 22. The lens has the following characteristics: Effective focal length 164.436 mm Numerical aperture 0.443 Magnification 0.25 The lens has mounted inside it a fiber optic ring light 44, which projects excitation to through the front lens element on the specimen (to the left in Figure 2). The lens is calculated so that the focus plane of the ring light is at B, while the plane of focus of the entire lens is at the front of that point, at A. The focus placement of the ring light at a point beyond the specimen minimizes specular reflections of the specimen. The emission filter slot 62 allows the insertion of an interference filter (a barrier filter) that removes the excitation illumination and other unwanted light from the incoming rays, leaving only the fluorescence or other relevant signal emitted by the specimen. The barrier filter is coated to minimize reflections. A critical advantage of the present lens is that the internal beam is aligned at an appropriate position for the insertion of a barrier filter. In this position, the rays are almost parallel. The ray aligned at this point is critical, because interference filters work best with rays that are incidents at 0 degrees. If the incoming illumination is at an angle, the filter exhibits both the widening of its transmission characteristics, and alterations in the wavelengths that it passes. It will be noted that the lens of the present invention implements a unique form of macroepiilumination, without the need for a dichroic mirror. In the microscopy of - > & £. epifluorescence, where small areas are illuminated, usually a dichroic ray splitter is placed behind the target. A dichroic ray splitter or mirror is a partially reflective surface that reflects a range of wavelength, while allowing it to pass through another wavelength range. In a fluorescence microscope, the illumination enters the dichroic mirror from the side. The mirror is reflective with respect to the excitation illumination, and is angled to reflect the excitation light down through the objective towards the specimen. The fluorescence emitted by the specimen (changed in the wavelength from the excitation) is collected by the target, which passes it upwards towards the dichroic mirror. The dichroic mirror is transparent to the emission wavelength, in such a way that the light proceeds through the dichroic to the detector plane. A different dichroic is required for each excitation / emission wavelength. There are major difficulties in applying a dichroic-based epi-illumination system to image macro-formation: 1. The dichroic mirror should be at least as large as the objective it must fill. Camera lenses are much larger than microscope lenses, and would require correspondingly large dichroic mirrors. Dichroic mirrors this big are not easily jfe- - »available, are expensive, and are subject to internal variability. 2. It is critical that the rear lens element of a macrolens be mounted as close as possible to the CCD. Any increase in the distance between the rearmost lens and the CCD markedly impairs the work number F and the light clustering efficiency. Therefore, there is no place for a dichroic to be mounted behind the lens, and the dichroic would have to be mounted inside the lens itself. Although dichroic ones have been mounted inside the microscope's objectives, dichroics mounted inside a macrolente have not been described. Our own tests with one of those configurations have revealed difficulties to achieve low funds and even lighting. 3. In a normal epi-illumination system, the dichroic reflects excitation through the entire lens. This has two disadvantages. o Each lens adds its own autofluorescence to the total background. This may not be a major problem with the small lens and the very bright lighting levels found in microscopy. With macrolenses and at lower light levels, however, the autofluorescence of the glass becomes a major component of the total signal that the detector sees. a The transmission of the excitation lighting is highly subject to the optical characteristics of the glasses used in the lens. Very expensive quartz glass optics (and difficult to work with) are required for ultraviolet ray epi-illumination. These UV-transparent optics can be constructed in the small sizes needed for a microscope objective, but they would be astronomically expensive in the large sizes described for the present invention. 4. Dichroic ray splitters absorb light. Typically, these are 80-90 percent efficient. The present invention operates without a dichroic mirror. Rather, the illuminator 44 is mounted in such a way that it shines directly on the front lens element X, from behind. This illuminates the specimen, without any need for a dichroic mirror. Illumination from within the lens is possible as a result of the novel features of the present invention. These characteristics include: a) the calculation to avoid the detection of internal reflections and autofluorescence; b) the calculation for the effective use of an emission barrier filter, located inside the lens and after the source of illumination; c) design in such a way that only one of the lens elements resides in front of the internal illuminator.

It is a key advantage that the lens of the present invention has only one X element in front of the illumination source 44. This feature has the advantages that reflections, reflections, and autofluorescence of the glass are minimized, and that only the element Frontal needs to be transparent to ultraviolet rays. A single lens transparent to ultraviolet rays is expensive, but it is not prohibitively so. An additional advantage of having only one lens in front of the lighting system is that the generation of a polarized lighting beam is made with much less difficulty. A polarizer can be placed in front of the light source, and only light has to pass through the front lens element before affecting the specimen. In the present invention, the front lens can be specified so that it has minimal effects on a polarized beam of light. If more lenses resided in front of the lighting system, it would be much more difficult to achieve minimal effects on polarized light. The front lens element is calculated in order to focus the light source at B, beyond the plane of the specimen at A. Defocusing of the illumination source on the specimen plane minimizes specular reflections. Since many well plates are constructed of polished plastic, and tend to generate specular reflections, this is an important feature.

The lens is highly efficient. A typical collection F / # of the lens is 4.5 (F / 1.15 of work). This implies a solid collection angle of 0.03891 sr, and a collection efficiency of 0.03891 / 4p = .3096 percent. For this typical lens, the expected transmission value is 0.85-0.90, giving an overall collection efficiency of 0.263-0.279 percent. This is almost equivalent to the efficiency of collecting a photographic lens of F / l.2. The present telecentric lens has a fixed field of view (12 centimeters in diameter, in this case), but if images of larger specimens need to be formed, a motorized translational table can be mounted inside the light-tight chamber. The translation table moves the specimen in relation to the lens, under the control of the computer. After each movement, only one "tile" is acquired. When the image of the entire specimen has been formed, all the tiles are remade (by means of the software) into a single large image, retaining the telecentricity, the freedom of parallax error, and high resolution over its entire surface. The lens can also be constructed in different sizes and magnification factors to allow images of larger or smaller specimens to form over the entire length of the CCD. When used with small, high-density DNA configurations, for example, the lens can be constructed without telecentricity, and at a much higher magnification factor in order to form the image of the high resolution configuration in a single acquisition. With this calculation of the new lens, the key principle for using epifluorescence to illuminate the entire specimen would be retained. Figure 3 shows better the telecentric property of the lens 22. Although an epi-illuminator lens can be constructed in accordance with the present invention, and without telecentricity, a telecentric lens collects parallel rays, over the entire area of a well plate. In this way, it does not scrutinize the interior of any well at an angle, and is free of parallax error. In contrast, images of deep, narrow lenses, made with standard lenses, exhibit severe parallax. That is, the geometric distortion increases at the edges of the field of vision. With a non-telecentric lens, circular lenses (such as wells) in the center of the image look like real circles. However, the lens scrutinizes the interior of the lateral lenses at an angle. Thus, these lateral objectives are seen as semilunar shapes, and the lens does not manage to form images of the sides of the lateral wells. Figure 4 shows the path of the light rays from the light of fiber optic ring 44. The light from the ring 44 strikes the front lens X element, which is coated to allow most of the light to pass through to the X specimen. However, a small portion of the incident light is reflected to the front lens of the specimen. back to the CCD-84 detector. Although these reflections are much less intense than the light source (eg, <0.2 percent), these are very intense in relation to the emission of weak fluorescence of the specimen. That is, the fluorescence from the specimen may well be in the order of 0.1 percent of the excitation intensity. Therefore, reflections inside the lens could conflict with the fluorescence from the specimen, and would degrade the sensitivity of the imaging system. In the present invention, reflections from the lens surfaces in front of the illumination source are not detected. In calculating the fate of the rays that are reflected from the front X-lens element, it can be determined that only some of the reflections that arise from the X-lens hit the CCD 84. For example, an X-ray emanating from the X-ray is shown. ring light 44 in position A, and striking the X lens in position B. The X reflection of this X ray is shown hitting a point C that is not found in the CCD 84. However, the rays from the position A in the ring light 44 that struck the front X-lens in the D position would be reflected to hit the CCD 84. The lens of the present invention contains a mechanism that blocks the excitation rays that would strike the front X-ray Bl in locations that they create annoying reflections. The rays of light emitted from ring light 44 are emitted at an angle (5 degrees in this case) directed inward toward the center of the X lens. Normally, these rays would strike the entire surface of the X lens, generating reflections that would be highly evident in the image formed in the X CCD, and that would degrade the sensitivity of the system. These undesirable rays are blocked by means of a lip X, placed as a ring just inside the ring light 44. Indeed, the lip X shades the front lens X of the rays that would generate reflections. The rays that are allowed to pass the lip X still generate internal reflections to the lens assembly 22, but they die inside the body of the lens assembly 22, or in the body of the camera assembly 18. These do not reach the CCD X. As a result of these novel design features, the present lens is capable of delivering an intense epi-illumination beam, without the presence of detectable internal reflections. Rather, the excitation angle incident to the front lens is calculated in such a way that its reflections strike the internal surface of the camera assembly X. In this way, the reflections die inside the camera body or the lens mechanism, without reaching the CCD X detector. Figure 4 shows better the cooled CCD camera. The camera 18 includes a CCD element 84 positioned behind an opening in the chamber. A particular problem that arises with the lenses of the present class (wide opening, rear element placed very close to the plane of the detector) is that the depth of field in the plane of the detector is very narrow. The lens of the present invention, for example, requires that the CCD be plane with respect to the optical axis of the lens, with a tolerance of 50 microns. Therefore, the present invention both implements and verifies during manufacture, to ensure that the surface of the CCD is flat, and a tilt adjustment at the tip (not shown) inside the body of the camera, which allows the CCD Adjust until it is flat. To reduce the dark noise produced by the electrons inside the CCD, the CCD element 84 is mounted to a heat drain 88, which in turn is thermally coupled to a cooling element and liquid circulation system to provide improved heat dissipation. . A lens 90 is placed over the aperture to focus the image on the CCD element 84. The rapid telecentric lens 22 (Figures 2 and 3) is mounted directly to the camera body by means of screws, after removing the photographic lens assembly . Similarly, the image intensifier 70 (when present) is mounted directly to the body of the camera. Preferably, the present invention uses a CCD configuration with high quantum efficiency, low interpretation and dark noise, and wide dynamic range. The quantum efficiency (QE) describes the ability of the photodetector to convert the incident photons to pairs of electron holes in the CCD. Consumer-grade CCDs typically exhibit an efficiency of 12-15 percent. The cooled, scientific grade, CCD cameras exhibit an efficiency of approximately 40 percent. A very limited number of thinned CCDs illuminated from the back can achieve a quantum efficiency as high as 80 percent at peak detection wavelengths. Interpretation noise originates from the CCD's pre-output amplifier, which measures small changes in the voltage produced each time the charge content of one or more CCD elements is transferred to it. Interpretation noise is directly related to the speed of interpretation, and is diminished by the use of slow interpretation. The dark noise is produced by thermally generated loads in the CCD, which increase the level of the bottom. A constant dark noise level can be subtracted from the image, but dark noise also has a random noise component (a variation in level) that can not be subtracted. This random component adds to the total random noise level of the detector, and has a direct effect on sensitivity. The dark noise is decreased by cooling the CCD. The size of the CCD element (typically approximately 2.25 square centimeters) is related to its ability to store photoelectrons (known as well capacity) and, consequently, its dynamic range. The larger each CCD element (pixel) in the configuration, the larger the full well capacity and the dynamic range of that pixel will be. A wide dynamic range is desirable, in the sense that it allows the detector to be used during longer exposure times, without saturation. This improves the detection of very small signals. Another advantage of large pixels is that the signal to the noise operation is improved, relative to that of the smaller pixels. Even so, most image forming systems by area use CCDs with small pixels, both because the cost is lower and because the resolution depends on the size of the pixel. Large pixels sacrifice resolution to gain signal. Small pixels sacrifice sensitivity and dynamic range to gain resolution. In short, CCD cameras have tended to negotiate sensitivity to light against resolution. To achieve high sensitivity, good resolution, and wide dynamic range, CCD cameras used in astronomy often incorporate a large CCD, which contains a large number of large pixels. The present invention (preferably) incorporates an unusually large CCD configuration (eg, 6.25 square centimeters), which contains many (preferably >1 million) large pixels. In addition, the CCD is selected for the highest possible efficiency (preferably about 70-80 percent). As a result, the detector of the present invention is characterized by having very high sensitivity, wide dynamic range (16 true bits), and high resolution. This synergistic combination of features is novel in its application to the formation of test images. To take full advantage of the CCD's features, preferred support electronics include a high-precision digitizer, with minimal interpretation noise. The camera provides a 16-bit digital signal output by means of the digitalization circuitry mounted inside the camera's control unit, and an interface card mounted inside the computer. The data from the CCD is digitized by means of the camera's control unit, digitized at appropriate speeds for the required sensitivity (for example, 200,000 pixels / second), and transferred directly to the computer's memory. After the integration period, the CCD camera accepts an activation pulse from the computer to start closing the electromechanical shutter. With the shutter closed, the image is transferred from the CCD to the internal arc buffer area of the computer. Although this chamber can be used without the cooling of the CCD element, extended periods of integration are achieved by the use of a CCD camera with an integral cooling element. The effectiveness of integration is limited by the degree of cooling. With an uncooled liquid cooling device, detector temperatures of about -50 ° C (below room temperature) can be achieved. At this temperature, dark noise accumulates at a typical velocity of about 7-10 electrons / second. This type of cooling has the advantage of low cost and easy implementation. It should be appreciated, however, that longer integration periods are possible if refrigerated liquid or cryogenic cooling is used. Figure 5 employs the optional intensifier for extreme low light and time resolved modes of imaging. The intensifier 70 is mounted between the telecentric lens 22 and the CCD camera 18. Intensification, as described, for example, in US Pat. No. 5,204,533 to Simonet, involves the coupling of an image intensifier to a camera. CCD The image intensifier typically includes a photocathode, S. fe 'a phosphor screen', and a microchannel plate (MCP) connected between the photocathode and the phosphor screen.With this type of device very high light amplification factors are available (for example, up to about 90,000 The intensifier also provides shutter capability for time resolved and photon counting applications With the intensifier inserted within the optical chain, the present invention becomes an intensified image CCD camera (ICCD) In all other respects the system is essentially identical to that of Figure 1 for extreme low light imaging.Figure 6 best shows the intensifier 70 as being of the GEN 4 type, and which includes a photosensitive cathode 72, a microchannel plate (MCP) 74, a phosphor screen 76, and a vacuum sealed body or housing 78. The epi-illuminator lens 22 (Figures 2, 3) is positioned in front of this assembly 70. In this In the output, the lens is focused on a cathode entry window 72, in order to transfer the specimen image thereto. The photosensitive cathode 72 is selected to emit electrons in proportion to the intensity of the light falling out of it. The microchannel plate 74 is placed inside the vacuum sealed body 78, between the cathode 72, and the phosphor screen 76, and coupled to the cathode 72 at each end. The microchannel plate 74 is provided with a small-channel microchannel plate channel configuration, each of which is coated with gallium arsenide. The electrons emitted from the cathode 72 are accelerated along the channels of the microchannel plate to the phosphor screen 76. As electrons are accelerated from the cathode along the small diameter channels, they strike the walls of the coated channels to produce additional electrons. As the multiplied electrons leave the channels of the microchannel plate, they strike the phosphor screen 76, and produce an intensified image of the specimen in an output window. This image is coupled to the CCD element 84 in the camera by means of a lens 80. It has been found that the use of low noise image intensifiers (e.g., GEN 4) is convenient over other types of intensifiers. It should be appreciated, however, that many types of intensifiers can be used (for example, those with cooling of the photocathode) with the present system. Similarly, devices with high intrinsic gain (such as back-illuminated CCD detectors bombarded with electrons) can be used instead of image intensifiers. The integrating chamber is configured in such a way that the highly amplified image, generated in the output window, is carried by an optical fiber on the CCD element. To form the images of low light specimens, the CCD element w r & is integrated during u? - * period. During the integration period, photons from the output window, incidents to the CCD element, are stored as negative charges (the signal) in numerous discrete regions of the CCD element. The amount of charge in each discrete region of the CCD element is accumulated as follows: Signal = Incident light x Quantum efficiency x Integration time. The greater the relative intensity of the incident light coming from the intensifier, the greater the signal stored in the corresponding region of the CCD element. With the system shown in Figures 5 and 6, only the CCD detector is cooled. This is sufficient for most purposes. It also has to be appreciated, however, that the photocathode of the intensifier can also be cooled, thereby improving the signal to the noise ratio of the intensifier. Similarly, the entire photosensitive device (intensifier + CCD) can be cooled. Nevertheless, the cooling of the entire photosensitive apparatus has the disadvantage that the phosphorus efficiency in the fiber optic output window is decreased. The CCD camera of the present invention uses an asynchronous reset that takes an external controlling signal from the computer. Since the CCD camera incorporates mechanisms that provide very low delay, short integration periods (eg, 1/100 second) can be used. If desired, these integration periods can be locked to a regulated power supply by gate in the image intensifier, with the result that the camera can be interpreted at very short intervals. Using the gate-regulating and fast-acting feature, and with the intensifier operating at a higher gain or with a multi-stage intensifier, the present invention can be operated by the same as a conventional photon counting chamber. In this way, the present system can be conveniently used both for direct imaging of pale specimens, or as a photon counting chamber, by changing its integration operation mode to gate regulation. Figure 7 is a flow chart illustrating the main data analysis process performed by computer 28 to control system 1 and acquire data therefrom. After the initiation of the process, a specimen is acquired for analysis in block 302, as described more fully below. When creating this image, the specimen is illuminated and acquired in such a way as to fully demonstrate the effect of interest. This image is referred to as the specific image. It will be noted that in the specific image, and in any other images acquired by system 1, there are many routine operations to remove errors (e.g., phase shift and polarization in the CCD) in the digital image formation system. In the present these routine options will not be shown. A grid consists of a probe template configuration. The "probe template" is the nominal definition of a single objective (for example, in terms of shape and area) that corresponds to a point on a membrane, a well on a plate or similar objective. Typically the probe template is a circular area, and there is a probe template for each objective in the specimen. A matching grid for a specimen image has all its probe templates matching the targets in the specimen image (whether they are visible or not). This correspondence grid is also created in block 302, through a process described more fully later. The system can simply analyze the specific image. Optionally, in block 304, an image of the specimen is acquired under conditions that best show the exact locations of the wells (typically illuminated with white light), and / or conditions that best show the filling in the discrete wells, while affecting minimal way by the reaction of interest. These images, which are used to improve the quality of data obtained from a specific image, are called "second-sight" images. A white light image can be used to create a grid that matches the specimen. Basically, the location of each well must be established, so that the data can be extracted from them. The small errors of superposition of errors between the well and the sampling opening can be confirmed seriously. The alignment procedure described above can be used, if there are reasonably good intensities from the specimen. If not, the plate itself is used as a reference, by capturing an image of the plate under dim light. This image shows where all the wells are, and this positioning information can be used to guide the extraction of data. Another form of second-sight image is used to correct for variation in filling. This is, the sample dosing systems do not put the same amount of sample into each well. In some cases the dosage error leads to large and irrelevant variations in observed intensities. To remove this error, a second fluorite is dosed. The intensity of the first fluorite is related to the reaction of interest. The intensity of the second fluorite is not related to the reaction of interest. Rather, its intensity must be the same from well to well. Variations in the intensity of the second fluorite can be used to correct variations in dosage. "In block 306, the second view image is used to find the best matching grid, remove non-specific background fluorescence, or to verify the consistency of the fill In block 308, an optional process is applied to correct the Variations in measurement efficiency This well-for-well calibration process can be used to correct uneven lighting measurements and / or harvesting efficiencies Well-by-well calibration can also be used to measure directly The concentration of the target substance is shown in Figure 10. The establishment of well-by-well calibration is illustrated in Figure 10. A second-sight image, if available, can provide a confidence factor for each objective. The system aligns the sampling grid with the specimen, measures all the targets in the specimen, and makes the data available.In block 312, the measurements taken from each objective discrete in the configuration of the specimen are decoded to different conditions. For example, a discrete target may be able to assume any of n conditions, and the process of block 312 may decode the sample in each probe to one of those conditions. The actual process is done using a simple intensity criterion (for example, goal X is more bríllLánte than 100 and is placed in category 1). At the discretion of the user, the decoding of the objectives to the conditions applies the process of statistical segmentation. In accordance with this process in block 312, the actual levels in each objective are measured within the specimen configuration. A mean and a standard deviation are determined for the set of samples, and this results in a statistical distribution of work. Then you can decode each discrete objective to a category, by using its position within the distribution of all the objectives. Typically, this decoding is based on the objective found in a calculated number of standard deviations from the mean. For example, an objective that is more than 5 units of standard deviation from the mean can be decoded to category 1. In block 314, a process is performed to generate a report of the configuration data, based on the process performed in block 312. It is contemplated that this may be any form of report writing software, which provides the operator with a substantial amount of flexibility to prepare reports of a desired format. Once the reports are generated, the data analysis process ends. Figure 8 is a flowchart illustrating the process performed for the acquisition of images and for aligning sampling grids with the lens configurations. After the initiation of the process, a specific image is acquired. There are known processes for acquiring skewed images of a specimen. These skewed images take into account all the distortions and significant errors introduced by the system itself when taking an image. In blocks 320, 322 and 324, the biased images are acquired for the specific, non-specific and second-sight images. These skewed images are acquired by using one of the known methods. Once the skewed images are acquired, these are stored and used thereafter, unless the environment of the system changes as a result of a replacement of the light bulb, and so on. In block 502, the specific image of the specimen is acquired and in block 508 the skewed image is removed. In block 504, the non-specific image of the specimen is acquired. This image determines the contribution of non-specimen components, such as the support substrate, to the image. This step is indicated as optional, since this would only be done in case the specimen had to be illuminated, in order to acquire the specimen image, in which case some light would also be reflected from the non-specimen elements. On the other hand, if the specimen were the source of light for the image (as in the € 6 chemiluminescence), the non-specific image would not be acquired. In block 510, the non-specific biased image is removed. In block 506, the second view image of the specimen is acquired, and in block 512 the skewed images are removed. This produces the second view image referenced in block 304. If the non-specific image of the specimen is available, in block 514 it is removed from the specific image of the specimen. This produces the specific image referenced in block 302. In block 516, if the second view image is not available, the specific image is used to create the correspondence grid. If the second view image is available, and has been acquired for grid alignment, the second view image will be selected. Otherwise, the selection can be made based on some known criteria, or the selection can be made simply by examining the images. In block 518, the system can automatically select one of a plurality of predetermined predefined grids (330) using known criteria, or a user can manually create a new default grid. The "default grid" represents a grid that contains a probe template configuration with uniform row and column spacing.

In the ^ ifeo block, a grid alignment process is performed, described more fully by probe with probe can be aligned with their corresponding objectives by means of directly aligning each template with its objectives, if most objectives contain signals detectable or if a second-sight image is used. Otherwise, the operator can define a configuration of "anchor points", which will help the system to locate the positions of the probe templates. It will be noted that, in some cases, it may be desirable to align the configuration using both methods. At the end of this process, a specific image of the specimen, a correspondence grid, and a second-sight image (optional) have been generated. The processes of goal or anchor alignment are essentially the same. In the case of anchor alignment, once the anchors are aligned, the system uses a known regression method to align the rest of the probe templates with the aligned anchor points. Figure 9 is a flow chart illustrating the process performed in block 520 of Figure 8. After the initiation of the process, in the block 602 background and image noise are estimated. In block 604, a determination is made as to whether a group alignment is necessary with the configuration of the objectives. This is either visually an operator, or through the system. The purpose of this test two patterns or regular shapes. If it is determined that there is adequate alignment of the group, the control is transferred to block 608. In block 606, a group alignment is performed. The purpose of this procedure is to align the grid of the probe template approximately with the respective objectives. The alignment can be done by means of the process described later with respect to block 608 to maximize the CS, except that the CS is maximized across the entire grid. In block 608, a stepwise process is performed within the area of each individual probe template, to locate that point that produces a maximum reliability score (CS), inside the probe template as determined by equation (1) ): CS (xO, yO) = (1) E-A ^? where: a is an eration (form) in the range of [0, 1]; Nf is a function &. normalization, which converts a value in the range of [-8, +8], to a new value in the range of [0, 1]; (x0, y0) is the center point of a probe template; S (x0, y0) is the area of the probe template at (x0, y0); D (x, y) is the density value (for example, brightness) in (x, y); and W (x, y) is a weighting function (for example, a two-dimensional Gaussian function with its maximum value in (0,0)). D is the average density value within S (xo, yo). W is the average value of the weighted function. The first term in equation (1) represents a score for the density values of the pixels within a discrete objective, while the second term represents a score for the correspondence of the density profile. The density profile will correspond to its expected shape closely as the noise decreases, such that the second term is very sensitive to the presence of noise and other artifacts inside the specimen contained in the ,; tc ^ & probe template. The first and second terms of equation (1) are added in a weighted way based on the weighting factor, a, which is a shape weighting factor. Preferably, the weighting factor is set by the system based on the estimated noise parameters. Alternatively, the weighting factor can have a default value and the operator can adjust it downward in the event that the process in Figure 9 does not produce good results. Equation (1) produces a "location A" for each probe template, which is that location that provides the maximum value in equation (1). Reference will be made to the location of the probe template before block 608 as "location G". In block 610, a comparison is made between location A and location G, in order to reach the final location of the center of each probe template. In effect, the weighting factors are used to determine the position of the center of the probe along the path between locations A and G, with the weighting determining how close the point is to location A. Preferably, this was done in a stepwise manner by moving from location G to the location that has the highest value for equation (1), only if that value exceeds a previously determined threshold confidence criterion, such as exceeding "* r the value of equation (1) at location G, by means of a previously determined amount or proportion As the threshold criterion increases, the movement of location G becomes less likely. illustrates the process of block 308 for establishing well-for-well calibrations for measurement efficiency.In block 702, the system acquires a set of reference plates to be used for calibration. reference plates, and to align sampling grids with the reference plates are as shown in Figures 7-9 Each well in a reference plate is filled with the same concentration or a known concentration of a substance. The set of reference plates has a different concentration of the substance, Typically, the substance will have similar fluorescence or luminance characteristics to those of the substance used for the substance. ticket the reaction of interest. In block 704, numerical data are obtained for the gray level of each well, in each reference plate. As a result of block 704, computer 28 has a set of data pairs (gray level value and concentration value for each reference plate) for each well. For example, well # 1 produces a gray level value of 100 on reference plate 1 with concentration 10, of 200 on reference plate 2 with concentration 20, and so on. In block 706, the calibration is created for each well by adjusting a regression, interpolation or other function to the data pairs associated with that well. As a result, each and every position of the well is calibrated to the relationship between the gray level value observed, and the concentration in a well plate. This calibration can then be used to classify and report data as shown in blocks 312-312 of Figure 7. Although the detailed description describes and illustrates the preferred embodiments of the present invention, the invention is thus not limited. Modifications and variations that do not depart from the scope and spirit of the invention will be apparent to those skilled in the art, as described by the appended claims.

Claims (15)

1. A method for use in a digital imaging system for tests to extract targets in a specimen, containing a configuration of objectives that may not be configured in perfect regularity, the method comprising the steps of: defining a matrix of locations of nominal targets, which includes a probe template of previously defined, two-dimensional size and shape in each of a plurality of fixed grid points, previously defined in the specimen; and determining the most probable location of the probe template corresponding to a specific objective by detecting both the intensity of the pixel and the spatial distribution of the pixel intensities in a specimen image, in a plurality of locations in the vicinity of a nominal target location. The method of claim 1, characterized in that it also comprises the step of, using a confidence value indicative of the reliability of the most probable location as a weighting factor, changing the location of the probe template of the nominal target location to the most likely location for the selected target. 3. The method of claim 2, wherein the determination step- ^ - r? Dt? Iza repeatedly for each objective. 4. The method of claim 2, wherein the determination step detects the confidence score, CS (xO, yO), in said plurality • frd * e locations, defined by the following formula: CS (xO, yO) where: a is a weighting factor in the range of [0, 1]; Nf is a normalization function, which converts a value in the range of [-8, - + 8], to a new value in the range of [0, 1]; (x0, y0) is the center point of a probe template; S (x0, y0) is the area of the probe template in (x0, y0); D (x, y) is the density value (for example, brightness) in (x, y); and W (x, y) is a weighting function (for example, a two-dimensional Gaussian function with its maximum value in - - - * - - ---- - - ^ tea- - ^ SJ- ^ t ^. . . * .... ^ ^^ .- ^ < ^ S «4Jh ^ D is the average density value inside S (xo, yo). W is the promise value of the weighted function, the most probable location being the one that maximizes the formula. The method of claim 4, wherein the image that is used in the definition step is produced by generating a primary image of the specimen that shows the effect of interest with greater advantage, generating a secondary image that shows minimum the effect of interest but showing at least one of a second effect and the location of the wells. The method of claim 1, wherein the determination step determines the confidence score, CS (x0, y0), at the plurality of locations, defined by the following formula: CS (xO, yO) = (the) N, [D (x, y) W (x-x0, y-y0) dxdy JS (xO. YO) f D [(x, y) -25] [W ( x ~ x0, y-y0) -W \ dxdy + g J s) xo, yo) 77 [D (x, y) -D] 2dxdyf [W (x, y) -W] 2dxdy yj if o.yo) J5 (0,0) where: a is a weighting factor in the range of [0, 1]; Nf is a normalization function, which converts • áltiS ^ bb, a value in the range of a new value in the range of [0, 1]; (x0, y0) is the center point of a probe template; S (x0, y0) is the area of the probe template in (x0, y0); D (x, y) is the density value (for example, brightness) in (x, y); and W (x, y) is a weighting function (for example, a two-dimensional Gaussian function with its maximum value in (0,0)). D is the average density value inside S (xo, yo). W is the average value of the weighted function, the most probable location being the one that maximizes the formula. The method of claim 6, wherein the image that is used in the derivation step is produced by generating a primary image of the specimen, which shows with greater advantage the effect of interest, generating a secondary image showing minimal way the effect of interest, and combining the secondary image with the primary image. The method of claim 1, wherein said determination step is performed repetitively for each objective. The method of claim 8, wherein the image that is used in said bypass step is produced by generating a primary image of the specimen, which shows the effect of interest with greater advantage, generating a secondary image showing minimal way the effect of interest, and combining the secondary image with the primary image. The method of any of claims 1 to 4, wherein the specimen is provided with previously defined anchor points, the matrix being initially aligned with the actual target locations in the specimen, by placing the specific probe templates on the specimen. one of: anchor points; and those points of objectives that contain detectable signals. The method of claim 2, wherein the confidence value for a target is determined by means of the detectability of the target. The method of claim 4 or 6, characterized in that it also comprises the step of changing the location of the probe template from the nominal target location to the most probable location for the selected target, only when the confidence value exceeds a value threshold previously defined. 13. In a digital imaging system for assays, an apparatus for extracting targets in a specimen that contains a configuration of objectives that may not be configured in perfect regularity, comprising: a grid generator that provides a matrix of nominal target locations, including a probe template of previously defined, two-dimensional size and shape in each of a plurality of fixed, defined grid points previously in the specimen; and a probe template locator, jointly responsive to pixel intensity and spatial distribution of pixel intensities in a specimen image, at a plurality of locations in the vicinity of a nominal target location, to determine the most likely location of the probe template that corresponds to a specific objective. The apparatus of claim 13, characterized in that it also comprises a changer, responsive to a trust signal related to the reliability of the most likely location, to change the location of the probe template from the nominal target location to the location more probable for the selected objective. The apparatus of claim 14, wherein said probe template locator comprises a processor that determines the confidence score, CS (x0, y0), at each of the plurality of locations, in accordance with the following formula: ®f > where: a is a weighting factor in the range of [0, 1]; Nf is a normalization function, which converts a value in the range of [-8, +8], to a new value in the range of [0, 1]; (x0, y0) is the center point of a probe template; S (x0, y0) is the area of the probe template in (x0, y0); D (x, y) is the density value (for example, brightness) in (x, y); and W (x, y) is a weighting function (for example, a two-dimensional Gaussian function with its maximum value in (0,0)). D is the average density value within a valid weighting function range. W is the average value of the weighted function, the processor selecting as the most probable location being the one that maximizes the formula. ^^