JP4137682B2 - Fluorescence spectrometer - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a fluorescence spectroscopy analyzer, and more particularly to a fluorescence spectroscopy analyzer that performs fluorescence statistical distribution analysis at a plurality of points.
[0002]
[Prior art]
Methods using the technique of statistical distribution analysis of fluorescent species are used for genome-related analysis such as DNA base sequence analysis, SNPs (single nucleotide polymorphism) analysis, and protein interaction analysis. For statistical distribution analysis, fluorescence correlation spectroscopy, fluorescence intensity distribution analysis, and the like are known.
[0003]
Fluorescence correlation spectroscopy is a technique for analyzing fluctuations in fluorescence intensity and obtaining the diffusion time and average number of molecules for each molecule (see, for example, Non-Patent Document 1).
[0004]
The fluorescence intensity distribution analysis method is a method for obtaining the photon count (brightness) and the average number of molecules for each molecule in the observation region from the fluorescence fluctuation data (for example, see Non-Patent Documents 2 and 3).
[0005]
In addition, the fluorescence lifetime analysis method is a method for obtaining the fluorescence lifetime and the average number of molecules for each molecule by analyzing the decay time of fluorescence (see, for example, Non-Patent Document 1).
[0006]
Conventionally, such an analysis has been performed in a solution. However, with the progress of biotechnology, it is required to perform the same analysis in a cell. FIG. 6 is a diagram showing a configuration of a conventional fluorescence spectroscopic analyzer. In this apparatus, a laser is introduced into a microscope, and fluorescence generated from fluorescent molecules in a solution excited by the laser light is measured to perform statistical distribution analysis.
[0007]
A laser light source 70 for exciting the fluorescent dye is provided on the upper part of the optical microscope, and laser light is incident thereon. The laser light is condensed by the objective lens 71 to create an observation region in the sample container 72. In the sample container 72, a solution of a fluorescence-labeled molecule to be observed (or a substance to be observed such as a vesicle) is placed. Fluorescence generated from the fluorescent label of the sample existing in the observation region passes again through the objective lens 71 and is guided to the detector 74 by the beam splitter 73. A condensing lens 75 and a pinhole 76 are provided in front of the detector 74 to perform confocal detection.
[0008]
With this configuration, fluorescence from only a very small observation area can be detected. The fluorescence signal obtained by the detector 74 is analyzed by an analysis device 77, and a desired measurement such as measurement of protein interaction is performed.
[0009]
In order to perform statistical distribution analysis on cells, it is conceivable to apply this fluorescence spectroscopic analysis technique to cells instead of solutions. That is, using this apparatus, a petri dish or the like is used instead of the sample container 72, and the cell is observed with an optical microscope to specify the measurement site. Then, the petri dish is moved by a stage moving mechanism (not shown), and the specified measurement site is moved to the observation region by the laser for observation.
[0010]
If the conventional technique is applied in this way, the technique of statistical distribution analysis can be applied to intracellular measurement. However, when the conventional technique is applied, sufficient observation is impossible because there is only one measurement point. Therefore, in order to solve this problem, a method of forming a plurality of laser beams and performing fluorescence statistical distribution analysis at a plurality of points has been proposed (for example, see Non-Patent Document 4).
[0011]
[Non-Patent Document 1]
Ch. Zander, J. Enderleing, RA keIler (eds.), Single molecule Detection in Solution, 69P. And 59p. (WILEY-VCH Verlag Berlin, 2002)
[0012]
[Non-Patent Document 2]
Yan Chan et.al., The Photon Counting Histogram in Fluorescence Fluctuation Spectroscopy, Biophysical Journal Vol.177 July 1999 553-567.
[0013]
[Non-Patent Document 3]
Hong Qian and Elliot L. Elson, Distribution of molecular aggregation by analysis of fluctuation moments, Proc. Natl. Acad. Sci, USA, Vol.87, pp.5479-5483, July 1990.
[0014]
[Non-Patent Document 4]
Applied Optics, vol.41, No.16, pp3336-3342 (2002).
[0015]
[Problems to be solved by the invention]
By the way, when observing cells, it is necessary to select observable cells, unlike the conventional observation in solution. That is, it is necessary to determine whether or not the fluorescence signal is reliable due to the intensity of fluorescence staining, to determine whether or not the shape of the cell is normal, and to determine whether or not the shape / size of the internal tissue is normal. .
[0016]
This cell pass / fail judgment operation has been determined by an expert based on experience and intuition based on the observation results of the cells described above. Therefore, whether or not the technique of single molecule fluorescence statistical distribution analysis is applicable to the cell is determined based on the observation result of the cell.
[0017]
In addition, when a user intends to observe a specific cell (for example, a cell having a specific shape or a cell expressing a specific abnormality), it is convenient if the cell to be observed can be easily specified. However, the identification of the cells to be observed takes a long time because there are large individual differences depending on the skill level.
[0018]
For this reason, when a large number of cells are observed / measured as routine work and the quality of the cells is determined or the cells are specified, it is required that the determination / specification work can be performed efficiently. Specifically, in order to elucidate intracellular protein interactions, transport problems, gene expression problems, etc. by single-molecule fluorescence statistical distribution analysis, the cells suitable for observation should be determined or identified quickly and accurately. Is required.
[0019]
The present invention has been made in view of such circumstances, and an object of the present invention is to provide a fluorescence analyzer that can quickly and accurately determine or specify a cell suitable for observation.
[0020]
[Means for Solving the Problems]
In order to solve the above problems, a fluorescence spectroscopic analyzer according to the first aspect of the present invention includes a container that holds a sample, an excitation optical system that irradiates the sample with a plurality of excitation lights, and each excitation light that is emitted. A plurality of fluorescence detection elements for detecting the respective fluorescence emitted from the obtained sample, a detection optical system for guiding the fluorescence to each detection element, an analyzer for analyzing the obtained plurality of fluorescence signals, and an image of the sample Whether each observation region of the sample is suitable for fluorescence detection and fluorescence signal analysis based on a photographing device to be photographed, an observation optical system for guiding an image to the photographing device, and an image of the sample photographed by the photographing device A determination / identification device that determines whether or not the observation region of the sample is suitable for fluorescence detection and fluorescence signal analysis, and the determination / specification device is based on the image of each observation region of the sample. cells of at least the sample More defined based staining shades, shape and size of the cell, and the image feature extraction unit for extracting a feature value representing the position and shape of the cells within the tissue, extracted the feature amount in each of the plurality of levels The image classification unit for classifying into any of the above aspects, the classification result of the feature quantity and the quality of the analysis result by the analyzer are associated with each other to learn the suitability / non-suitability of each observation region for each analysis. A determination learning unit that determines whether the observation region is appropriate or inappropriate based on the result of the classification based on the result .
[0021]
A fluorescence analyzer according to another aspect of the present invention is the above-described fluorescence spectrometer, wherein the determination / specification device uses at least one new feature suitable for the determination or identification from the feature amount. A feature amount selecting unit that extracts an amount; and a control unit that executes the determination or the specification as the feature amount for performing the determination or the specification of the new feature amount.
[0022]
A fluorescence analyzer according to another aspect of the present invention is the fluorescence spectrometer according to the above-described invention, wherein the feature quantity selection means is at least one suitable for the determination or the identification from the feature quantity. It further has a feature amount learning means for improving the reliability of logic for extracting a new feature amount by learning.
[0025]
A fluorescence analyzer according to another aspect of the present invention is the fluorescence spectrometer according to the above-described invention, having a plurality of containers for holding the sample, and sequentially moving the positions of the containers to the observation position. Thus, fluorescence detection and fluorescence signal analysis of the sample are performed.
[0026]
A fluorescence analyzer according to another aspect of the present invention is the fluorescence spectrometer according to the above-described invention, provided in a common optical path of the excitation optical system and the detection optical system, and condensed in the sample. A light beam deflecting means for moving the fluorescence measurement space in the container is provided.
[0029]
DETAILED DESCRIPTION OF THE INVENTION
[Configuration of fluorescence spectroscopic analyzer]
FIG. 1 is a configuration diagram showing a first embodiment of a fluorescence spectroscopic analyzer according to the present invention.
[0030]
In order to perform statistical distribution analysis by single-molecule fluorescence spectroscopy measurement at a plurality of locations, the fluorescence spectroscopy analyzer is configured using a plurality of laser light sources 1. The plurality of laser light sources 1 are connected to an optical fiber 2, and the plurality of optical fibers 2 are installed so that the end surface arrangement thereof is an imaging surface of the imaging lens. In other words, the fiber end is arranged one-dimensionally or two-dimensionally on the imaging plane of the lens, and this is reduced and projected into the sample by the microscope optical system to create a plurality of observation regions. This microscope optical system corresponds to an excitation optical system.
[0031]
The plurality of laser beams emitted from the fiber end are transmitted through the beam splitter 3 for branching to the detection optical system, the imaging lens 4 and the branching beam splitter 5 for visual observation, and the sample lens 7 is passed through the objective lens 6. Is reduced and projected into the sample. This state can be observed by being superimposed on the transmission image of the sample by the visual optical system. The observer can move the observation region to an arbitrary position in the sample while observing the transmission image. This observation may be performed by the display device 9 through the camera 8. The movement can also be performed by the electric stage 10.
[0032]
As described above, the fluorescence spectroscopic analysis apparatus according to the present embodiment constitutes a so-called inverted microscope, and a sample such as a cell to be observed in the sample container 7 illuminated by the observation light source 12 and the condenser 13 is used as the objective lens 6. It can be magnified and observed. In addition, it is possible to observe a phase contrast microscope, a differential interference microscope (DIC), and the like provided in a normal biological observation microscope. In addition to visual observation, an image can be captured by the camera 8 and displayed on the display device 9. It has become. That is, it has the function of a normal microscope.
[0033]
Furthermore, the fluorescence analyzer of the present embodiment includes a determination device 30 that automatically determines whether or not a cell to be observed is suitable for observation.
[0034]
Appropriate / inappropriate determination or identification of the cell to be observed according to the present invention is performed by using a fluorescence image or a visible light observation image (transmission image) of the cell to be observed captured by the camera based on the information on the quality of the fluorescence distribution analysis result from the analyzer. , Phase difference image, differential phase image (DIC), and the like) by appropriately classifying the image analysis results.
[0035]
The determination device 30 determines whether the cell to be observed is suitable for observation based on a fluorescence image or a visible light observation image (transmission image, phase difference image, differential phase image (DIC), etc.) of the cell to be observed captured by the camera 8. It is determined automatically. On the other hand, the information about the quality of the fluorescence distribution statistical analysis result is input from the analysis device 10 to the determination device 30. The determination device 30 optimizes the determination logic by repeatedly learning the comparison between the analysis result and the determination result for each observation cell. The determination device 30 will be described in detail later.
[0036]
Next, acquisition of fluorescence statistical distribution information will be described.
[0037]
FIG. 2 is a diagram showing a detection optical system. The detection optical system 11 includes a plurality of detectors 15 corresponding to the pinhole plate 14 and performs confocal detection. In the present embodiment, the plurality of detectors 15 are used. However, the present invention is not limited to this embodiment, and a single detector having a plurality of detection elements may be used.
[0038]
The position of each pinhole on the pinhole plate 14 is adjusted to coincide with the observation region. In other words, the position of the pinhole is at a position conjugate with a plurality of light sources composed of a plurality of fiber end faces. The plurality of fluorescence intensity signals obtained in this way are subjected to statistical distribution analysis processing such as fluorescence correlation spectroscopy and fluorescence intensity distribution analysis in the analysis device 16.
[0039]
In this way, information on the state of protein reaction in the observation region, measurement of specific protein concentration, detection of specific molecules, and the like can be obtained.
[0040]
FIG. 3 is a diagram showing a state of the observation region of the cells in the sample container.
There are four observation regions 23 a, 23 b, 23 c, and 23 d in the cell 22 in the sample container 7. As described above, the observation regions 23a to 23d can be moved to arbitrary positions while performing observation with the microscope optical system. Since the statistical distribution analysis information of each point can be obtained simultaneously from the plurality of observation regions 23a to 23d, for example, by analyzing the observation region A23a and the observation region B23b, the progress direction or speed of the biochemical reaction is analyzed. Is possible.
[0041]
In order to perform statistical distribution analysis by single molecule detection, it is desirable that about one molecule exists in the observation time per observation region. In addition, since a very large number of measurements are required for statistical distribution data analysis, it is preferable to have a condition in which the movement time due to the Brownian motion of molecules can be observed in a short time.
[0042]
For example, in the fluorescence intensity distribution analysis, one observation time of about 40 μsec is used. Also, almost the same observation time is used in fluorescence correlation spectroscopy. The size of the observation region where one molecule exists within this time is about 1 μm cube. Therefore, the diameter of the excitation laser beam 24 needs to be condensed to 0.5 to 1 μm. The wavelength of visible light normally used for excitation is about 0.5 μm, and the numerical aperture of the condenser objective lens 6 needs to be about 1.0. A normal single mode fiber has a numerical aperture of about 0.1, and the magnification of the system needs to be about 5 times in order to enter the objective lens 6 while maintaining a Gaussian beam shape. Further, in order to make the numerical apertures coincide with each other, it is necessary to make the optical system of the system 10 times.
[0043]
[Determination device configuration]
FIG. 4 is a block diagram illustrating a signal flow related to the configuration of the determination device 30.
[0044]
The determination device 30 includes an image feature extraction unit 31, an image classification unit 32, and a recording / learning unit 33. The image feature unit 31 processes the cell image sent from the camera 8 and extracts a characteristic parameter. The image classification unit 32 classifies the cell image based on this parameter. The determination / learning unit 33 automatically determines the suitability / unsuitability of the cells based on the classification result, and optimizes the decision logic by taking in the good / bad result information from the analysis device 16 and learning.
[0045]
[Operation]
Next, the operation of the determination device 30 will be described.
[0046]
From the camera 8, an image of a cell to be observed is input. As the image, a transmission image by visible light (non-fluorescence), a phase difference image, a differential phase image (DIC), a fluorescence image, or the like can be used.
[0047]
The image feature extraction unit 31 extracts the feature amount (parameter) of the cell from the input image. The feature quantity to be extracted includes information on the intensity of fluorescence staining such as the intensity of staining of intracellular molecules, or information on the shape of cells. The amount that characterizes the cell shape includes cell size, shape, color, color density, contour curvature, or the location, size, shape, color, color density, contour curvature, number of intracellular tissues, etc. There is.
[0048]
Furthermore, a value obtained by converting these items using various functions may be used as the feature amount. For example, these items may be used as feature quantities in the frequency space by Fourier transform. Such conversion is convenient because feature quantities that do not depend on the direction or position of cells can be extracted.
[0049]
The image classification unit 32 classifies the cell image based on the plurality of extracted feature amounts.
[0050]
FIG. 5 is a diagram showing image classification. In FIG. 5, images are classified by two parameters, “cell size” and “cell shape”.
[0051]
Then, the determination / learning unit 33 determines whether or not the observed cell is suitable for the fluorescence statistical distribution analysis based on the classification result. In the present embodiment, the determination / learning unit 33 determines that the cell to be observed is “medium” and “shape is Δ”, or the cell to be observed is “small” and “shape is □”. If so, judge that it is appropriate. Then, this determination result is presented to the user.
[0052]
Next, the learning operation of the determination / learning unit 33 will be described.
[0053]
The image feature extraction unit 31 and the image classification unit 32 perform the above-described operation for each cell to be observed, extract the size, shape, etc. of the drawing cells and classify the images. On the other hand, an analysis result indicating whether the observed cell has been properly analyzed by statistical fluorescence distribution analysis (good) or failed in analysis (bad) is input from the analysis device 16 to the determination / learning unit 33.
[0054]
Therefore, the determination / learning unit 33 associates the previous classification result with the analysis result, and learns, for example, that the classification result is “good”. The determination / learning unit 33 optimizes the determination logic by repeatedly executing the classification and analysis result correspondence and learning.
[0055]
Note that the analysis result is not limited to the one automatically input from the analysis device 16, and may be one manually input by the user at an arbitrary time, for example. Further, the user may input not only the quality of the distribution analysis result but also a determination result for specifying whether or not the cell can be an observation target such as abnormal gene expression. In this case, learning of the determination logic for cell identification different from the above-described determination logic is performed.
[0056]
In the present embodiment, as the above learning is repeated several times, the determination / learning unit 33 determines that “the size is medium”, “the shape is Δ (triangle)”, and “the size is small” and “the shape is small”. Is a square (square), it is learned that the cell is suitable (or can be specified) for fluorescence statistical distribution analysis.
[0057]
For example, a neural network may be used as a method for realizing the learning of the present embodiment.
[0058]
FIG. 5 shows an application example for at most two parameters, but more parameters are actually used. At this time, it is important to select what parameters are suitable for the determination, that is, parameters that can effectively determine appropriate and inappropriate cells. For this purpose, for example, techniques of multivariate analysis and factor analysis may be used, and effective parameters may be selected by learning. Then, these methods may be incorporated into the determination / learning unit 33, and appropriate parameters may be automatically extracted, and appropriate / unsuitable cells may be determined based on the extracted parameters.
[0059]
[effect]
The determination logic can be optimized by the learning function of this embodiment. Then, it is possible to easily and automatically determine the suitability / unsuitability of the cell to be observed or specify the cell to be viewed by the optimized judgment logic.
[0060]
In the present embodiment, for the above purpose, it is determined whether the sample is suitable for analysis by determining the intensity of fluorescent staining for the sample, or the intensity of intracellular molecules and the density of intracellular tissues are determined. It is preferable to determine whether the sample (cell) is suitable for analysis by determining the arrangement and shape. A sample image may be used as a comparison target.
[0061]
Furthermore, in the present embodiment, by learning the characteristic amount such as the shape and size of a cell suitable for fluorescence measurement and analysis, the position and shape of intracellular tissue, and the staining concentration, and the result of fluorescence spectroscopic analysis at every analysis. Each time it is used, the judgment / specific performance can be improved.
[0062]
In this embodiment, the shape and size of cells suitable for fluorescence measurement and analysis, the position and shape of intracellular tissues, and the amount of features such as staining concentration and the result of fluorescence spectroscopic analysis are learned each time analysis is performed. It has a function (learning function) that improves the reliability by selecting an effective parameter for the determination. For this reason, the “intuition” of the judge who has relied on the conventional method is introduced into the analysis apparatus by the learning function, which contributes to elimination of individual differences of the judge and improvement of efficiency.
[0063]
Next, a second embodiment of the fluorescence spectrometer according to the present invention will be described.
[Configuration of fluorescence spectroscopic analyzer]
Since the configuration of the second embodiment is the same as that of the first embodiment shown in FIG. 1, the same reference numerals are given, and detailed description thereof is omitted.
[0064]
[Operation]
In the second embodiment, when there are a plurality of cells in the observation visual field, the electric stage 10 shown in FIG. 1 is automatically moved. Then, appropriate / inappropriate determination is performed for each cell, or cells to be viewed are sequentially identified.
[0065]
Further, it may be configured to have a plurality of containers for holding a sample and sequentially detect and analyze fluorescence by moving the positions of the containers sequentially.
[0066]
[effect]
According to the second embodiment, the necessary fluorescence statistical distribution analysis can be automatically performed, so that work efficiency can be improved.
[0067]
Next, a third embodiment of the fluorescence spectroscopic analyzer according to the present invention will be described.
[Configuration of fluorescence spectroscopic analyzer]
Since the configuration of the third embodiment is the same as that of the first embodiment shown in FIG. 1, the same reference numerals are given and detailed description thereof is omitted.
[0068]
In the third embodiment, when there are a plurality of cells in the observation visual field, the observation point is moved by a laser beam deflecting means in an optical path (not shown). That is, the fluorescence condensed in the sample by providing a light beam deflecting means in the common optical path of the optical system (excitation optical system) leading to the excitation light source and the sample and the optical system (detection optical system) leading from the sample to the fluorescence detector. The measurement space can be moved.
[0069]
[effect]
Fluorescence measurement space focused in the sample by the beam deflecting means provided in the common optical path of the excitation light source and the optical system leading to the sample (excitation optical system) and the optical system leading from the sample to the fluorescence detector (detection optical system) By having the function of moving the observation area, the observation area can be quickly moved to an arbitrary position of the cell.
[0070]
In addition, by having a plurality of fluorescence measurement spaces with a plurality of excitation light sources and a plurality of fluorescence detectors, it becomes possible to measure a plurality of observation points in the selected observed cell at the same time (no time difference is a problem). Various analyzes such as transportation problems and temporal changes in gene expression distribution are possible.
[0071]
In addition, since the above-described embodiments include various stages of the invention, various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. For example, even if some constituent requirements are deleted from all the constituent requirements shown in the embodiment, the problem described in the column of the problem to be solved by the invention can be solved, and the effect described in the column of the effect of the invention Can be extracted as an invention.
[0072]
【The invention's effect】
As described above, according to the fluorescence spectroscopic analyzer of the present invention, cells suitable for observation can be determined or specified quickly and accurately.
[Brief description of the drawings]
FIG. 1 is a configuration diagram showing a first embodiment of a fluorescence spectrometer according to the present invention.
FIG. 2 is a diagram showing a detection optical system.
FIG. 3 is a view showing a state of an observation region of cells in a sample container.
FIG. 4 is a block diagram showing a signal flow related to the configuration of the determination apparatus.
FIG. 5 is a diagram showing image classification.
FIG. 6 is a diagram showing a configuration of a conventional fluorescence spectroscopic analyzer.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Laser light source, 2 ... Fiber, 3 ... Beam splitter, 4 ... Imaging lens, 6 ... Objective lens, 9 ... Display apparatus, 10 ... Electric stage, 11 ... Detection optical system, 15 ... Detector, 16 ... Analysis apparatus , 23 ... observation region, 30 ... determination device, 31 ... image feature extraction unit, 32 ... image classification unit, 33 ... determination / learning unit.

Claims (5)

A container for holding a sample;
An excitation optical system for irradiating a sample with a plurality of excitation lights; and
A plurality of fluorescence detection elements for detecting each fluorescence emitted from the sample irradiated with each excitation light;
A detection optical system for guiding fluorescence to each detection element;
An analyzer for analyzing the obtained plurality of fluorescence signals;
A photographing machine for photographing an image of the sample;
An observation optical system for guiding an image to the photographing machine;
Based on the image of the sample photographed by the photographing machine, it is determined whether each observation region of the sample is suitable for fluorescence detection and fluorescence signal analysis, or the sample suitable for fluorescence detection and fluorescence signal analysis Equipped with a judgment and identification device that identifies the observation area of
The determination / specification device is:
An image feature extraction unit that extracts at least light and shade of fluorescence staining of the cells of the sample based on the image of each observation region of the sample, a shape and size of the cell, and a feature amount representing the position and shape of the intracellular tissue ;
An image classifying unit that classifies the extracted feature quantities into any of a plurality of modes defined based on a plurality of levels ;
Corresponding the classification result of the feature quantity and the quality of the analysis result by the analyzer, learning appropriate / inappropriate of each observation region for each analysis, and based on this learning result, A fluorescence spectroscopic analyzer, comprising: a determination learning unit that determines suitability / unsuitability .   The determination / identification device is configured to extract at least one new feature quantity suitable for the determination or the identification from the feature quantity, and to perform the determination or the identification of the new feature quantity. The fluorescence spectroscopic analysis apparatus according to claim 1, further comprising a control unit that executes the determination or the identification as the feature amount.   The feature quantity selecting means further comprises feature quantity learning means for improving the reliability of logic for extracting at least one new feature quantity suitable for the determination or the identification from the feature quantity by learning. The fluorescence spectroscopic analyzer according to claim 2. A plurality of containers for holding the sample, by moving sequentially the observation position the position of the vessel, any of claims 1 to 3, characterized in that the fluorescence detection and fluorescence signal analysis of the sample The fluorescence spectroscopic analyzer according to claim 1. 2. A light beam deflecting unit that is provided in a common optical path of the excitation optical system and the detection optical system and moves a fluorescence measurement space condensed in the sample within the container. 5. The fluorescence spectroscopic analyzer according to any one of 4 .

Images