A kind of label-free cell detection device and method based on mating plate illumination
Technical field
The present invention relates to it is a kind of based on mating plate illumination label-free cell detection device and method, it can be achieved that sub-micron it is micro-
Particle size identification, and can be used for the classification of senile cell and normal cell, and cell ageing is related to a variety of senile diseases such as
Malignant tumour and cardiovascular disease etc..
Background technique
Flow cytometer clinically is depended on to the detection classification of cell.Conventional flow cytometer passes through using complicated
Laser beam shaping at excitation area is limited with elliptical beam similar in cell volume, is reduced interference, and tight by light path system
Lattice control sheath streaming system, keep laser beam orthogonal with sample flow, to achieve the purpose that single sample signal detection and acquisition, control
Process processed is complicated, at high cost.In terms of the detection and classification to biological cell, conventional cell instrument is usually required using various fluorescence
Reagent or dyestuff carry out complicated dye marker to cell, and dying operation process very complicated, external force operation can be to the knot of cell
Structure and function generate certain damage, in addition, fluorescence signal is weaker, optical path is relative complex.Conventional flow cytometer passes through
Granular one dimension light scattering signal, i.e. forward light scattering intensity are detected, it can be achieved that the size to particle identifies.It is this according to luminous intensity
The dimension measurement method of relative size to the stability of excitation light source, detector sensitivity with the performance of signal adapter propose compared with
High request.
Mating plate illumination is a kind of technology by beam modulation flakiness lighting source.The technology is mainly used in fluorescence microscopy
Imaging, the advantages of fluorescent molecule on focal plane is only excited using mating plate improve imaging resolution, it can be achieved that biological tissue,
The high-resolution imaging of embryo, neural network, cell mass etc..But at present mating plate means of illumination application depend on it is glimmering
Signal technology, there are certain limitations for application range.
Clinically to the detection of senile cell mainly by means of biomarker technology, i.e. SA- β-Gal dyeing detection.Aging
Cell due to beta galactosidase catalysis present navy blue, thus by optical microscopy microscopy can distinguish senile cell with not
Senile cell, the technology have been widely used in biology agings research.Another method is by low cytometric analysis point
Select senile cell and non-senile cell.In flow cytomery, the transcription regulating region building table in cell ageing is needed
Up to the reporter gene of green fluorescent protein (GFP), thus the non-senile cell of quantitative detection and senile cell.Measured by flow cytometry
Method can more clearly distinguish senile cell, but the radom insertion of reporter gene will appear additional differential expression, cause
Reporter gene is inhomogenous, influences detection effect.To sum up, above method is easy to detect, accuracy is higher, but process steps are multiple
It is miscellaneous, and the used time is longer, higher cost, while artificial diagosis is subjective and takes time and effort, and cell marking, dyeing course
Eucaryotic cell structure may be changed, influence experimental result.
Summary of the invention
The present invention to solve the above-mentioned problems, propose it is a kind of based on mating plate illumination label-free cell detection device and side
Method.The present invention uniformly excites single microparticle or cell as mode of excitation using mating plate lighting engineering, obtains the low back of sample
The two-dimentional light scattering diagram sample of scape noise, and it is applied to the rapid dimensional identification and the label-free inspection of senile cell of single microparticle
Survey classification.
To achieve the goals above, the present invention adopts the following technical scheme:
A kind of label-free cell detection device based on mating plate illumination, including the mating plate by laser beam shaping at thin photocatalytic film
Generating unit, the thin photocatalytic film uniformly excite the three-dimensional space movement on vertically and horizontally face driven by sample control unit
Sample to be tested, the form micro-image of sample and two-dimentional light scattering diagram picture are captured by image acquisition units, transfer to imaging analysis list
Member realizes automatic classification on the basis of extracting image features;
The mating plate generating unit, including laser light source, reflecting optics, neutral density piece and cylindrical lens, laser light
Source emits cylindric light beam, and after reflecting mirror controls the direction of propagation and neutral density piece adjustment light intensity, it is saturating to project cylinder
Mirror, cylindrical lens modulate width of light beam generation with a thickness of several microns to tens microns of illumination mating plate.
The sample control unit, the sample chip including carrying sample suspensions, for the fixator of fixed sample chip,
With the electricity driving displacement platform for driving sample to move along x-axis, y-axis and z-axis.
Described image acquisition unit includes microcobjective and cmos detector, and image acquisition units have focusing mode and go
Two kinds of operating modes of burnt mode;Under focusing mode, the sample to be tested after object lens alignment is accurately positioned is focused, and is detected by CMOS
The form micro-image of device record single sample;Under the mode of defocusing, object lens are capturing sample far from sample to be tested certain distance
It is distributed in the scattering light in space, which is that forward direction defocuses distance, and captured scattering light projection to cmos detector is put down
Face, detector record the two-dimentional light scattering diagram picture of the sample.
The imaging analysis unit includes Mie photon diffusion models analog module, characteristic parameter extraction module and supporting vector
The automatic categorization module of machine, the Mie photon diffusion models analog module are realized according to Mie light scattering theory model to the two of particle
Tie up light scattering analogue;The characteristic parameter extraction module extracts characteristic parameter, packet according to the sample two dimension light scattering diagram sample of capture
Include the luminous intensity scanning and Fourier methods analysis of pattern;The automatic categorization module of support vector machines, by the sample knot of acquisition
Fruit is divided into training set and test set, and according to the characteristic value of input, by finding optimized parameter, Automatic Optimal classification function is realized
The automatic classification of sample.
Label-free cell detection method based on mating plate illumination, comprising the following steps:
(1) sample chip is constructed, upper and lower surface of the two panels Conventional glass thin slice as sample chamber, two panels coverslip are selected
As gasket, both ends are respectively placed in, and are clamped between glass flake, the micro chamber that can carry sample liquor is constituted.
(2) sample suspensions are configured and are imported in sample micro chamber by pipette, ready sample chip is passed through certainly
Fixator processed is fixed on precision three-dimensional electricity driving displacement platform;
(3) the mobile sample of control electricity driving displacement platform, determines sample to be imaged, so that sample to be tested is located at object lens center, and right
Sample focuses, and triggering cmos detector records the form micro-image of the sample;
(4) start mating plate generating device, calibrate optical path, determine that laser beam passes sequentially through and reflected in mating plate generating unit
The optical axis of mirror, neutral density piece and cylindrical lens keeps beam level to propagate, and horizontal adjustment cylindrical lens position makes to illuminate
The beam waist position of mating plate is in detection object lens center;
(5) illumination mating plate excitation suspension in sample to be tested, the scattering light distribution from individual particle or individual cells in
Three-dimensional space adjusts microcobjective far from sample to be tested, object lens is made to acquire the scattering within the scope of respective angles under the mode of defocusing
Light, triggering cmos detector record the two-dimentional light scattering diagram sample of the sample;
(6) the two-dimentional light scattering diagram sample that cmos detector captures is input to Image analysis system and carries out image procossing and sample
This identification and classification.
In the step (2), self-control fixator is made of intermediate plate and fixed screw, can stable holding sample chip, and it is simultaneous
It is dissolved in optics inverted microscope.
In the step (6), classified automatically using algorithm of support vector machine to experimental result, pretreated experiment is tied
Fruit is grouped learning training and prediction, provides classification results and classification assessed value.
The Dimensions recognition system of a kind of pair of submicron particle is filled using the above-mentioned label-free cell detection based on mating plate illumination
It sets.
The categorizing system of a kind of pair of senile cell and normal cell is examined using the above-mentioned label-free cell based on mating plate illumination
Survey device.
The invention has the benefit that
(1) different scale of the invention using the lateral two-dimentional light scattering diagram sample realization particle pattern of sub-micron level of detection is sensitive
Detection, it is relatively more stable than the method in conventional flow cytometer according to one-dimensional forward direction light intensity signal;
(2) present invention uses quiescence cells suspension scanning imagery mode, gets rid of liquid complicated in conventional flow cytometer
Flow control system, easy to operate, relative inexpensiveness has broad applicability;
(3) it in traditional microscopy, is generally required when preparing biological sample smear through drying, fixation and freezing etc.
Reason, will cause certain destruction to the structure of sample, influences observing effect.In the present invention, sample to be tested is placed under suspension state
Observation and imaging, provide preferable structure and function and protecting for sample, and sample can be allowed in its closer reset condition
Under measure, be particularly suitable for living cells imaging, ensure that the authenticity and validity of DATA REASONING;
(4) present invention provides the excitation mating plate limitation lasing region of high quality using mating plate lighting method as excitation light source
Domain avoids while other cells being excited to cause to interfere, and provides preferable imaging signal to noise ratio, and flexible mating plate thickness and position are controlled
System is easy to that cell is quickly positioned and scanned, and exciting power needed for efficient capacity usage ratio can reduce imaging reduces light
Influence of the toxicity to cell activity;
(5) acquisition device of the present invention has two operating modes, can capture the morphological image and two of individual cells simultaneously
Light scattering diagram sample is tieed up, the corresponding relationship for establishing eucaryotic cell structure and optical property is facilitated, realizes the imaging to individual cells behavior
And analysis, the deficiency of light intensity signal is only provided when compensating for conventional flow cytometer to individual particle or cell analysis;
(6) present invention detects senile cell using label-free method, can be realized to the quick of aging and normal cell
Automatic classification, gets rid of conventional method to the complicated processes of cell dyeing and artificial diagosis;
(7) present invention is suitable for the imaging identification and classification of other biological cell, has popularity.
Detailed description of the invention
Fig. 1 is the structure and schematic illustration of apparatus of the present invention;
Fig. 2 (a)-Fig. 2 (c) is the different mating plate thickness mating plates and its measurement result that present example provides;
Fig. 3 (a)-Fig. 3 (f) is the micro-image and two-dimentional light for the different-grain diameter polystyrene microsphere that present example provides
Scattering pattern;
Fig. 4 is fast Fourier (FFT) the analysis result schematic diagram of experiment and simulation polystyrene microsphere light scattering;
Fig. 5 (a)-Fig. 5 (f) is that the micro-image of normal cell and senile cell that present example provides and two-dimentional light dissipate
Penetrate pattern;
Fig. 6 is characteristic parameter distribution schematic diagram of the human body at fiber aging and normal cell light scattering diagram sample.
Wherein, 1, laser, 2, reflecting mirror, 3, neutral-density filter, 4, cylindrical lens, 5, sample chamber, 6, sample
Room fixator, 7, three-D electric displacement platform, 8, microcobjective, 9, cmos detector, 10, data analysis system.
Specific embodiment:
The invention will be further described with example with reference to the accompanying drawing.
A kind of label-free cell detection device and method based on mating plate illumination, the mating plate illumination that the present invention uses can have
Effect limits excitation area, inhibits the background interference in light scattering imaging, and control process is simple and flexible.Two-dimentional light scattering technique mentions
For the light scattering on sample to be tested polar angle and two, azimuth angular range, it can get and scatter letter more abundant than one-dimensional light
Breath is realized and is identified in the microparticle size of submicron resolution level, and can be avoided complicated dying operation and fluorescence signal
Detection process realizes label-free identification and classification to senile cell.In image acquisition process of the invention, sample to be tested is natural
It is placed under suspension state, has not both needed the sheath flow control system of flow cytometer, also got rid of in optical microscopy microscopy and handle
Drying that smear is related to, fixed equivalent damage operation, ensure that the authenticity and validity of measurement.
A kind of label-free cell detection device and method based on mating plate illumination, including mating plate generating unit, sample control
Unit, image acquisition units and data processing unit.Mating plate generating unit is by laser beam shaping at the thin of micron order thickness
Mating plate, the mating plate of generation enter sample control unit excitation sample to be tested, and sample control unit is being hung down for controlling sample to be tested
Three-dimensional space on straight and horizontal plane is mobile, and image acquisition units capture the form micro-image and two-dimentional light scattering diagram of sample
Picture, and be transmitted to imaging analysis unit and carry out image procossing, feature extraction and data analysis.
Mating plate generating unit includes laser light source, reflecting optics, neutral density piece and cylindrical lens.Laser light source transmitting
Cylindric light beam projects cylindrical lens after reflecting mirror and neutral density piece, cylindrical lens squeezed light in one-dimensional direction
Beam width forms the illumination mating plate with a thickness of several microns to tens microns.Laser light source selected type is that diode semiconductor is solid
Body laser, wavelength are 532nm, and the laser beam spot sizes of generation are 1.052mm.Neutral density piece is for adjusting laser light beam intensity
Degree, optional transmitance are 50%, 32%, 10%, 1%, 0.1%.
Sample control unit includes special sample chip, fixator and electricity driving displacement platform.Sample chip is by four 170 μm
Thick glass flake is composed, wherein two panels glass flake is as sample chip upper and lower surface, and in addition two panels is put as gasket
It sets at both ends, clamping between the upper and lower surfaces, forms the chamber that volume is about 25.4mm x 10.0mm x 0.17mm, is used for
Sample loading suspension.Sample chip intermediate plate is clamped and is fixed on electricity driving displacement platform.The electricity driving displacement platform is three-axis accurate
Displacement platform can be controlled separately sample in x-axis, and y-axis and z-axis are mobile, and displacement resolution is up to 20nm or more.
Image acquisition units include microcobjective and cmos detector.Image acquisition units have focusing mode and the mould that defocuses
Two kinds of operating modes of formula.Under focusing mode, the sample to be tested after object lens alignment is accurately positioned is focused, and is remembered by cmos detector
Record the form micro-image of single sample;Under the mode of defocusing, object lens work far from sample to be tested certain distance, and the distance is fixed
Justice is that forward direction defocuses distance, and the two-dimentional light scattering diagram picture of sample to be tested is obtained in the case where forward direction removes defocus distance.In the present invention, it defocuses
Distance is 200 μm.Object lens collect the two-dimensional scattering light of the sample in corresponding angular range, pass through cmos detector record
The two-dimentional light scattering diagram sample of a sample.
Imaging analysis unit includes the simulation of Mie photon diffusion models, and the feature extraction of image and support vector machines (SVM) are certainly
Dynamic sorting algorithm.In algorithm of support vector machine, the sample results of acquisition are divided into training set and test set, by finding optimal ginseng
Number, Automatic Optimal classification function obtain classification results and assessment result.
As shown in Figure 1, a kind of label-free cell detection device based on mating plate illumination, including laser 1, reflecting mirror 2, in
Property density filters 3, cylindrical lens 4, sample chamber 5, sample room's fixator 6, three-D electric displacement platform 7, microcobjective 8,
Cmos detector 9, data analysis system 10.
Specifically include following procedure:
The laser beam that laser 1 issues adjusts the direction of propagation by reflecting mirror 2, and is controlled by neutral-density filter 3
Laser beam after laser intensity processed, adjustment direction and energy projects cylindrical lens 4,4 vertical direction modulation light of cylindrical lens
Beam compresses light beam in one-dimensional direction, is shaped to the thin photocatalytic film of micron order thickness, and the size of mating plate thickness depends on light
Beam projects the effective numerical aperture size of cylindrical lens;
Mating plate is illuminated from 5 side illumination sample suspensions of sample chamber, to excite the sample to be tested in suspension.Fixator 6
Sample room 5 is fixed on three-D electric displacement platform 7, three-D electric displacement platform 7 controls sample in vertical direction and horizontal plane
On accurate movement;
After sample is excited by laser light-piece, the scattering light for being distributed in three-dimensional space is detected the detection of object lens 8 and collects, CMOS
Detector 9 captures two-dimentional light scattering diagram sample and records;
The image data of record is transmitted to analysis system 10 and carries out data processing and analysis.
Embodiment 1
Using the label-free cell detection device and method illuminated based on mating plate, the Uniform Illumination light of different-thickness is realized
Piece.In light scattering imaging, to reduce the ambient noise in imaging process, and guarantee simultaneously sample to be tested by Uniform Illumination,
The mating plate for providing different-thickness for the sample of different size scale in the present invention excites, and has pointedly control mating plate thickness
Degree limits excitation area, avoids the interference from other sample scatter light.By control light beam expand multiple and cylinder is saturating
The focal length of mirror generates the mating plate of required thickness.The light of different-thickness is realized in this example by conversion different focal length cylindrical lens
Piece.Different thickness is visualized by rhodamine solution, and measures mating plate thickness by vertical scanning image pixel.
Concrete operation step:
(1) rhodamine 6G solution is configured, and solution is imported in the sample chamber built;
(2) open laser light source, calibrate optical path, make laser beam pass sequentially through optical element include reflecting mirror, it is neutral close
The optical axis of piece and cylindrical lens is spent, horizontal transmission is kept.90 ° of Rotating cylindrical surface lens, compress light beam in the horizontal direction,
Cylinder lens position is adjusted, makes to illuminate mating plate beam waist position in detection object lens center;
(3) increase wavelength filter before cmos detector, filter out excitation wavelength (532nm), allow rhodamine solution
Launch wavelength (575nm) pass through, triggering cmos detector record mating plate visualization after image;
(4) by vertical scanning image pixel, gray value is obtained, determines the with a tight waist of mating plate, and measure its mating plate thickness, this
Place uses halfwidth (FWHM) as thickness and portrays parameter;As a result as shown in Fig. 2 (a), Fig. 2 (b), Fig. 2 (c), mating plate thickness
Respectively 5.6 μm, 13.7 μm and 53.3 μm.
Embodiment 2
In order to verify the sensitivity and accuracy that the present invention identifies micron level particle size, polystyrene mark is used
Quasi- microballoon carries out experimental verification and device calibration.In the present invention, the partial size is selected to be for the standard microballoon of 3.87 μm and 4.19 μm
Imaged samples obtain its micro-image and two-dimensional scattering figure, and the result of experimental result and Mie simulation are carried out FFT comparison
Analysis.Selected microcobjective is 20 times of object lens that numerical aperture is 0.4, corresponding detectable angular range in the present invention
It is 72.5 ° -107.5 °, so in Mie simulation, scatters polar angle and azimuth value range thus.Mating plate is adjusted in this example
With a thickness of 13 μm or so.
Concrete operation step:
(1) appropriate polystyrene microsphere stoste is drawn, is diluted with ultrapure water, and microsphere suspensions are imported into sample microcavity
Room;
(2) sample chip for being loaded with microsphere suspensions prepared is fixed on three-D displacement platform by fixator, is controlled
Displacement platform, focusing objective len determine the imageable target in visual field, finely tune displacement platform, targeted microspheres is made to be located at object lens middle position;
(3) cmos detector is triggered, the form micro-image under focusing mode of microballoon is recorded;
(4) laser light source is opened, mating plate generating unit device is adjusted, keeps laser beam holding horizontal and in optical element
On central optical axis, the mating plate and targeted microspheres of formation are on sustained height, and mating plate beam waist position is maintained at the object lens visual field
Center;
(5) object lens are finely tuned, forward direction defocuses 200 μm, makes object lens work collection of scattered light under the mode of defocusing, and triggering CMOS is visited
Device is surveyed, the two-dimentional light scattering diagram sample of the microballoon is recorded;
(6) according to Mie light scattering theory model, the two-dimentional light scattering result of microballoon is simulated.Simulated conditions are optical source wavelength
532nm, microballoon refractive index 1.59, solution medium refractive index 1.334 locating for microballoon, scatters polar angle and azimuthal angular range is
72.5°-107.5°;
(7) horizontal sweep is carried out to the two-dimentional light scattering diagram picture of experiment and simulation respectively, obtains the gray scale of light scatter intensity
It is worth curve.The quantitative analysis of one-dimensional FFT is carried out by Fourier transformation, the frequency main peak value in FFT map can quantify light scattering
The fringe distribution of pattern.According to the difference for the frequency main peak value that the distribution of the light scattered striation of various sizes of microballoon generates, realize
Size identification and classification to microballoon.In this example, the frequency step of FFT sampling is set as 0.0138 (1/degree).
Shown in the experimental result of this example such as Fig. 3 (a)-Fig. 3 (f).Fig. 3 (a) and Fig. 3 (d) is 3.87 μm and 4.19 μ respectively
The micro-image that the microballoon of m obtains under 20 times of object lens focusing task modes.Fig. 3 (b) and Fig. 3 (e) is under the mode of defocusing respectively
The two-dimentional light scattering diagram sample of the two kinds of beads obtained.The size difference of two kinds of microballoons is 320nm, focuses mould in ordinary optical object lens
More difficult intuitive differentiation two kinds of particles under formula.In contrast, the two-dimentional light scattering diagram sample of two kinds of beads is in striped quantity and distribution
Notable difference is presented, there are 7 bright fringes, 8 bright fringes occurs in 4.19 μm of microballoons in the light scattering diagram sample of 3.87 μm of microballoons,
Therefore, two kinds of microballoons directly can quickly be distinguished by judging the striped quantity of light scattering diagram sample.Fig. 3 (c) and Fig. 3 (f) is root
According to the light scattering diagram sample of two kinds of beads of Mie modeling as a result, the result consistent scattering strip compared with experimental result is presented
Line distribution.Further utilize the difference of two kinds of beads of fast Fourier FFT method quantitative analysis and experiment and analog result
Identical property, as a result as shown in Figure 4.For 4.19 μm of microballoons, simulation and the FFT frequency main peak value of experimental result are 1.432 (1/
degree);For 3.87 μm of microballoons, the frequency main peak value of experimental result is 1.2963 (1/degree), and analog result is
1.2825 (1/degree), the difference of only one sampling step length, also than more consistent.This example demonstrates the present invention for sub-micro
The feasibility and accuracy that rice resolution ratio particle size identifies.
Embodiment 3
Using the label-free cell detection device and method illuminated based on mating plate, to human body at fiber senile cell and normally
Cell carries out cell recognition and automatic classification.It is about 50 μm that mating plate thickness is adjusted in this example.Aging is induced to by hydrogen peroxide
Human body carry out cell morphology image at fiber senile cell and normal cell and its two-dimentional light scattering diagram sample acquisition, every class are random
Acquire 55 experimental results.Characteristic parameter extraction is carried out to 110 light scattering diagram pictures of acquisition, and inputs support vector machines
(SVM) algorithm classifies automatically to two class cells as characteristic parameter.
Specific implementation:
(1) prepare the processed human body of hydrogen peroxide into fiber senile cell and normal cell, be configured to carefully with PBS buffer solution
Born of the same parents' suspension imports sample micro chamber, control bit moving stage and imaging unit device, and the micrograph of cell is obtained under focusing mode
Picture;Such as Fig. 5 (a), Fig. 5 (b).
(2) starting mating plate lighting device obtains the two-dimentional light scattering diagram of two class cells under the 200 μm of modes that defocus respectively
Sample, such as Fig. 5 (c), Fig. 5 (d).
(3) algorithm detects and extracts the characteristic parameter of speckle distribution.Local intensity maxima point is detected first, determines spot
Position and center.In the present invention, pixel more than intensity 6 units bigger than surrounding pixel is defined as maximum intensity value point.
After determining spot centers, the pixel region being less than within the scope of 6 strength differences of the central point is defined as with the local maxima
Speckle area centered on hot spot.110 experimental results are scanned, the numerical value ginseng of corresponding speckle count and speckle area is extracted
Number characterizes different cells according to speckle count and the distribution relation of speckle area, and realization is to senile cell and normal cell
Classification, as shown in Figure 6.
(4) using the numerical value of speckle count and speckle area as the characteristic of division value of support vector machines, when executing SVM algorithm
Linear kernel function is chosen, and classifier is constructed using 5- folding cross-validation method.110 data of two class cell results are divided at random
For 5 data sets, each data set has 22 samples, in turn using 4 data sets therein as training set, a remaining number
After being used as test set, repeated overlapping to verify 5 times according to collection, using the average value of the accuracy of 5 prediction results as finally accurate
Degree.Classification results are as shown in table 1.Wherein, the cell quantity that accuracy rate is defined as correctly being classified accounts for whole cell quantities
Percentage;Sensitivity definition is the ratio of the quantity that senile cell can correctly be classified and senile cell total quantity;Specificity
It is defined as the ratio of quantity and normal cell total quantity that normal cell is correctly classified.AUC parameter is for assessing classifier
Performance is worth closer to 1, and presentation class device performance is better.
1 human body of table at fiber aging and normal cell svm classifier result
Above-mentioned, although the foregoing specific embodiments of the present invention is described with reference to the accompanying drawings, not protects model to the present invention
The limitation enclosed, those skilled in the art should understand that, based on the technical solutions of the present invention, those skilled in the art are not
Need to make the creative labor the various modifications or changes that can be made still within protection scope of the present invention.