WO2002099734A1

WO2002099734A1 - Automated image analysis for detecting microalgae and other organisms in the environment

Info

Publication number: WO2002099734A1
Application number: PCT/US2002/017587
Authority: WO
Inventors: Triantafyllos Tafas; Michael Kilpatrick; Petros Tsipouras
Original assignee: Ikonisys Inc.
Priority date: 2001-06-04
Filing date: 2002-06-04
Publication date: 2002-12-12
Also published as: CA2469850A1

Abstract

Computer controlled image analysis methods and apparatus (103-106) for detecting or detecting and quantifying microorganisms in a sample (109) from the environment are provided. In one preferred embodiment, the microorganism detected is an Alexandrium cell (417). The methods may further provide determination of the concentration of a substance or compound associated with the microorganism.

Description

AUTOMATED IMAGE ANALYSIS FOR DETECTING MICRO ALGAE AND OTHER ORGANISMS IN THE ENVIRONMENT

This application claims priority benefits of U . S . Patent Application No .60/295 ,586 filed June 4, 2001, the entire disclosure of which is incorporated herein by reference.

1. FIELD OF THE INVENTION

The present invention relates to computer controlled methods and apparatus for identifying or identifying and quantifying microorganisms and compounds from environmental samples using image analysis.

2. BACKGROUND OF THE INVENTION

Citation or identification of any reference in this section or any section of this application shall not be construed as an admission that such reference is available as prior art to the present invention.

In the constantly changing global environment, the properties of various microorganisms in natural and other bodies of water have also changed. "Blooms" of microorganisms have occurred often making them harmful to other organisms of the surrounding environment and, to terrestrial animals, including humans. See, Anderson et al. 1993. Woods Hole Ocean. Instit., Technical Report 93-02. Within this category of microorganisms lies an important group of planktonic microalgae whose harm mainly results from the toxins they produce, though some examples exist of species that have detrimental effects to wildlife due simply to their abundance and physiology. Examples of microorganisms that fall into this category include: Pseudo-nitzschia, which is responsible for the death of sea- lions due to neurological dysfunction caused by do oic acid ("DA") (Scholin et al. 2000, Nature 403:80); Microcystis, which is responsible for the killing of farm animals caused by microsystin toxins (Chengappa et al. 1989, J.A. KM. A. 194:1724); Gymnodinium, which is responsible for shellfish poisoning (Nielsen, M.V. 1993, Mar. Ecol. Prog. Ser. 95:273): Pfiesteria, which is responsible for fish death and neurological dysfunction in humans (Steidingeretal. 1996, J. Phycol. 32: 157); mάAlexandrium, which is responsible for fish death caused by Paralytic Shellfish Poisoning ("PSP") toxins (Robineau et al. 1991, Marine Bio 108:293).

In New Zealand, both phytoplankton monitoring and shellfish flesh testing programs have led to an extensive database which has helped link species of Pseudo-nitzschia to specific DA outbreaks. P. pungens and P. turgidula have been associated with DA contamination of shellfish, and cultured isolates of these species proved to be toxin producers. The use of species-specific rRNA-targeted oligonucleotide probes and DA immunoassays led to the discovery of toxin production by P. fraudulenta, and showed the nontoxic P. heimii to be a major bloom former. Pseudo-nitzschia delicatissima, P. pseudodelicatissima and P. multiseries, also identified using rRNA-targeted probes, have been linked to DA contamination of New Zealand shellfish, while P. australis is the main cause of DA in scallops. Rhodes et al. 1998, Natural Toxins 6:105.

In a survey on Australia, using rRNA fluorescent probes the presence of Pseudo- nitzschia australis was identified for the first time. Other species also seen were P. multiseries, P. pseudodelicatissima, P. pungens, P. delicatissima, P. fraudulenta, P. subfraudulenta, P. subpacifica, P. heimii and P. cuspidata. Preliminary analysis using Enzyme-Linked Immunosorbent Assay has confirmed significant levels of domoic acid in extracts from P. australis. Of the five P. pseudodelicatissima cultures tested, one produced detectable levels of domoic acid. The regular presence in Southern Australian waters of low concentrations Pseudo-nitzschia and domoic acid highlight the need for monitoring programs to be set up in marine farm areas. Lapworth etal.2000, Ninth Conference on Harful Algal Blooms, Tasmania. For all these and other similar microorganisms that are increasing in abundance worldwide, proper detection allows for measures to avoid harmful effects that arise from "bloom" formation and/or toxicity. The importance of detection of such organisms has significance in three major overlapping sectors: (1) economic effects on industries such as aquaculture, tourism, and the seafood industry; (2) human and animal health resulting from the consumption of seafood; and (3) environmental / ecological effects.

Because all these sectors interact in direct and indirect ways, the significance in detection and appraisal of possible danger or harm is large. Demand spans from the public to the private sector as environmental protection, human safety, and quality control activities issues are advocated.

In the United States alone, hundreds of millions of dollars have been invested by public and private institutions in the past decade in an attempt to overcome the problems arising from the existence of harmful microorganisms. At the moment, economic assessment of the effects of harmful algae is actively increasing since the effects of alterations in the environment due to harmful microorganisms are no longer negligible. Though investment was not large from 1980 to 1992, the over 100 million dollars per year in damages created by harmful algae within the last 8 years, forced many agencies and companies to take matters more seriously (Hoagland et al. 1999. ECOHAB Research Projects). The task of distinguishing, identifying and counting the harmful species within the plethora of other microalgae, detritus and near-identical non-harmful species in atimely fashion is difficult using methods developed to date. The task is made even more difficult when having to distinguish between toxic and non-toxic species, or when monitoring the time course of toxin production is necessary or desired.

There is a need for development of fast, economically feasible, and automated techniques for identification or identification and quantification of microorganisms in environmental samples.

3. SUMMARY OF THE INVENTION

In its most general aspect, the present invention provides a computer-implemented image analysis method of detecting at least one signal, which provides information regarding the presence or quantity of a microorganism in an environmental sample.

In its most general embodiment, the method of the present invention includes acquiring image data of an environmental sample, in a preferred embodiment, from a body of water, processing the image data to select and record images of a detectable signal indicative of microorganism(s) or a substance or compound, such as a toxic substance or compound, produced by the microorganism. Counts may be maintained of the number and/or strength of the detectable signal identified. The microorganism or substance or compound e.g. a toxic compound, or both may be detected, in certain embodiments, using more than one detectable signal.

In the method, the image data is transformed from one color space, e.g., an RGB (Red Green Blue) color space into a different color space, e.g., an HLS (Hue Luminescence Saturation) color space. Filters and/or masks are utilized to distinguish those images that meet pre-selected criteria, i. e. contain a detectable signal(s), and eliminate those that do not, and thus identify or identify and quantify a microorganism or a toxic substance or compound produced by the microorganism.

According to the methods of the invention, a monolayer of microorganisms, fixed to a suitable solid substrate is observed by a computer controlled microscope system. The monolayer can be obtained merely by spreading a sample, e.g. , of a body of water suspected of containing a microorganism of interest, on a solid substrate, such as a slide. Alternatively, a monolayer can be obtained by spreading a sample suspected of containing an enriched population of microorganisms of interest on a solid substrate.

According to one embodiment of the method of the invention, computer controlled image analysis is conducted on an environmental sample, preferably a sample of a body of water suspected of containing microorganisms of interest, treated to provide at least two different signals which can both be detected and quantified. At least two signals are required for this embodiment. Of course, more than two signals may be employed. A first signal is employed to identify a microorganism of interest ("identification signal") and a second signal is employed to identify a characteristic substance or compound within an identified microorganism of interest ("characterization signal"). As would be understood by those skilled in the art, the designation "first" and "second" signals is not meant to mean one signal must be detected prior in time to the other. Rather, the terms indicate that two different signals are detected and analyzed. Thus, a characterization signal may be processed before an identification signal and vice versa. When the characteristic compound is a toxic compound, detection and quantification of the two different signals provides for determination of: (1) the number of microorganism cells, e.g., per unit volume of an environmental sample, (2) the number of microorganisms producing the characteristic substance or compound, or (3) both the number of microorganism cells and the amount of toxic substance or compound present i.e., the extent of the presence of toxic microorganisms. When the characteristic compound is indicative of a particular genus or particular species of microorganism, detection and quantification of the two different signals provides for determination of the extent of the presence of a particular genus or particular species of microorganism. In a specific embodiment, the characterization signal is indicative of a toxic substance or compound produced by the microorganism. In another specific embodiment, the characterization signal, in conjunction with the identification signal is indicative of a particular genus or a particular species of microorganism. In another specific embodiment, the characterization symbol is indicative of a storage compound, such as a lipid or fatty acid stored in a storage vacuole of the microorganism. A physical feature of the microorganism cells can be used to provide a first signal, or more preferably the microorganism cells are treated, e.g. stained, to produce a first signal. The fixed microorganism cells are also treated to produce a second signal specific to a substance or compound characteristic of the microorganism or a particular stage of the life cycle of the microorganism. As used herein, "treatment" of a sample to provide or generate a desired (identification or characterization) signal may be accomplished simply by fixing a sample to a substrate. A signal may be generated by a substance present in the microorganism and detected e.g., by detecting the autofluorescence of a substance, e.g., chlorophyl, present in the microorganism. A signal may be generated by a physical feature of the microorganism, such as size, shape and detected, e.g., by a size/shape recognition system. Alternatively, treatment of sample to produce or generate a desired (identification or characterization) signal may be accomplished by contacting the sample with an agent, e.g., a stain, a labelled ligand, etc. which produces a signal.

In the illustrative description below and illustrated e.g. , in Figure 4, merely for ease of

5 description and not by way of limitation, a first signal, i.e. identification signal is detected which fluoresces in the "green" wavelength range and a second signal, i.e., characterization signal, fluoresces in the "red" wavelength range. As would be understood by those skilled in the art, identification of signal(s) fluorescing at different wavelengths can be detected by adjusting the cut off parameters of the illustrative algorithm as necessary. For example, and

10 not by way of limitation, if a fluorophore such as aminomethylcoumarin (AMC) fluorescing in the "blue" wavelength is used, the grey level value(s) can be fixed at 155-180; if a fluorophore, such as fluoresceinisothiocyanate (FITC) fluorescing in the "green" wavelength is used, the grey level value(s) can be fixed at 40-70; if a fluorophore such as rhodamine

(PRITC) fluorescing in the "red" wavelength is used, the grey level value(s) can be fixed at 0-

15 15 and 245-255, etc.

The computer controlled image analysis of this embodiment of the invention is accomplished using a computer software product including a computer-readable storage medium having fixed therein a sequence of instructions which, when executed by a computer directs the performance of method steps comprising:

20 • A microscope image of an optical field of a substrate having fixed thereon a monolayer of cells treated to produce a first signal specific to a microorganism or group (e.g., a genus) of microorganisms of interest and a second signal specific to a particular species of microorganism of interest or a substance or compound produced by a microorganism of interest is acquired and transferred to the computer as a color space or image, e.g.,

25 an RGB (red, green blue) image. RGB is a color system that is used more frequently in computer graphics. Other color systems that can be used in the present invention include: CMY (cyan, magenta, yellow); HSV (hue, saturation, value), etc. Rogers,

1985, Procedural elements for computer graphics, McGraw Hill Int'l Editions, p. 433.

• The image is transformed to the HLS color space.

30 • The Luminance component of the HLS image is transferred to a new monochrome grey- level image and clipped for pixel values of less than 20 to cut down background pixels.

• The grey-level image is then transformed to a "binary" image: this is a black and white image in which pixels with corresponding pixels in the Luminance image having grey- level values lower than the cut off point are set to 255 (white). An "opening" filter is applied: Opening is a successive application of an "erosion" filter followed by a "dilation" filter. This allows for the removal of small noise particles from the binary image.

• A processing function is applied that removes blobs or erases border blobs that touch the end of the image.

• The area of each microorganism is measured and all microorganisms that have an area of more than the appropriate cell mask of interest, e.g. , 5000 pixels, are excluded from further processing. The pixel number, or area of the remaining candidate cells and location are recorded and saved for further processing. • The Hue component from the original RGB image is transferred to a new monochrome image and binarized, so that pixels having grey-level values at a set level, e.g., 255, while all pixels are set to 0. Pixels with such Hue values represent areas of the image providing a second, i.e. characterization signal.

• A custom processing function is applied that removes candidate blobs that extend beyond the field of view.

• Another custom processing function is applied that "fills holes" in the remaining candidate blobs.

• The candidate blobs are confirmed to lie within the candidate cells to determine that a microorganism of interest is present. In addition, if quantification of the microorganism or substance or compound of interest is desired, the intensity of the identification or characterization signal or both signal can be determined and compared to a standard curve as appropriate.

The methods and apparatus of the present invention are used to detect or detect and quantify the presence, in an environmental sample, of any of a number of microorganisms, including but not limited to the following:

Pseudo-nitzschia species, Alexandrium species, Anabaena species, Chrysochromulina species, Dinophysis species, Gyrodinium species, Gymnodinium species, Microcystis species, Nodulania species, Pfiesteria and Pfiesteria-like species, Prymnesium species, Prorocentrum species, Gambierdiscus species, Ostreopsis species, Coolia species, Thecamdinium species, Amphidinium species, Pyrodinium species, Cynlindrospermosis species, Heterosigma species, Gyrodinium species, Chaetoceros species, etc.

One particular embodiment of the present invention is a computer-implemented method for detecting or detecting and quantifying a Pseudo-nitzschia species, e.g., P. australis in an environmental sample. In particular, made of this embodiment, the presence of a toxic compound, e.g., domoic acid, is detected or detected and quantified. Another particular embodiment of the present invention is a computer-implemented method for detecting or detecting and quantifying Alexandrium species, e.g., A. exeavatum, A. minutum, in an environmental sample.

Another particular embodiment of the present invention is a computer-implemented method for detecting or detecting and quantifying a Pfiesteria species., e.g., P. piscicida in an environmental sample.

Another particular embodiment of the present invention is a computer-implemented method for detecting or detecting and quantifying a Gymodinium species, e.g., G. breve, G. nagasakieuse, G. galatheanum, Another particular embodiment of the present invention is a computer-implemented method for detecting or detecting and quantifying Dinophysis species, e.g., D. acuminata, D. acuta, D. caudata, andD. rotuda, in an environmental sample.

The present invention also encompasses a computer software product including a computer-readable storage medium having fixed therein a sequence of instructions which, when executed by a computer, direct the performance of steps for conducting the methods of the invention as described herein.

The present invention also encompasses a method of operating a laboratory service for screening environmental samples to detect or detect and quantify microorganisms in environmental samples for providing useful information for assessing whether a treatment of a portion of the environment, e.g. , a body of water, is effective to control or inhibit "bloom" of a microorganism. The method encompasses the steps of receiving a prepared substrate, e.g. a slide, that has a monolayer of microorganism cells from an environmental sample placed on the substrate, treating the sample to generate at least one signal, obtaining an image of the monolayer of cells, and operating a computerized microscope according to the method(s) described herein to provide useful information regarding identity and/or quantity of a microorganism or a compound or substance produced by the microorganism.

The present invention also provides a system for screening environmental samples for the presence of a microorganism or a toxic substance or compound produced by a microorganism. The basic elements of the system include an X- Y stage, a mercury light source, a fluorescence or bright field microscope, a digital camera system such as a color CCD camera or a complementary metal-oxide semiconductor (CMOS) image system, a personal computer (PC) system, and one or two monitors and most importantly a computer software product including a computer-readable storage medium having fixed therein a sequence of instructions which, when executed by a computer, direct the performance of steps for conducting the methods of the invention as described herein. 3.1. OBJECTS AND ADVANTAGES OFTHE INVENTION

It is an object of the invention to provide methods and systems which provide identifying or identifying and quantitative information regarding microorganisms present in an environmental sample. The present methods and systems advantageously provide identifying or identifying and quantitative information with respect to microorganisms which may be present in relatively low concentration. It is an object of the invention to establish a fast approach for the identification and quantification of both a microorganism and a toxic substance or compound produced by the microorganism. This provides an important step forward in the development of a technology for the detection of microorganisms and/or their associated compounds.

It is another object of the invention to provide automated image analysis methods and systems to identify or identify and quantitate the presence of a toxic substance or compound produced by a microorganism. In certain embodiments, not only may the toxic substance or compound be identified but also the microorganism which produces it. This two step embodiment of the process is advantageous because the associated substance or compound may be present only during certain stages of the microorganism's life cycle. Such is the case with some harmful microalgae.

In certain embodiments, the quantitative information provided by the methods of the invention is used to assess efficacy of a treatment which is employed to control the microorganism in a portion of the environment, e.g., body of water.

Unlike conventional methods of identification which require either concentration of microorganisms in an environmental sample, e.g., by condensing a sample from a volume of many liters to a single drop on a microscope slide or by culturing a sample to increase microorganisms present to a detectable level, in certain advantageous embodiments, the present invention provides methods of identifying or identifying and quantifying microorganisms in an environmental sample without concentration. Of course, as would be understood by those skilled in the art, concentration of microorganisms, e.g. by filtration of a sample using a sieve or filter, can be used with the methods and systems of the present invention. The methods and systems of the invention provide the ability to detect or detect and quantify large numbers of different species of microorganisms as well as particular species which may be present in low number(s) in environmental samples. 4. BRIEF DESCRIPTION OF THE DRAWINGS

The present invention may be understood more fully by reference to the following detailed description of the invention, illustrative examples of specific embodiments of the invention and the appended figures in which:

Figure 1 is schematic representation of an apparatus or system useful for the methods of the invention. The automated slide feeder (109) depicted in Figure 1 is an optional component of a preferred system useful for the methods of the invention.

Figure 2 is a flow chart summarizing the method of one aspect of the invention in which a microorganism is detected or detected and quantified in an environmental sample. As shown in Figure 2, image analysis is used to detect a microorganism of interest based upon the detection of at least two signals, i.e., an identification signal which permits the selection of a "candidate cell", and a characterization signal which permits the selection of a "candidate blob". A micro-organism is confirmed detected upon determination that a candidate blob lies within a candidate cell.

Figure 3 is a flow chart summarizing the method of one embodiment of the aspect of the method illustrated in Figure 2.

Figure 4 is a flow chart summarizing the method of another aspect of the invention in which a microorganism, exemplified by Alexandrium species, is detected or detected and quantified in an environmental sample.

5. DETAILED DESCRIPTION OF THE INVENTION

The invention will be better understood upon reading the following detailed description of the invention and of various exemplary embodiments of the invention, in connection with the accompanying drawings. It will be clear to those skilled in the art that the invention encompasses methods and systems using automated image analysis for screening samples for the presence of a microorganism or a substance or compound, e.g. , a toxic substance or compound, produced by the microorganism. The methods are based upon detection of any detectable characteristic of the microorganism or substance or compound. The invention also encompasses methods and systems using automated image analysis for screening environmental samples for the presence of a microorganism, which may be an indicator of the environment's present condition or may affect the environment. If a substance or compound, such as a toxic substance or compound, is associated with the microorganism, the substance or compound may be identified and quantified. For the purpose of this invention, an environmental sample is obtained from any naturally occurring or man-made body of water or soil (whether in its "natural state" or under cultivation). As used herein, a "body of water" includes both moving and more sedentary waters, including streams, rivers, ponds, lakes, estuaries, bays, sounds, oceans, etc. When an environmental sample is a sample of a soil to be screened for a microorganism of interest or a substance or compound produced by a microorganism of interest, water or an aqueous solution is added to the sample to extract any microorganism(s) or substance(s), forming a mixture, energy may be added to the mixture to form a slurry, the mixture is filtered to remove soil particles and other debris and the filtrate is used according to the methods described herein. 0 According to the present invention, an environmental sample is placed on a solid substrate, such as a glass slide or a filter, etc., so that any microorganism(s) of interest present therein can be observed as a "monolayer" of cells. As used herein, a monolayer simply means that the cells are arranged whereby they are not viewed on top of one another. It does not require confluence. It simply requires a layer of cells that has a thickness of a single cell. In 5 certain embodiments, an environmental sample is passed through a filter or sieve and any microorganism(s) present therein can be observed on the filter or sieve as mono-layer.

5.1. METHODS AND SYSTEMS FOR DETECTION OR DETECTION AND QUANTIFICATION _Q The present invention provides methods, apparatus and systems for the rapid detection and, more preferably, detection and quantification of a microorganism in an environmental sample. The methods employ apparatus and systems which afford computer controlled automated image analysis.

Figure 1 illustrates the basic elements of a system suitable for use according to the 5 methods of the invention. A system such as illustrated in Figure 1 can be used in any of the methods described in the sub-sections below. The basic elements of the system include anX-Y stage (101), a mercury light source (102), a epi fluorescence and/or bright field microscope (103), a color CCD camera or a CMOS image sensor (105), a personal computer (PC) system (106), and one or two monitors (107 and 108). Figure 1 also illustrates an optional component Q (109), an automated slide feeder used in a preferred system of the present invention. In a preferred embodiment, the automated slide feeder is configured with the stage so that slides can be automatically moved in and out of position for image capture and/or analysis.

The individual elements of the system can be custom built or purchased off-the-shelf as standard components. Each element is described in somewhat greater detail below. The X- Y stage can be any motorized positional stage suitable for use with the selected microscope. Preferably, the X-Y stage can be a motorized stage that can be connected to a personal computer and electronically controlled using specifically compiled software commands. When using such an electronically controlled X-Y stage, a stage controller circuit card plugged into an expansion bus of the PC connects the stage to the PC. The stage should also be capable of being driven manually. Electronically controlled stages such as described here are produced by microscope manufacturers, for example including Olympus (Tokyo, Japan), as well as other manufacturers, such as LUDL (NY, USA).

The microscope can be any fluorescence microscope equipped with a reflected light fluorescence illuminator or any bright-field microscope and a motorized objective lens turret with a 20X, 40X, and 100X objective lens, providing a maximum magnification of 1000X. The motorized nosepiece is preferably connected to the PC and electronically switched between successive magnifications using specifically compiled software commands. When using such an electronically controlled motorized nosepiece, a nosepiece controller circuit card plugged into an expansion bus of the PC connects the stage to the PC. The microscope and stage are set up to include a mercury light source, capable of providing consistent and substantially even illumination of the complete optical field.

The microscope produces an image viewed by the camera. The camera can be any color 3 -chip CCD camera or other camera or digital image sensor system connected to provide an electronic output and providing high sensitivity and resolution. The output of the camera is fed to a frame grabber and image processor circuit board installed in the PC. A camera found to be suitable is the OPTRONICS 750 (OPTRONICS, CA.). Another digital camera found to be suitable is the complementary metaloxide semiconductor system (CMOS) image sensor available from Photobit (Pasedena, CA).

Various frame grabber systems can be used in connection with the present invention. The frame grabber can be, for example, the MATROX GENESIS available from MATROX (Montreal, CANADA). The MATROX GENESIS module features on-board hardware supported image processing capabilities. These capabilities compliment the capabilities of the MATROX IMAGING LIBRARY (MIL) software package. Thus, it provides extremely fast execution of the MIL based software algorithms. The MATROX boards support display to a dedicated SVGA monitor. The dedicated monitor is provided in addition to the monitor usually used with the PC system. Any monitor SVGA monitor suitable for use with the MATROX image processing boards can be used. One dedicated monitor usable in connection with the invention is a ViewSonic 4E (Walnut Creek, CA) SVGA monitor.

In order to have sufficient processing and storage capabilities available, the PC can be any PC such as an INTEL PENTIUM-based PC having at least 256 MB RAM and at least 40GB of hard disk drive storage space. The PC preferably further includes a monitor. Other than the specific features described herein, the PC is conventional, and can include keyboard, printer or other desired peripheral devices not shown.

Alternatively, automated sample analysis may be performed by an apparatus and system for distinguishing, in an optical field, objects of interest from other objects and background, suchastheautomatedsystemexemplifiedinU.S. PatentNo.5,352,613, issued October4, 1994 (incorporated herein by reference) . Furthermore, once an obj ect, i.e., a relevant animal cell has been identified, the color, e.g. the combination of the red, green, blue components for the pixels that comprise the obj ect, or other parameters of interest relative to that obj ect, i.e., whether it is infected or not by an infectious agent, can be measured and stored. Other examples of alternative apparatus and systems for automated sample analysis are illustrated in WO99/58972 and PCT/USOO/31494 (incorporated herein by reference).

One suitable system consists of an automatic microscopical sample inspection system having:

• a sample storage module and loading and unloading module • a sample transporting mechanism to and from an automated stage that moves the sample under a microscope objective lenses array

• an array of CCD cameras

• a processing unit having a host computer, multiple controllers to control all mechanical parts of the microscopy system and • a high speed image processing unit where the CCD cameras are connected (see,

PCT/USOO/31494).

An innovative feature of this embodiment of a computer controlled system is an array of two or more objective lenses having the same optical characteristics. The lenses are arranged in a row and each of them has its own z-axis movement mechanism, so that they can be individually focused. This system can be equipped with a suitable mechanism so that the multiple obj ective holder can be exchanged to suit the same variety of magnification needs that a common single-lens microscope can cover. Usually the magnification range of light microscope objectives extends from IX to 100X .

Each objective is connected to its own CCD camera. The camera field of view characteristics are such that it acquires the full area of the optical field as provided by the lens.

Each camera is connected to an image acquisition device. This is installed in a host computer. For each optical field acquired, the computer is recording its physical location on the microscopical sample. This is achieved through the use of a computer controlled x-y mechanical stage. The image provided by the camera is digitized and stored in the host computer memory. With the current system, each objective lens can simultaneously provide an image to the computer, each of which comprises a certain portion of the sample area. The lenses should be appropriately corrected for chromatic aberrations so that the image has stable qualitative characteristics all along its area.

The imaged areas will be in varying physical distance from each other. This distance is a function of the distance at which the lenses are arranged and depends on the physical dimensions of the lenses. It will also depend on the lenses' characteristics, namely numerical aperture and magnification specifications, which affect the area of the optical field that can be acquired. Therefore, for lenses of varying magnification /numerical aperture, the physical location of the acquired image will also vary. The computer will keep track of the features of the obj ectives-array in use as well as the position of the motorized stage. The stored characteristics of each image can be used in fitting the image in its correct position in a virtual patchwork, e.g. "composed" image, in the computer memory.

The host computer system that is controlling the above configuration, is driven by software system that controls all mechanical components of the system through suitable device drivers. The software also comprises properly designed image composition algorithms that compose the digitized image in the computer memory and supply the composed image for processing to further algorithms. Through image decomposition, synthesis and image processing specific features particular to the specific sample are detected.

5.1.1. METHODS USING AT LEAST TWO SIGNALS According to one embodiment of the methods of the invention, a microorganism is detected, or preferably detected and quantified, in an environmental sample treated to provide at least two different signals which can both be detected and quantified. In this embodiment, at least two signals are required. As would be understood by those skilled in the art, if desired, more than two signals can be employed. A first signal is employed to identify a microorganism or a group (such as a genus, etc.) of microorganisms of interest ("identification signal") and a second signal, different from the first, is employed to identify a characteristic substance or compound within the microorganism of interest ("characterization signal"). If more than two signals are employed, one or more signals are employed as a identification signal and one or more signals are employed as a characterization signal.

In certain modes of this embodiment, the identification signal indicates that a member of a group of microorganisms is present and the characterization signal indicates that the microorganism belongs to a particular genus of microorganisms. In certain other modes of this embodiment, the identification signal indicates that a microorganism belongs to a particular genus of organisms and the characterization signal indicates the presence of a compound indicative of a particular species of that genus. In other words, the combination of identification and characterization signals indicates the presence of microorganisms of a particular genus or species. In certain other modes of this embodiment, the identification signal indicates the presence of a particular microorganism and the characterization signal indicates the presence of a substance or compound produced by the microorganism, e.g., a substance or compound associated with a particular stage in the life cycle of the microorganism. In a specific example of this mode, the substance or compound is atoxic substance or compound, h another specific example, the substance or compound is a storage substance indicative, e.g., of a specific microorganism or a stage of the life cycle of microorganism. In a particular illustrative example, the storage substance is a lipid containing storage substance such as apolyunsaturated fatty acid. Such substance maybe identified using image analysis, e.g. , by the distinctive shape of a storage vacuole containing such a substance. Evaluation of the strength of one or both of the signals is used to quantify either the concentration of microorganisms or substance or compound or both in the environmental sample in any of the modes of this embodiment.

As used herein, "signal" should be taken in its broadest sense, as a physical manifestation which can be detected and identified, thus carrying information. One simple and useful signal is the light emitted by a fluorescent dye selectively bound to a structure of interest. That signal indicates the presence of a structure, e.g., indicative of a microorganism which might be difficult to detect absent the fluorescent dye.

According to this embodiment of the method, detection is based on both an identification signal and a characterization signal. The approach requires the ability to generate a specific identification signal and a characterization signal for the desired target microorganism or substance or compound. The identification signal may be generated by any of several methods. For example, recognition of the characteristic morphology of a particular organism or a particular portion of an organism, e.g. , a chloroplast, vacuole, etc. by cell (size and/or shape) recognition algorithms, immunostaining of the organism with an antibody specific for a cell specific antigen or recognition of organism-specific nucleic acid sequences. The signal produced by binding of the specific antibody could be generated by the primary antibody itself or by a subsequent binding of a secondary antibody and could be fluorescent or colorimetric in nature. The recognition of organism-specific nucleic acid sequences may be by in situ amplification techniques such as direct or indirect in situ Polymerase Chain Reaction ("PCR") amplification or rolling circle amplification, which may also be designed to produce fluorescent or colorimetric signals. Rolling Circle Amplification (RCA) technology, as described by Lizardi etal, 1998,Nature Genetics 19:225-232, entitled "Mutation Detection and Single-Molecule Counting Using Isothermal Rolling-Circle Amplification", the entire disclosure of which is incorporated by reference in its entirety, see, id., Figs. 1, 4, and 6, incorporated specifically herein by reference; see also, Nilsson et al., 1994, Science 265 :2085- 2088; Nilsson et al., 1997, Nature Genet. 16:252-255; Fire et al., 1995, Proc. Nat'l Acad Sci. USA, 92:4631-4645; and Liu et al., 1996, J. Am. Chem. Soc. 118:1587-1594, each of which is incorporated herein by reference in its entirety, can be used to generate an identification or characterization signal. Cells that emit both signals, i.e., the cell is a microorganism cell and contains a characteristic substance tested for, will be scored. Counts may be maintained of the number and strengths of the first and second signals detected.

RCA generates multiple copies of a selected region of a target gene as a single stranded DNA. In an exemplary embodiment, detection of the ssDNA produced by RCA comprises in situ probing, with one or multiple probes, partially deproteinized or deproteinized cytological samples suspected of containing a rare cell. The DNA in the sample is denatured prior to RCA. In one mode of the embodiment, detection is achieved by condensation of the ssDNA after hybridization of complementary oligonucleotide tags to the ssDNA (RCA-CACHET format). In yet another embodiment, a hapten such as BUDR is incorporated into the ssDNA. The ssDNA is then contacted with an antibody that recognizes the hapten. The ssDNA-antibody complex can be detected by a label that is conjugated to the antibody, conjugated to a second antibody or, if the antibody is coupled to avidin, by a label that is conjugated to biotin. The detectable label is preferably a fluorophore, but can also be an enzyme that catalyzes a colorimetric reaction such as alkaline phosphatase (AP) or horseradish peroxidase (HRP). It should now be evident that any detectable indicator of the presence of a microorganism may serve as the identification signal, subject to certain constraints noted below.

The characterization signal indicates the presence of aparticular substance or compound being tested for and may be generated by any methods known to those skilled in the art, by immunostaining or colorimetric inhibition assay. The signal produced by binding of the specific antibody could be generated by the primary antibody itself or by a subsequent binding of a secondary antibody and could be fluorescent or colorimetric in nature. It should now be evident that any detectable indicator of the presence of a microorganism may serve as the identification or characterization signal, subject to certain constraints noted below.

In certain automated sample analysis embodiments of the invention, if the generation of the identification signal is measured first, indicating cell identity, the sample will be observed using an automated optical microscope to delineate coordinates of a desired number of identified microorganisms. Only those samples found to contain identification signal need be treated to generate the characterization signal, indicating the presence of the particular compound being assessed. The automated image analysis algorithms will search for the presence of the second signal at predetermined coordinates of microorganisms and also at predetermined coordinates of control microorganism cells. This process may be reversed, whereby the characteristic signal is observed first, and if a cell is emitting that signal then that cell may be treated with the identification signal to determine whether it is the microorganism of interest. According to a preferred embodiment, a sample is treated at one time to generate at least two signals, i.e., identification and characterization signals. It is even possible to observe both signals simultaneously, searching only for the simultaneous presence of two signals at a single set of coordinates or even a single signal which results from the interaction of two components (e.g. a quenching of a first signal by a partner 'signal', the first signal being for the microorganism and the partner 'signal' being for the compound).

The requirements and constraints on the generation of the identification and characterization signals are relatively simple. The materials and techniques used to generate the identification signal should not interfere adversely with the materials and techniques used to generate the characterization signal, to an extent which compromises unacceptably the diagnosis, and visa versa. Nor should they damage or alter the cell characteristics sought to be measured to an extent that compromises unacceptably the diagnosis. Finally, any other desirable or required treatment of the cells should also not interfere with the materials or techniques used to generate the identification and characterization signals to an extent that compromises unacceptably the detection or detection and quantification of microorganisms. Within those limits, any suitable generators of the identification and characterization signals may be used. See, infra Section 5.2 for more detailed specifics regarding particular microorganisms.

Figures 2-4 schematically illustrate certain aspects of various embodiments of the methods of the present invention.

Figure 2 schematically illustrates one embodiment of one method of the invention for detection or detection and quantification of a microorganism in an environmental sample. A pair of lenses, e.g., an objective and ocular lens providing, e.g., 100 to 1000X total magnification optionally 100 to 400X total magnification, optionally 400 to 1000X total magnification, preferably 400X or 1000X total magnification is used and a solid substrate, such as a slide or a filter, etc., containing a sample suspected of containing a microorganism of interest, treated so that at least an identification signal and a characterization signal can be detected or detected and quantified if a microorganism of interest is present in the sample, is analyzed as follows:

• An optical field on the substrate is selected and a microscope image is acquired using a CCD camera or CMOS image system and transferred to a computer as an RGB image. (201)

• The image is transformed to the HLS model. (202)

• The Luminance component of the HLS image is transferred to a new monochrome grey- level image and clipped for pixel values of less than 20 to cut down background pixels. (203) • The grey-level image is then transformed to a "binary" image: this is a black and white image in which pixels with corresponding pixels in the Luminance image having grey- level values lower than the cut off point are set to 255 (white). (204)

• An "opening" filter is applied: Opening is a successive application of an "erosion" filter followed by a "dilation" filter. This allows for the removal of small noise particles from the binary image. (205)

• A processing function is applied that removes blobs, designated "cells" or erases border cells that touch the end of the image. (206)

• The area of each microorganism is measured and all microorganisms that have an area of more than the appropriate cell mask of interest, e.g. , 5000 pixels, are excluded from further processing. The pixel number, or area of the remaining candidate cells and location are recorded and saved for further processing. (207)

• The Hue component from the original RGB image is transferred to a new monochrome image and binarized, so that pixels having grey-level values at a set level, e.g., 255, while all pixels are set to 0. Pixels with such Hue values represent areas of the image providing a second, i.e. characterization signal. (208)

• A custom processing function is applied that removes border blobs that extend beyond the field of view. (209)

• Another custom processing function is applied that "fills holes" in the remaining candidate blobs. (210) • Noise particles are removed. (211)

• Candidate blobs are selected based on detection of an appropriate signal. (212)

• The candidate blob is confirmed to lie within an appropriate candidate cell and hence indicates that a microorganism of interest is detected. (213)

According to the above described embodiment of the method of the invention, the candidate cell selected in steps 203 -207 is based on detection of an identification signal relevant to the microorganism of interest and the blob selected in step 212 is based on detection of a characterization signal relevant to the microorganism of interest.

The method illustrated in Figure 2 can be used to detect or detect and quantify any of the microorganisms including but not limited to Pseudo-nitzschia species; Alexandrium species, Anabaena species, Chrysochromulina species, Dinophysis species, Microcystis species, Nodulania species, Pfiesteria and Pfiesteria-like species, Prymnesium species, Prorocentrum species, Gambierdiscus species, Ostreopsis species, Coolia species, Thecadinium species, Amphidinium species, Pyrodinium species, Cynlindrospermosis species, Heterosigma species, Gymnodinium species, Gyrodinium species, Chaetoceros species, etc. While the description below explains the method of a preferred mode of the invention with respect to a representative chlorophyl-containing microorganism, e.g. , Alexandrium, it is understood that the description is merely so explained and illustrated in Figure 3, for ease of explanation. It will be understood by those skilled in the art that this mode of the invention can be applied to and, in fact, encompasses computer controlled automated image analysis of any type of environmental sample suspected to contain any type of microorganism which, like Alexandrium, contains chlorophyll which auto-fluoresces in the visible wavelength range and which comprises chlorophyl containing body encompassing an area or area and shape of the cells of the microorganisms which can be determined by those skilled in the art. Such organisms include but are not limited to: Alexandrium species, Chrysochromulina species, Dinophysis species, Microcystis species, Prymnesium species, Pseudo-nitzschia species, Prorocentrum species, Ostreopsis species, Amphidinium species, Prodinium species, etc. For certain microorganisms, e.g., Microcystis, etc. which are contained within an amorphous mucous substance secreted by the cells, such mucous sheath must be first disrupted prior to contact of the cells with an agent specific for the cell surface in order to generate an identification or characterization signal.

This mode of one embodiment of the method of the invention can also be modified, e.g. , by varying the criteria for cell shape and size of the chlorophyl body, to detect or detect and quantify microorganisms such as Pseudo-nitzschia species, etc., which contain chlorophyl body or bodies that occupy a size and/or area of the microorganism cells that can be determined by those skilled in the art.

Figure 3 schematically illustrates one embodiment of one method of the invention for detection of a chlorophyl containing microorganism.

An objective lens which provides, e.g., 100 to 1000X, preferably 100 to 400X total magnification, is used and a slide containing a sample suspected of containing Alexandrium, a chlorophyl-containing microorganism of which, e.g. , treated with an antibody specific to the cell surface of the microorganism conjugated to a fluorophore which appears "green" on an automated stage is analyzed. The chlorophyl contained in the organism will fluoresce in the red range and generates a characterization signal. Thus, a microorganism of interest present will be immunostained with a cell surface antibody that produces an image where the cell perimeter is bright green and the internal, chlorophyll containing part of the cell, is autofluorescing in a red. The detection of a microorganism of interest entails:

• An optical field is selected and a microscopic image is acquired and transferred to the computer as an RGB image. (301)

• The image is transformed to the HLS model. (302) • The Luminance component of the HLS image is transferred to a new monochrome grey- level image. (303)

• The grey-level image is then transformed to a "binary" image. This is a black and white image in which pixels with corresponding pixels in the Luminance image having grey- level values lower than twenty represent background pixels (intercellular image area) and are set to 0 (black). All other pixels with corresponding pixels in luminance image having grey level values equal to or higher than 21 represent cells and are set to 255 (white). (304)

• An "opening" filter is applied: Opening is a successive application of an "erosion" filter followed by a "dilation" filter. This allows for the removal of small noise particles from the binary image. (305)

• A processing function is applied that removes candidate cells that extend beyond the field of view. The remaining candidate cells are microorganisms. (306)

• The area of each microorganism is measured and all microorganisms that have an area of more than the cell mask are excluded from further processing. In an example, the cell mask is set at 500 pixels. These represent microorganisms which are not of interest. The pixel number, or area of the remaining candidate cells, is recorded and saved for further processing. (307)

• The Hue component from the original RGB image is transferred to a new monochrome image and binarized, so that pixels having grey-level values between 40 and 70 are set to 255, while all pixels are set to 0. Pixels with such Hue values represent areas of the image depicting membrane specific staining. The area, in pixels, of the cell images, the monocytes, are recorded and saved for further processing. (308)

• A "closing" filter is applied: Closing is a successive application of a "dilation" filter followed by an "erosion" filter. This allows for the connection of small gaps between separated membrane depicting pixel groups in the binary image. (309) • Candidate blobs are selected with a set shape depending upon the shape and size of the microorganism and the chlorophyl containing structure (body) in such microorganism. (310)

• The candidate bodies are verified to lie within the limits of the interesting cell mask that were identified to be less than, e.g. , 5000 pixels. (311)

• The Hue component from the transformed HLS image is transferred to a new monochrome image and binarized, so that pixels having grey-level values between 0 and 20 are set to 255, while all the rest are set to 0. Pixels with such Hue values represent areas of the image depicting red chlorophyll autofluorescence. (312) • A custom processing function is applied that removes border blobs that extend beyond the field of view. (313)

• Another custom processing function is applied that "fills holes" in the remaining candidate bodies. (314)

• An opening filter is applied to remove small noise particles from the binary image. (315)

• Then the size of the remaining candidate bodies is measured to ascertain that it corresponds to the size expected of the candidate cells identified above. (316) Figure 4 schematically illustrates one embodiment of one method of the invention using

Alexandrium as the microorganism, merely for ease of explanation. An objective lens producing a total, e.g. , 400X total, magnification is used and a slide containing a sample suspected of con aining Alexandrium, treated with an antibody-conjugated to a fluorophore which appears "green" on an automated stage is analyzed. Thus, an Alexandrium present will be immunostained with a cell surface antibody that produces an image where the cell perimeter is bright green and the internal, chlorophyll containing part of the cell, is autofluorescing in a red. The detection of Alexandrium entails:

• An optical field is selected and the microscopic image is acquired and transferred to the computer as an RGB image. (401)

The image is transformed to the HLS model. (402)

• The Luminance component of the HLS image is transferred to a new monochrome grey- level image and clipped for pixel values of less than 20 to cut down background pixels.

(403)

• The grey-level image is then transformed to a "binary" image: this is a black and white image in which pixels with corresponding pixels in the Luminance image having grey- level values lower than the cut off point are set to 255 (white). (404) An "opening" filter is applied: Opening is a successive application of an "erosion" filter followed by a "dilation" filter. This allows for the removal of small noise particles from the binary image. (405)

• A processing function is applied that removes candidate cells that extend beyond the field of view. The remaining candidate cells are microorganisms. (406)

• The area of each microorganism is measured and all microorganisms that have an area of more than the cell mask, 5000 pixels, are excluded from further processing. These represent non-Alexandrium microorganisms. The pixel number, or area of the remaining candidate cells, is recorded and saved for further processing. (407) • The Hue component from the original RGB image is transferred to a new monochrome image and binarized, so that pixels having grey-level values between 40 and 70 are set to 255, while all pixels are set to 0. Pixels with such Hue values represent areas of the image depicting membrane specific staining. The area, in pixels, of the cell images, the monocytes, are recorded and saved for further processing. (408) • A "closing" filter is applied: Closing is a successive application of a "dilation" filter followed by an "erosion" filter. This allows for the connection of small gaps between separated membrane depicting pixel groups in the binary image. (409)

• Candidate blobs are selected with a "closed doughnut shape." (410)

• The candidate blobs are confirmed to lie within a candidate cell. (411) • The Hue component from the transformed HLS image is transferred to a new monochrome image and binarized, so that pixels having grey-level values between 0 and 20 are set to 255, while all the rest are set to 0. Pixels with such Hue values represent areas of the image depicting red chlorophyll autofluorescence. (412)

• A custom processing function is applied that removes candidate bodies that extend beyond the field of view. (413)

Another custom processing function is applied that "fills holes" in the remaining candidate bodies. (414)

• An opening filter is applied to remove small noise particles from the binary image. (415) • The size of the remaining candidate bodies is measured to ascertain that it is more than

50% of the size of the candidate cells identified above. (416)

• If the candidate blob lies within the candidate cell and the size of the chlorophyl body is at least 50% (or greater) of the area of the candidate cell, the microorganism detected is determined to an Alexandrium species. (417) As would be understood by those skilled in the art, detection of one or a few microorganisms in a substrate may be sufficient and the process is deemed "done." (418) Alternatively, the process can be repeated until all optical fields on the substrate have been assessed and analyzed and the process is deemed "done." (418)

5.2. MICROORGANISMS AND APPLICATIONS

A number of nucleic acid probes have been developed that have shown utility in the identification of specific microorganisms. These types of probe are short nucleic acid strands, oligonucleotides, designed to bind or hybridize to their complementary target sequence. This is done either through whole cell hybridization ("WCH"), lysed cell hybridization ("LCH") or lysed cell "sandwich" hybridization ("LCSH"). For WCH, the probes are hybridized directly in whole cells (on a slide or in solution) that have been made permeable to the probe. For LCH, cells need to be lysed, their DNA "extracted", and the probe allowed to hybridize. For LCSH, cells also need to be lysed, while a set of probes does the following: hybridization of the capture probe to the complementary rRNA sequences immobilizes target sequences in lysate, and is followed by hybridization of the biotinylated signal probe to the target rRNA, adsorption of avidin-enzyme conjugate to the signal probe and finally visualization of capture probe-target- signal probe sandwiches by an enzyme driven color reaction. The labels with which the probes are tagged are readable by a range of detection methods. The probe can be conjugated to a colorimetric driven reactions and signals & producing the probe can moiety; alternatively, be directly conjugated to a fluorescent reporter molecule such as fluorescein isothiocyanate ("FITC"), or it can be biotinylated and detected using fluorescent avidin conjugates.

Alexandrium cells can be tagged with oligonucleotide probes that identify particular nucleotide sequences from various species of Alexandrium. Such probes are described by Anderson et al. in U.S. Patent No. 5,582,983, which issued on December 10, 1996. Genetic markers for Dinophyceas are described in U.S. Patent No. 5,958,689, which issued on September 28, 1999.

Specific pre-rRNA species have been shown to be useful molecular targets for the detection and identification of specific Pseudo-nitzschia species (Cangelosi et al. 1997, Applied & Environm. Micro. 63:4859; Anderson, D.M. 1995, Rev. Geophys. 33:Suppl.; Scholin et al. 1994, Natural Toxins 2:152). Nucleic acid probes directed to pre-rRNA spacer regions were demonstrated to specifically identify individual Pseudo-nitzschia species including P. australis, P. multiseries and P. pungens. DNA probes based on specific ribosomal RNA sequences have been demonstrated to detect the presence of Pseudo-nitzschia species in marine samples, and their potential for biotoxin monitoring assessed. Rhodes et al. 1998, Natural Toxins 6:105. Any of these nucleic acid probes can be used to generate a signal useful to a detect a microorganism according to the methods of the present invention.

As would be understood by those skilled in the art, for each microorganism or each compound or substance to be detected, specific reagents are required to generate the signals needed for detection. For example, for the embodiment in ^•which.Alexandrium species cells are detected, a labelled antibody directed, e.g. to a cell surface antigen of Alexandrium can be used in generating a first signal. As would be understood by those skilled in the art, antibodies or antibody fragments specific for any microorganism of interest or any compound or substance of interest can be produced by any method known in the art for the synthesis of antibodies (or binding fragments thereof), in particular, by chemical synthesis or preferably, by recombinant expression techniques.

Polyclonal antibodies to a microorganism of interest or a substance or compound of interest can be produced by various procedures well known in the art. For example, a microorganism of interest, an antigen thereof or a compound or substance of interest can be administered to various host animals including, but not limited to, rabbits, mice, rats, etc. to induce the production of sera containing polyclonal antibodies specific for the microorganism, compound or substance or antigen thereof. Various adjuvants may be used to increase the immunological response, depending on the host species, and include but are not limited to, Freund's (complete and incomplete), mineral gels such as aluminum hydroxide, surface active substances such as lysolecithin, pluronic polyols, polyanions, peptides, oil emulsions, keyhole limpet hemocyanins, dinitrophenol, and potentially useful human adjuvants such as BCG (bacille Calmette-Guerin) and corynebacterium parvum. Such adjuvants are also well known in the art. Monoclonal antibodies can be prepared using a wide variety of techniques known in the art including the use of hybridoma, recombinant, and phage display technologies, or a combination thereof. For example, monoclonal antibodies can be produced using hybridoma techniques including those known in the art and taught, for example, in Harlow et al, Antibodies: A Laboratory Manual, (Cold Spring Harbor Laboratory Press, 2nd ed. 1988); Hammerling, et al, in: Monoclonal Antibodies and T-Cell Hybridomas 563-681 (Elsevier, N.Y., 1981) (said references incorporated by reference in their entireties). The term "monoclonal antibody" as used herein is not limited to antibodies produced through hybridoma technology. The term "monoclonal antibody" refers to an antibody that is derived from a single clone, including any eukaryotic, prokaryotic, or phage clone, and not the method by which it is produced.

Methods for producing and screening for specific antibodies using hybridoma technology are routine and well known in the art. Briefly, mice can be immunized with a microorganism of interest or antigenic part thereof, a compound or substance of interest or antigen thereof and once an immune response is detected, e.g., antibodies specific for the microorganism of interest or a compound or substance of interest are detected in the mouse serum, the mouse spleen is harvested and splenocytes isolated. The splenocytes are then fused by well known techniques to any suitable myeloma cells, for example cells from cell line SP20 available from the ATCC. Hybridomas are selected and cloned by limited dilution. The hybridoma clones are then assayed by methods known in the art for cells that secrete antibodies capable of binding an infectious agent or host cell of interest. Ascites fluid, which generally contains high levels of antibodies, can be generated by immunizing mice with positive hybridoma clones.

Antibody fragments which recognize specific epitopes of a microorganism or of a compound or substance of interest may be generated by any technique known to those of skill in the art. For example, Fab and F(ab')2 fragments may be produced by proteolytic cleavage of immunoglobulin molecules, using enzymes such as papain (to produce Fab fragments) or pepsin (to produce F(ab')2 fragments). F(ab')2 fragments contain the variable region, the light chain constant region and the CHI domain of the heavy chain. Further, the antibodies useful to generate a signal (e.g., when labelled) can also be generated using various phage display methods known in the art.

In phage display methods, functional antibody domains are displayed on the surface of phage particles which carry the polynucleotide sequences encoding them. In particular, DNA sequences encoding VH and VL domains are amplified from animal cDNA libraries (e.g., human or murine cDNA libraries of lymphoid tissues). The DNA encoding the VH and VL domains are recombined together with an scFv linker by PCR and cloned into a phagemid vector (e.g., p CANTAB 6 or pComb 3 HSS). The vector is electroporated in E. coli and the E. coli is infected with helper phage. Phage used in these methods are typically filamentous phage including fd and Ml 3 and the VH and VL domains are usually recombinantly fused to either the phage gene III or gene VIII. Phage expressing an antigen binding domain that binds to a microorganism or compound or substance of interest can be selected or identified with a microorganism or compound or substance antigen thereof, e. g. , using labeled antigen or antigen bound or captured to a solid surface or bead. Examples of phage display methods that can be used to make the antibodies useful in the method of the present invention include those disclosed in Brinkman et al., 1995, J. Immunol. Methods 182:41-50; Ames et al., 1995, J. Immunol. Methods 184:177-186; Kettleborough et al, 1994, Eur. J. Immunol. 24:952-958; Persic et al., 1997, Gene 187:9- 18; Burton et al., 1994, Advances in Immunology 57:191-280; PCT applicationNo. PCT/GB91/O1 134; PCT publication Nos. WO 90/02809, WO 91/10737, WO 92/01047, WO 92/18619, WO 93/1 1236, WO 95/15982, WO 95/20401, and WO97/13844; and U.S. Patent Nos. 5,698,426, 5,223,409, 5,403,484, 5,580,717, 5,427,908, 5,750,753, 5,821,047, 5,571,698, 5,427,908, 5,516,637, 5,780,225, 5,658,727, 5,733,743 and 5,969,108; each of which is incorporated herein by reference in its entirety. As described in the above references, after phage selection, the antibody coding regions from the phage can be isolated and used to generate whole antibodies or any other desired antigen binding fragment, and expressed in any desired host, including mammalian cells, insect cells, plant cells, yeast, and bacteria, e.g., as described below. Techniques to recombinantly produce Fab, Fab' and F(ab')2 fragments can also be employed using methods known in the art such as those disclosed in PCT publication No. WO 92/22324; Mullinax et al., 1992, BioTechniques 12(6):864-869; Sawai et al, 1995, AJRI 34:26-34; and Better et al., 1988, Science 240: 1041-1043 (said references incorporated by reference in their entireties).

To generate whole antibodies, PCR primers including VH or VL nucleotide sequences, a restriction site, and a flanking sequence to protect the restriction site can be used to amplify the VH or VL sequences in scFv clones. Utilizing cloning techniques known to those of skill in the art, the PCR amplified VH domains can be cloned into vectors expressing a VH constant region, e.g., the human gamma 4 constant region, and the PCR amplified VL domains can be cloned into vectors expressing a VL constant region, e.g., human kappa or lambda constant regions. Preferably, the vectors for expressing the VH or VL domains comprise an EF-lα promoter, a secretion signal, a cloning site for the variable domain, constant domains, and a selection marker such as neomycin. The VH and VL domains may also cloned into one vector expressing the necessary constant regions. The heavy chain conversion vectors and light chain conversion vectors are then co-transfected into cell lines to generate stable or transient cell lines that express full-length antibodies, e.g. , IgG, using techniques known to those of skill in the art. The signal produced by binding of the specific antibody can be generated by the primary antibody itself or by subsequent binding of a secondary antibody either in both attached to a Cahel which could be fluorescent or colorimetric in nature. Polyclonal antisera has been demonstrated to be especially useful for Pseudonitzschia pungens when distinguishing toxic and non-toxic species. Anderson, 1995, Rev. Geophys. 33:Suppl. Antibodies to intracellular components can also be used for signal generation. In this case the cell must be permeabilized to permit entry of the antibody.

Similar to antibodies, lectin probes are proteins that bind to various cell surface sugars with high specificity for a specific type of sugar molecule. Lectin probes can also incorporate non-radioactive label molecules for detectability, and can be used to help distinguish between toxic and non-toxic species. (Rogers, D.J. and Fish B.C., 1991, Marine algal lectin. frr.Lectin Reviews, Vol. 1 Eds. D.C. Kilpatrick, E. VanDriessche, T. Bog-Hansen, Sigma Library, St.

Louis, USA, pp. 129-142; Rogers, D.J. and Hori K., 1993, Marine algal lectins: New developments, Hydrobiologia, 260:589-593.

As detailed above, the methods and systems of the invention are useful to detect or detect and quantify any of a number of microorganisms, including but not limited to the following:

Pseudo-nitzschia species; Alexandrium species, Anabaena species, Chrysochromulina species, Dinophysis species, Microcystis species, Nodulania species, Pfiesteria and Pfiesteria- like species, Prymnesium species, Prorocentrum species, Gambierdiscus species, Ostreopsis species, Coolia species, Thecadinium species, Amphidinium species, Pyrodinium species,

Cynlindrospermosis species, Heterosigma species, Gymnodinium species, Gyrodinium species,

Chaetoceros species, etc.

One particular embodiment of the present invention is a computer-implemented method for detecting or detecting and quantifying a Pseudo-nitzschia species, e.g., P. australis in an environmental sample. In a specific example of this embodiment the presence of domoic acid is detect or detected and quantified in the Pseudo-nitzschia organisms.

Another particular embodiment of the present invention is a computer-implemented method for detecting or detecting and quantifying Alexandrium species, e.g., A. exeavatum, A. minutum, in an environmental sample. Another particular embodiment of the present invention is a computer-implemented method for detecting or detecting and quantifying a Pfiesteria spp., e.g., P. piscicida in an environmental sample.

Another particular embodiment of the present invention is a computer-implemented method for detecting or detecting and quantifying a Dinophysis species in an environmental sample.

Table 1 below presents a non-limiting illustrative list of specific microorganisms

(and/or substances) which can be detected or detected and quantified according to the present invention together with specific toxic effects as well as the toxin(s) observed in the environment associated with the organisms. TABLE 1

Barth, H. & Nielsen A. (1989), The occurrence of Chrysochromulina polylepis in the Sagerrak and Kattegat in May/June 1988: an analysis of extent, effects and causes. CEE, Water Poll. Res. Rep. 10:v-96p. 2. Belin, C, Berthome, J.P., Lassus P. (1989) Dinoflagelles toxiques et phenomenes d'eaux colorees sur les cotes Francaises. Revue Hydroecol. Equinoxe 25:31-38.

3. Chengappa MM, Pace LW, McLaughlin BG (1989) Blue-green algae (Anabaena spiroides) toxicosis in pigs. J. Am. Vet. Med. Assoc. 15:194.

5 4. Chinain, M. et al., Genetic diversity in French Polynesian strains of the ciguatoxic dinoflagellate, Gambierdiscus toxicus: RFLP and sequence analysis on the SSU and LSL rRNA genes in Harmful and Toxic Algal Bloom, UNESCO Publ. Paris, 1998, pp. 287-290.

5. Guo, Mingxin; Harrison, P.J.; Taylor, F.J.R. (1996) Fish kills related to Prymnesium parvum N. Carter (Haptophyta) in the People's Republic of China J. Appl. Phycol. 8,

10 2:111-117.

6. Kotaki, Y; Koike, K.; Sato, S.; Ogata, T.; Fukuyo, Y; Kodama, M. (1999) Confirmation of domoic acid production of Pseudo-nitzschia multiseries isolated from Ofunato Bay, Japan. Toxicon., 37(4):677-82.

7. Nielsen, M.V. (1993) Toxic effect of the marine dinoflagellate Gymnodinium galatheanum of juvenile cod Gadus morhua, Mar. Ecol. Prog. Ser. 95, 3:273-277.

15

8. Robineau, B.; J.A. Gagne; L. Fortier; and A.D. Cembella, 1991, Potential impact of a toxic dinoflagellate (Alexandrium excavatum) bloom on survival offish and crustacean larvae, Mar. Biol. 108:293-301.

9. Runnegar, M.T.; Andrews, J.; Gerdes, R.G.; Falconer, I.R. (1987) Injury to hepatocytes induced by a peptide toxin from the cyanobacterium Microcystis aeruginosa, Toxicon.

20 25(11):1235.

10 Kaebernick M.; Neilan, B.A.; Borner, T.; Dittmann, E. (2000) Light and the transcriptional response of the microcystin biosynthesis gene cluster.

11. Runnegar, M.T.; Jackson, A.R.; Falconer, I.R. (1988) Toxicity of the cyanobacterium Nodula ia spumigena Meτtens . Toxicon. 26(2): 143-51.

25 12, Scholin, C.A., et al. (2000) Mortality of sea lions along the central California coast linked to a toxic diatom bloom. Nature, Jan. 6, 403 (6765): 80-4.

13. Sournia, A., et al. (1991) Le phytoplankton nuisible des cotes de France, Editions CNRS, pp. 154.

14. Steidenger, K.A.; Burkholder, H.B.; Glasgow, Jr., E.W.; Truby, J.K.; Garrett, E.J.;

30 Nogaand Smith, S.A., (1996) Pfiesteria piscicida, a new toxic dinoflagellate genus and species in the order Dinamoebales, J. Phycology, 32:157-164.

15, Trainer, V.L.; Adams, N.G.; Bill, B.D.; Anulacion, B.F.; Wekell, J.C. (1998) Concentration and dispersal of a Pseudo-nitzschia bloom in Penn Cove, Washington, USA, Nat. Toxins, 6(3-4):l 13-26.

16. Faith, S.A.; Miller, CA. (2000) A newly emerging toxic dinoflagellate, Pfiesteria

35 piscicida: natural ecology and toxicosis to fish and other species, Vet. Hum. Toxicol., Feb. 42(l):26-9. The following non-limiting examples illustrate specific aspects of the present invention.

6. EXAMPLE: DETECTING ALEXANDRIUM This example illustrates the detection of the toxic dinoflagellate Alexandrium

(Dinophiceae) in a water sample. Alexandrium is responsible for paralytic shellfish poisoning (PSP).

6.1. SAMPLE PREPARATION: CELL FIXATION AND CELL STAINING A sample from a natural body of water, e.g. , suspected of containing Alexandrium cells is fixed on a substrate.

The sample is stained with fluorophore-conjugated monoclonal antibody specific for the cell surface of Alexandrium for 15-60 minutes at room temperature in the dark. In this example, anti- Alexandrium monoclonal antibody conjugated fluorescein isothiocyanate (FITC) is used to generate an identification signal. The chlorophyl of the Alexandrium which auto- fluoresces generates a characterization signal. The sample is washed with buffered saline and allowed to air dry.

6.2. AUTOMATED SAMPLE ANALYSIS

Any Alexandrium cells present are immunostained with the cell surface antibody that produces an image where the cell perimeter is bright green (identification signal) and the internal, chlorophyll containing part of the cell, autofluoresces in red providing a second signal (characterization signal).

A PC executes a sample analysis software program compiled in MICROSOFT C++ using the MATROX IMAGING LIBRARY (MIL). MIL is a software library of functions, including those which control the operation of the frame grabber and which process images captured by the frame grabber for subsequent storage in PC as disk files. MIL comprises a number of specialized image processing routines particularly suitable for performing such image processing tasks as filtering, object selection and various measurement functions. The analysis software program runs as a WINDOWS 95 application. The program prompts and measurement results are shown on the computer monitor, while the images acquired through the imaging hardware are displayed on the dedicated imaging monitor. Automated image analysis of the fixed sample is conducted according to the method of the invention as described in Section 5.1.1., supra, and as illustrated in Figure 4.

Cell positions are identified at 400X total magnification. More particularly, the detection of Alexandrium entails (numbers in parentheses indicate step depicted in Figure 4): • An optical field is selected and the microscopic image is acquired and transferred to the computer as an RGB image. (401)

• The image is transformed to the HLS model. (402)

• The Luminance component of the HLS image is transferred to a new monochrome grey- level image and clipped for pixel values of less than 20 to cut down noise or background pixels. (403)

• The grey-level image is then transformed to a "binary" image: this is a black and white image in which pixels with corresponding pixels in the Luminance image having grey- level values lower than the cut off point are set to 255 (white). (404) • An "opening" filter is applied: Opening is a successive application of an "erosion" filter followed by a "dilation" filter. This allows for the removal of small noise particles from the binary image. (405)

• A processing function is applied that removes or "erases border blobs" in the image, i.e., "blobs" that are touching the edge of the image. The remaining images (designated "candidate cells") represent "candidate" microorganism (algae) cells.(406)

• The area of each candidate cell is measured and all candidate cells that have an area of more than the Alexandrium cell mask, 5000 pixels, are excluded from further processing. These removed blobs represent non-Alexandrium microorganisms. The pixel number, or area of the remaining candidate cell images representing candidate Alexandrium cells (identified by the identification signal) is recorded and saved for further processing. (407)

• The Hue component from the original RGB image is transferred to a new monochrome image and binarized, so that pixels having grey-level values between 40 and 70 are set to 255, while all pixels are set to 0. Pixels with such Hue values represent areas of the image depicting membrane specific staining. The area, in pixels, of the cell images is recorded and saved for further processing. (408)

A "closing" filter is applied: Closing is a successive application of a "dilation" filter followed by an "erosion" filter. This allows for the connection of small gaps between separated membrane depicting pixel groups in the binary image. (409) • Candidate blobs are selected with a "closed doughnut" shape, i.e., a small circular hole or dark image (associated with a characterization signal generated by Alexandrium chlorophyl) within a larger, peripheral circular image (associated with an identification signal generated by fluorophore conjugated monoclonal antibody specific Alexandrium cell surface) (410) The candidate blobs are verified to be Alexandrium cells as follows: • The Hue component from the transformed HLS image is transformed to a new monochrome image and binarized, so that pixels having grey-level values between 0 and 20 are set to 255, while all the rest are set to 0. Pixels with such Hue values represent areas of the image depicting red chlorophyll autofluorescence. (412) • A custom MATROX function is applied that removes candidate blobs that extend beyond the field of view. (413)

Another custom MATROX function is applied that "fills holes" in the remaining candidate blobs. (414)

• An opening filter is applied to remove small noise particles from the binary image. (415)

• The size of the remaining blob is measured to ascertain that it is more than 50% of the size of the candidate cells identified above. (416)

• If the blob is more than 50% of the size of the candidate cell and conforms to the "closed doughnut" shape, it is confirmed to be an Alexandrium cell. (417) Therefore, the candidate cell is identified that has pixels less than 5000 has a membrane that fluoresces green and has within it an internal candidate body area of chlorophyll that fluoresces red and meets the criteria for a cell of the Alexandrium. The algorithm may be repeated for additional fields.

In an alternative example, after Alexandrium cells are identified, a second characterization signal may be generated to measure the concentration of a compound associated with. Alexandrium cell, such as the concentration of PSP toxin(s).

Computer and image processing technologies are constantly changing. Newer technologies which meet the needs of the above-described methods and apparatus, while not specifically described here, are clearly contemplated as within the invention. For example, certain conventional pixel and image file formats are mentioned above, but others may also be used. Image files may be compressed using JPEG or GIF techniques now known in the art or other techniques yet to be developed. Processing may be performed in an RGB color description space instead of the HLS space currently used. Other color spaces may also be used, as desired by the skilled artisan, particularly when detection of a sought-after characteristic is enhanced thereby.

While the embodiments of the invention have been described in connection with a sample of water, aspects of the invention may be practiced on other environmental samples as well, such as soil samples, etc. The use of a computer-controlled microscopic vision system to identify microorganisms within the sample is applicable to samples covering a full range of cell concentrations. The present invention has now been described in connection with a number of particular embodiments thereof. Additional variations should now be evident to those skilled in the art, and are contemplated as falling within the scope of the invention.

All references cited herein are incorporated herein by reference in their entirety.

Claims

What is claimed is:

1. A computer software product for detection of a microorganism or a substance produced by a microorganism in an environmental sample, comprising a computer-readable storage medium having fixed therein a sequence of instructions which when executed by a computer directs the performance of a method which comprises: a) acquiring a microscope image of an optical field of a substrate having fixed thereon an environmental sample suspected of containing microorganismsor a substance of interest, said sample treated to produce a first signal specific to a microorganism or a group of microorganisms (identification signal), if present, and to produce a second signal specific to a species of said group of microorganisms or a substance specific to said microorganism (characterization signal), if present, and transferring the image to a first color space; b) transforming the first color space to an HLS space; c) transforming the Luminance component of the HLS space to a new monochrome grey-level image and then to a binary image; d) applying a processing function to the binary image to remove blobs; e) applying a mask to exclude areas greater than characteristics of the area of the microorganism or group of microorganisms of interest and recording the location of remaining candidate areas, i.e., candidate cells; f) transforming the Hue component from the first color space to a new monochrome image and binarizing same to detect the second signal to determine candidate blobs; g) applying at least one processing function to remove candidate blobs extending beyond the field of view; and h) conforming candidate blobs that lie within a candidate cell location to detect the presence of a microorganism or substance of interest.

2. The computer software product of Claim 1 , wherein the sequence of instructions further comprises: determining the intensity of the first or second signal or both signals to quantify the microorganisms or substance of interest in the sample.

3. The computer software product of Claim 1 , wherein the microorganism detected is Pseudo-nitzschia species, Alexandrium species, Anabaena species, Chrysochromulina species, Dinophysis species, Gyrodinium species, Gymnodinium species, Microcystis species, Nodulania species, Pfiesteria or Pfiesteria-like species, Prymnesium species, Prorocentrum species, Gambierdiscus species, Ostreopsis species, Coolia species, Thecamdinium species, Amphidinium species, Pyrodinium species, Cynlindrospermosis species, Heterosigma species, Gyrodinium species, or Chaetoceros species.

4. A method for detecting a microorganism or substance of interest in an environmental sample, comprising: a) acquiring a microscope image of a n optical field of a substrate having fixed thereon an environmental sample suspected of containing microorganisms or substance of interest, said sample treated to produce a first signal specific to a microorganism or a group of microorganisms (identification signal), if present, and to produce a second signal specific to a species of said group of microorganisms or a substance specific to said microorganism (characterization signal) and transferring the image to a first color space; b) transforming the first color space to an HLS space; c) transforming the Luminance component of the HLS space to a new monochrome grey-level image and then to a binary image; d) applying a processing function to the binary image to remove blobs; e) applying a mask to exclude areas greater than characteristics of the area of the microorganisms or group of microorganisms of interest and recording the location of remaining candidate areas, i.e., candidate cells; f) transforming the Hue component from the first color space to a new monochrome image and binarizing to detect the second signal to determine candidate blobs; g) applying at least one processing function to remove candidate blobs extending beyond the field of view; and h) conforming candidate blobs that lie within a candidate cell location to detect the presence of a microorganism or substance of interest.

5. The method of Claim 4, further comprising: determining the intensity of the first or second signal or both signals to quantify the microorganisms or substance of interest in the sample.

6. The method of Claim 4, wherein the microorganism detected is Pseudo-nitzschia species, Alexandrium species, Anabaena species, Chrysochromulina species, Dinophysis species, Gyrodinium species, Gymnodinium species, Microcystis species, Nodulania species, Pfiesteria or Pfiesteria-like species, Prymnesium species, Prorocentrum species, Gambierdiscus species, Ostreopsis species, Coolia species, Thecamdinium species, Amphidinium species, Pyrodinium species, Cynlindrospermosis species, Heterosigma species, Gyrodinium species, or Chaetoceros species.

7. The method of Claim 4, wherein the characterization signal is associated with a toxic compound product by the microorganism of interest.

8. The method of Claim 4, which does not require concentration of the sample.

9. A method for computer controlled detection of a microorganism of interest, comprising conducting image analysis of a microscope field of a substrate having fixed thereon a sample, said method as depicted in Figure 3.

10. A method for computer controlled detection of a microorganism of interest, comprising conducting image analysis of a microscope field of a substrate having fixed thereon a sample, said method as depicted in Figure 4.