CN106644902A - Evaluation method of stability of laminar flow of flow cytometer - Google Patents

Abstract

The invention provides an evaluation method of stability of a laminar flow of a flow cytometer. The evaluation method comprises the following steps of (1) detecting 90-degree Mie scattered light of microspheres in a flowing chamber by utilizing a high-speed microscopic image acquisition system; (2) performing cluster analysis on situations that the length of the light intensity/smear is insufficient, normal, diffracted, overlapped and the like of the acquired massive images by utilizing a grey cluster analyzing method, and obtaining a standard normal smear; (3) determining the boundary of the smear by utilizing a mid-point method, and calculating the corresponding flow velocity of the microspheres; (4) characterizing the stability of a liquid path system of the flow cytometer by utilizing the stability of the flow velocity of the microspheres.

Description

Flow cytometer laminar flow stability evaluation method

Technical Field

The invention relates to the field of microsphere velocity measurement statistical analysis, in particular to the field of evaluation of the stability of laminar flow in a flow chamber of a flow cytometer.

Background

The flow cytometer is a clinical examination analyzer which can realize high-speed one-by-one multi-parameter quantitative analysis of single cells or other particles flowing in a straight line at a high speed in a suspension by detecting scattered light signals and/or fluorescence signals. The main purpose of the liquid path system is to form a stable laminar flow by wrapping a sample liquid containing a sample to be detected (cells or microspheres) with a sheath liquid, thereby achieving the purpose of obtaining a single cell flow. The stability of the liquid path system will directly affect the position and time of the cells/microspheres passing through the detection area of the flow cell, and further affect the signal intensity and the light pulse duration of the corresponding scattered light and fluorescence signals. The stability of the liquid path system is evaluated, particularly the speed stability of cells/microspheres passing through a detection area of the flow chamber is evaluated, and the stability of the whole instrument can be quickly pre-judged.

At present, the method for judging the stability of a liquid path system of a flow cytometer mainly comprises a pressure method and a pulse signal characteristic analysis method. The pressure method is characterized in that the stability of the liquid path system can be judged by observing the most critical sample liquid pressure and sheath liquid pressure in the liquid path system and aiming at different detection rate requirements, and the change amplitude of the two is within a certain range. However, the pressure method is to detect the gas pressure acting on the sample liquid and the sheath liquid, rather than directly detecting the liquid flow rate, and therefore, the influence of the subsequent sample introduction structure and pipeline on the laminar flow and the cell/microsphere speed cannot be measured. The pulse signal characteristic analysis method is to detect scattered light and fluorescence signals generated when cells pass through a detection area of a flow chamber through a data acquisition module, and to represent the stability of the cell speed by using the stability of the obtained pulse width. The method needs to complete a series of operations such as scattered light excitation, collection, photoelectric conversion, pulse processing, parameter extraction and the like of cells/microspheres, and relates to an optical path system and an electronic circuit processing system, so that uncertain factors of a measurement process are increased, and the flowing condition of the cells/microspheres in a flow chamber cannot be truly reflected.

The high-precision flow field characteristic analysis method is mainly a Particle Image Velocimetry (PIV), the velocity measurement of the method depends on tracer particles dispersed in a flow field, the instantaneous velocity distribution of the flow field is indirectly measured by measuring the displacement of the tracer particles within a known short time interval, and rich flow field space structures and flow characteristics can be provided. However, the size of the microspheres used in the liquid flow field analysis by the PIV technique is similar to the size of the sample detected by the flow cytometer, so that the flow characteristics of the laminar flow and the single cell flow in the flow chamber cannot be analyzed by using multiple tracer particles.

Disclosure of Invention

In order to solve the above problems, an object of the present invention is to provide a flow cytometer laminar stability evaluation method, which includes the steps of: 1) detecting 90-degree Mie scattered light of the microspheres in the flow chamber by using a high-speed microscopic image acquisition system; 2) the method comprises the following steps of performing cluster analysis on conditions of insufficient light intensity/trailing length, normality, diffraction, overlapping and the like in a large number of collected images by using a gray cluster analysis method to obtain a standard normal trailing image, wherein the method comprises the following specific steps: n observation objects are arranged, m evaluation indexes and s different gray classes are arranged, each observation object has m characteristic data to be observed, and the obtained sequence is as shown in formula (1):

X₁＝(x₁(1),x₁(2),…,x₁(n))

X₂＝(x₂(1),x₂(2),…,x₂(n))

……

X_m＝(x_m(1),x_m(2),…,x_m(n)) (1)

determining the center point λ of the gray class 1,2, …, s₁,λ₂,...,λ_sOf each indexThe value range is correspondingly divided into s gray classes; extending the ash class in different directions, considering adding 0 ash class and s +1 ash class, and determining the central point lambda of the ash class₀And λ_s+1Thus, a new sequence of centroids is obtained: lambda [ alpha ]₀,λ₁,λ₂,...,λ_s,λ_s+1Connection point (lambda)_k1) and the central point (lambda) of the kth-1 small ash class_k-10), point of attachment (λ)_l1) and center point (lambda) of the l +1 th small ash class_l+10), get the trapezoidal whitening weight function of j index with respect to k gray classAn observed value x for an index j can be determined from

Calculating the membership degree of the gray class k (k is 1,2, … s)Calculating the comprehensive clustering coefficient of the object i (i is 1,2, …, n) about the gray class k

Wherein,k subclass whitening weight function for j index, η_jFor the weight of the index j in the comprehensive cluster,

byJudging that the object i belongs to a gray class k;

3) determining a trailing boundary by using a midpoint method, and calculating the corresponding microsphere flow rate; 4) and (3) representing the stability of the flow path system of the flow cytometer by using the stability of the flow velocity of the microspheres.

Preferably, the index for performing cluster analysis in step 2) is determined as follows:

summing gray values of pixel points of each line in four types of images, namely, an image with insufficient light intensity/trailing length, a normal image, a diffraction image and an overlapping image respectively to obtain a transverse gray sum curve; solving a first derivative of the transverse gray sum curve; setting positive and negative thresholds, counting the number of extreme points in the threshold range, and taking the number of effective extreme points as an index of cluster analysis;

respectively summing the gray values of pixel points in each row in the four types of images, namely, the images with insufficient light intensity/trailing length, normal images, diffraction images and overlapping images to obtain a longitudinal gray sum curve; solving a first derivative of the longitudinal gray scale sum curve; setting positive and negative thresholds, counting the times of the curve passing through the positive and negative thresholds, and taking the number of intersection points with the positive and negative thresholds as an index of cluster analysis.

Preferably, the microsphere flow rate in the step 3) is obtained by the formula v ═ l/t, where l is the microsphere tail length and t is the camera exposure time.

Preferably, the step 4) further comprises the formulaAnd calculating a standard deviation, representing the microsphere speed by using the average value of the trailing length, and evaluating the stability of the microsphere speed by using the standard deviation of the trailing length.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed.

Drawings

Further objects, features and advantages of the present invention will become apparent from the following description of embodiments of the invention, with reference to the accompanying drawings, in which:

FIG. 1 is a schematic diagram of a high-speed image acquisition microsphere velocity measurement;

FIG. 2 is a microsphere smear image;

fig. 3 is a lateral gray summation curve: FIG. 3(a) Normal Ash; FIG. 3(b) ash class insufficient in length; FIG. 3(c) overlapping gray classes;

fig. 4 is a transverse gray scale sum derivative curve: FIG. 4(a) Normal Ash class; FIG. 4(b) ash class insufficient in length; FIG. 4(c) overlapping gray classes;

fig. 5 vertical gray summation curve: FIG. 5(a) Normal Ash; FIG. 5(b) ash class insufficient in length; FIG. 5(c) overlapping gray classes;

FIG. 6 is a longitudinal gray scale sum derivative curve: FIG. 6(a) Normal Ash class; FIG. 6(b) ash class insufficient in length; FIG. 6(c) overlapping gray classes;

FIG. 7 is a graph of column element gray value rising edge;

fig. 8 is a graph of column element gray value falling edges.

Detailed Description

The objects and functions of the present invention and methods for accomplishing the same will be apparent by reference to the exemplary embodiments. However, the present invention is not limited to the exemplary embodiments disclosed below; it can be implemented in different forms. The nature of the description is merely to assist those skilled in the relevant art in a comprehensive understanding of the specific details of the invention.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the drawings, the same reference numerals denote the same or similar parts, or the same or similar steps.

The invention provides a method for evaluating the stability of laminar flow in a flow chamber of a flow cytometer, which is characterized in that a high-speed microscopic image acquisition system is used for detecting 90-degree Mie scattered light of microspheres in the flow chamber, and a gray cluster analysis method is used for carrying out cluster analysis on conditions of insufficient light intensity/trailing length, normality, diffraction, overlapping and the like in a large number of acquired images to obtain a standard normal trailing image. The trailing boundary was then determined using the midpoint method and the corresponding microsphere flow rate was calculated. And finally, representing the stability of the flow cytometry liquid path system by using the stability of the flow velocity of the microspheres.

The invention selects the 90-degree lateral scattered light as the observation object, can avoid the direct light interference of the excitation light source, and removes the background light source collected by the traditional microscopic image, thereby reducing the background light information and improving the contrast of the image. When the microspheres pass through the laser excitation area of the flow chamber, the trailing images of the microspheres can be collected by changing the exposure time of the high-speed camera, and then the flow velocity of the microspheres is obtained. A schematic diagram of a high-speed image acquisition microsphere velocity measurement method based on 90-degree Mie scattering is shown in fig. 1.

Since the flow cytometer can detect tens of thousands of cells per second, and the position of the microsphere in the flow cell at the moment of exposure of the high-speed camera is random, the acquired microsphere trailing image generally includes 4 cases of blank, normal, short length and overlapping, as shown in fig. 2.

The invention adopts a clustering analysis method based on a trapezoidal whitening weight function to classify microsphere trailing images, and the method comprises the following concrete implementation steps:

n observation objects, m evaluation indexes and s different gray classes are set. Then each observation object has m pieces of feature data to be observed, and the obtained sequence is as shown in formula (1):

X₁＝(x₁(1),x₁(2),…,x₁(n))

X₂＝(x₂(1),x₂(2),…,x₂(n))

… …

X_m＝(x_m(1),x_m(2),…,x_m(n)) (1)

determining the center point λ of the gray class 1,2, …, s₁,λ₂,...,λ_sThe value range of each index is also divided into s gray classes correspondingly. Extending the ash class in different directions, considering adding 0 ash class and s +1 ash class, and determining the central point lambda of the ash class₀And λ_s+1Thus, a new sequence of centroids is obtained: lambda [ alpha ]₀,λ₁,λ₂,...,λ_s,λ_s+1Connection point (lambda)_k1) and the central point (lambda) of the kth-1 small ash class_k-10), point of attachment (λ)_l1) and center point (lambda) of the l +1 th small ash class_l+10), get the trapezoidal whitening weight function of j index with respect to k gray classAn observed value x for an index j can be determined from

Calculating the membership degree of the gray class k (k is 1,2, … s)Calculating the comprehensive clustering coefficient of the object i (i is 1,2, …, n) about the gray class k

Wherein,k subclass whitening weight function for j index, η_jIs the weight of the index j in the comprehensive cluster.

ByAnd judging that the object i belongs to the gray class k.

In one embodiment of the invention, a laser diode with the wavelength of 605nm is selected as the exciting light, and the power is 80 mW; the microscope objective is SPAHL-50 of Japan Sigma Koki company, the numerical aperture is 0.42, the magnification is 50, and the working distance is 20.5 mm; the image acquisition system uses a Q450 high-speed image acquisition system of Dantec company, a matched high-speed CMOS camera is V310 of VisionResearch company, the maximum resolution is 1280 multiplied by 800, the highest speed is 50 ten thousand fps, the minimum shutter time is 1us, and the geometric dimension of a single pixel point is 20 mu m multiplied by 20 mu m; the detection sample was prepared using a standard quality control microsphere Flow-Check Pro Fluoro-spheres A69183 from Beckman Coulter, Inc., the microsphere diameter being 20. + -. 1 μm. The exposure time was set to 40 μm, and the sample loading operation was performed using a flow cytometer. And after the laminar flow state is stable, image acquisition is started, the sampling frame rate is set to 3300 frames/s, and the total number of sampling pictures is set to 12000.

The gray values of the pixel points in each row in the 4 types of images are summed respectively to obtain a transverse gray sum curve as shown in fig. 3. The gray value threshold value is set to be 35, and the number of rows above the threshold value is counted to be used as an index of cluster analysis. The insufficient gray class has a gray value gradual change process in the longitudinal direction, and the rising edge in the transverse gray sum curve is slowly changed; normal gray does not have a longitudinal gray level gradual change process, so that the change of a rising edge and a falling edge is faster; the overlapped ash does not have a gradual change process in the longitudinal direction, but a plurality of trails are overlapped, so that the number of rising edges and falling edges of the overlapped ash is more than 1. The first derivative is solved for the curve in fig. 3, as shown in fig. 4. Setting the positive threshold value as 55 and the negative threshold value as-50, and counting the number of extreme points in the threshold value range. That is, when the amplitude of the extreme point is greater than 55 or less than-55, the statistics are taken as the valid extreme point. The numbers of the insufficient, normal and overlapped effective extreme points are respectively 1,2 and 4. The number of the effective extreme points can be used as an index of cluster analysis.

If the trailing length of the deficient gray class is seriously insufficient or the gray value is insufficient, the number of effective extreme points is 0; there are many possibilities for overlapping gray classes, and the number of trails, overlapping positions, overlapping manners, etc. of overlapping gray classes are uncertain, and the number of valid extreme points of overlapping gray classes may be other integers greater than 2.

Similarly, the gray values of each row of pixel points in the 4 types of images are summed respectively to obtain a longitudinal gray sum curve as shown in fig. 5. The gray value threshold is set to be 200, and the number of the rows above the threshold is counted to be used as an index of the cluster analysis. The insufficient gray class has a gray value gradual change process in the longitudinal direction, so that the peak value of the longitudinal gray sum is smaller; the probability of complete overlap of overlapping grays in the longitudinal direction is low, so the number of rows with gray values of 0 is greater than that in the normal case. The first derivative is solved for the curve in fig. 5, as shown in fig. 6. The positive threshold is set at 150 and the negative threshold is set at-250, and the number of times the curve passes through the positive and negative thresholds is counted. The trailing image gray distribution of normal gray is relatively uniform, so that the longitudinal gray sum first derivative curve has a monotone increasing and decreasing characteristic, no jitter exists near the positive and negative threshold values, and the number of the intersection points of the longitudinal gray sum first derivative curve and the positive and negative threshold values is 4. The trailing image gray distribution of the deficient and overlapped gray has gradual change or jump, the jitter exists near the positive and negative threshold values, and the number of fault points is not less than 4. The number of the intersection points with the positive and negative threshold values can be used as an index of cluster analysis.

The microsphere velocity is obtained by the formula v ═ l/t, where l is the microsphere tail length and t is the camera exposure time. In order to ensure accurate evaluation of the stability of the liquid path system, the rising and falling processes of the tailing gray value need to be analyzed, the tailing boundary is reasonably selected, and the calculation error of the tailing length of the microsphere is reduced.

And carrying out column summation on the gray values of the normal image, and symmetrically selecting 5 columns of pixel points by taking the maximum value of the gray value summation as a center. The process of raising the gray level of the selected 5 columns of pixels is shown in fig. 7. As can be seen from fig. 7, the gray values of the 5-column elements before the 303 th row have only slight jitter and remain substantially flat. From column 304 to column 309, the gray value rises rapidly and changes linearly. Similarly, the process of decreasing the gray-scale value of the 5 columns of pixels is shown in fig. 8. The gray value of the selected 5 columns of pixel points is kept stable before the 420 th column, and the gray value is rapidly reduced from the 421 st column to the 428 th column and is linearly changed.

Based on the characteristic of rapid linear change of gray values of pixel points in each row, the method selects a midpoint method to determine the trailing boundary. The midpoint method is to select a pixel point with the gray value closest to the average value of the edge change process as a trailing boundary. Taking fig. 7 and 8 as an example, the average gray values of the 5 columns of pixels in the ascending process are 36.5, 34, 33.1 and 35, which are the closest to the gray values (38, 37, 35, 34 and 35) of the pixels in the 307 th row, so that the 307 th row is used as the starting boundary point of the tailing; the average gray scale values of the 5 columns of pixels in the descending process are 33.2, 31.7, 31.2, 31.4 and 27.4, which are closest to the gray scale values (32, 30, 31 and 27) of the 307 row of pixels, so that the 424 th row is taken as the starting boundary point of the tailing.

Determining the trailing boundary of the normal image by using a midpoint method, calculating the average value of the trailing length to the number of pixel points to be 116.9, and calculating the average value according to a formulaThe standard deviation σ was calculated to be 1.7. Because the true value of the microsphere speed in the flow chamber cannot be obtained, the microsphere speed can be represented by using the average value of the trailing length, and the stability of the microsphere speed is evaluated by using the standard deviation of the trailing length, so that the evaluation of the laminar flow stability of the flow cytometer is completed.

Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.

Claims (4)

1. A flow cytometer laminar flow stability assessment method, the method comprising the steps of:

1) detecting 90-degree Mie scattered light of the microspheres in the flow chamber by using a high-speed microscopic image acquisition system;

2) the method comprises the following steps of performing cluster analysis on conditions of insufficient light intensity/trailing length, normality, diffraction, overlapping and the like in a large number of collected images by using a gray cluster analysis method to obtain a standard normal trailing image, wherein the method comprises the following specific steps:

n observation objects are arranged, m evaluation indexes and s different gray classes are arranged, each observation object has m characteristic data to be observed, and the obtained sequence is as shown in formula (1):

X₁＝(x₁(1),x₁(2),…,x₁(n))

X₂＝(x₂(1),x₂(2),…,x₂(n))

……

X_m＝(x_m(1),x_m(2),…,x_m(n)) (1)

determining the center point λ of the gray class 1,2, …, s₁,λ₂,...,λ_sCorrespondingly dividing the value range of each index into s gray classes; extending the ash class in different directions, considering adding 0 ash class and s +1 ash class, and determining the central point lambda of the ash class₀And λ_s+1Thus, a new sequence of centroids is obtained: lambda [ alpha ]₀,λ₁,λ₂,...,λ_s,λ_s+1Connection point (lambda)_k1) and the central point (lambda) of the kth-1 small ash class_k-10), point of attachment (λ)_l1) and center point (lambda) of the l +1 th small ash class_l+10), get the trapezoidal whitening weight function of j index with respect to k gray class(j-1, 2, …, m; k, l-1, 2, …, s), one observation x for index j may be determined from

$f_j^k\left(x\right)=\left\{\begin{array}{cc}0& x\∉\[{\mathrm{\λ}}_{k-1},{\mathrm{\λ}}_{l+1}\]\\ \frac{x-{\mathrm{\λ}}_{k-1}}{{\mathrm{\λ}}_k-{\mathrm{\λ}}_{k-1}}& x\∈({\mathrm{\λ}}_{k-1},{\mathrm{\λ}}_k\]\\ 1& x\∈\[{\mathrm{\λ}}_k,{\mathrm{\λ}}_l\]\\ \frac{{\mathrm{\λ}}_{l+1}-x}{{\mathrm{\λ}}_{l+1}-{\mathrm{\λ}}_l}& x\∈\left({\mathrm{\λ}}_l,{\mathrm{\λ}}_{l+1}\right)\end{array}\right.---\left(2\right)$

Calculating the membership degree of the gray class k (k is 1,2, … s)Calculating the comprehensive clustering coefficient of the object i (i is 1,2, …, n) about the gray class k

${\mathrm{\σ}}_i^k=\underset{j=1}{\overset{m}{\Σ}}{f}_j^k\left(x_{ij}\right)\·{\mathrm{\η}}_j---\left(3\right)$

Wherein,k subclass whitening weight function for j index, η_jFor the weight of the index j in the comprehensive cluster,

byJudging that the object i belongs to a gray class k;

3) determining a trailing boundary by using a midpoint method, and calculating the corresponding microsphere flow rate;

4) and (3) representing the stability of the flow path system of the flow cytometer by using the stability of the flow velocity of the microspheres.

2. The method of claim 1, wherein the index for performing cluster analysis in step 2) is determined as follows:

summing gray values of pixel points of each line in four types of images, namely, an image with insufficient light intensity/trailing length, a normal image, a diffraction image and an overlapping image respectively to obtain a transverse gray sum curve; solving a first derivative of the transverse gray sum curve; setting positive and negative thresholds, counting the number of extreme points in the threshold range, and taking the number of effective extreme points as an index of cluster analysis;

respectively summing the gray values of pixel points in each row in the four types of images, namely, the images with insufficient light intensity/trailing length, normal images, diffraction images and overlapping images to obtain a longitudinal gray sum curve; solving a first derivative of the longitudinal gray scale sum curve; setting positive and negative thresholds, counting the times of the curve passing through the positive and negative thresholds, and taking the number of intersection points with the positive and negative thresholds as an index of cluster analysis.

3. The method of claim 1, wherein the microsphere flow rate in step 3) is determined by the formula v-l/t, wherein_lThe length of the microsphere tail, and t the camera exposure time.

4. The method of claim 1, wherein the step 4) further comprises the step of calculating a formulaAnd calculating a standard deviation, representing the microsphere speed by using the average value of the trailing length, and evaluating the stability of the microsphere speed by using the standard deviation of the trailing length.