Machine learning based multi-indexed biological sample synchronous microfluidic analysis system

Abstract

The application discloses a biological sample multi-index synchronous microfluidic analysis system based on machine learning, relates to the technical field of biological sample index microfluidic analysis, and comprises a sample marking module, a microfluidic segmentation module, a light splitting range analysis module, a light splitting extraction module and a synchronous microfluidic analysis module; the sample marking module is used for marking biological samples through markers; the microfluidic segmentation module is used for segmenting microdroplets and shooting comprehensive microdroplet images; the light splitting extraction module is used for shooting marker microdroplet images after screening laser through a light filter; the light splitting range analysis module is used for dividing marker regions and background regions, then analyzing TD thresholds, and adjusting to obtain a marker adjustment image; the synchronous microfluidic analysis module is used for extracting the fluorescence intensity of the markers and calculating the concentrations of different indexes in the biological samples; the application is used for solving the problems that the existing biological sample index microfluidic analysis technology still excessively depends on manual adjustment and needs to be independently analyzed for different indexes, leading to a long analysis period and low precision.

Description

Technical Field

[0001] This invention relates to the field of microfluidic analysis technology for biological sample indicators, specifically a machine learning-based multi-indicator synchronous microfluidic analysis system for biological samples. Background Technology

[0002] Microfluidic analysis of biological samples is a cutting-edge interdisciplinary technology that uses microfluidic chips as a core platform to manipulate, react to, and detect one or more specific indicators in trace amounts of biological samples.

[0003] Existing microfluidic analysis techniques for biological samples typically require independent experiments and analyses for different indicators. Independent analysis consumes a large amount of sample and has a long analysis cycle. Furthermore, these techniques require significant manual adjustments, especially when adjusting the image threshold, necessitating repeated manual adjustments to accurately represent cell regions. Microfluidics often handles a large number of images, making manual adjustments slow and inaccurate. This not only increases processing time but also introduces significant errors in the indicator detection results. For example, Chinese patent application CN113466006A discloses a "microfluidic method for screening pathogenic microorganisms in medical samples," but this method is only suitable for screening a single indicator. If different indicators need to be screened, the entire process of sample processing, flow control, image acquisition, and analysis confirmation must be repeated, consuming a large amount of sample and resulting in a long analysis cycle. Therefore, existing microfluidic analysis techniques for biological samples suffer from over-reliance on manual adjustments and the need for independent analysis of different indicators, leading to long analysis cycles and low accuracy. Summary of the Invention

[0004] This invention aims to at least partially solve one of the technical problems in the prior art. It involves labeling biological samples with biomarkers, cutting mixed samples into different droplets, then exciting the biomarkers in the mixed sample with a laser to induce fluorescence and capturing a composite droplet image. The laser is then filtered before capturing labeled droplet images. The composite droplet image is analyzed to obtain a regional grayscale fluctuation function. Based on this function, biomarker and background regions are divided. The threshold of the composite droplet image is then analyzed based on the division results to obtain the TD threshold. The threshold of the biomarker droplet image is adjusted to the TD threshold to obtain a biomarker-adjusted image. Finally, the fluorescence intensity of different biomarkers is extracted from the biomarker-adjusted image, and the concentration of different indicators in the biological sample is calculated based on the fluorescence intensity. This addresses the problems of existing microfluidic analysis techniques for biological sample indicators, which rely too heavily on manual adjustment and require independent analysis of different indicators, resulting in long analysis cycles and low accuracy.

[0005] To achieve the above objectives, this application provides a machine learning-based multi-index synchronous microfluidic analysis system for biological samples, including a sample labeling module, a microfluidic segmentation module, a spectroscopic range analysis module, a spectroscopic extraction module, and a synchronous microfluidic analysis module; the sample labeling module, microfluidic segmentation module, spectroscopic range analysis module, and spectroscopic extraction module are respectively connected to the synchronous microfluidic analysis module for data transmission.

[0006] The sample labeling module is used to label biological samples using markers;

[0007] The microfluidic segmentation module is used to segment biological samples into different droplets, and based on microfluidic technology, different lasers are used to simultaneously irradiate the droplets and capture composite droplet images;

[0008] The spectral extraction module is used to capture images of the marked microdroplets after filtering the laser light through a filter;

[0009] The spectral range analysis module is used to analyze the composite droplet image, divide the marker region and background region, and then analyze the threshold of the composite droplet image based on the division results to obtain the TD threshold. Then, the threshold of the marker droplet image is adjusted to the TD threshold to obtain the marker adjusted image.

[0010] The synchronous microfluidic analysis module is used to extract the fluorescence intensity of different markers from the image by adjusting the markers and to calculate the concentration of different indicators in the biological sample based on the fluorescence intensity of the markers.

[0011] Furthermore, the sample labeling module is configured with a sample labeling strategy, which includes:

[0012] Biomarkers are prepared for indicators that need to be detected in biological samples, and each biomarker has a different excitation wavelength;

[0013] The biomarker is thoroughly mixed with the biological sample to obtain a mixed sample.

[0014] Furthermore, the microfluidic segmentation module includes a microfluidic cutting unit and a flag excitation unit;

[0015] The microfluidic cutting unit is used to cut the mixed sample into different droplets;

[0016] The marker excitation unit is used to excite the markers in the mixed sample with a laser, causing the markers to fluoresce and to capture a composite microdroplet image.

[0017] Furthermore, the microfluidic cutting unit is configured with a microfluidic cutting strategy, which includes:

[0018] At the microscale, the surface tension of the mixed sample is disrupted by the fluid shear force of the continuous phase, and the mixed sample is cut into different microdroplets.

[0019] Each droplet has the same volume.

[0020] Furthermore, the flag activation unit is configured with a flag activation strategy, the flag activation strategy including:

[0021] The markers are numbered using the symbol MR. n This indicates that n is a positive integer and n is the index of MR;

[0022] Will have MR n The laser excitation wavelength is labeled LR. n Use all LR n Simultaneously, the droplets are irradiated;

[0023] Images of the droplets were captured and named composite droplet images.

[0024] Furthermore, the spectral extraction module is configured with a spectral extraction strategy, which includes:

[0025] Starting with n=1, a filter is used to filter the laser, so that only LR is visible. n It can irradiate microdroplets through a filter, capture images of the microdroplets, and obtain the nth marker image;

[0026] By incrementing n and repeating the process, different nth label images are obtained. All nth label images are collectively referred to as label droplet images.

[0027] Furthermore, the spectral range analysis module includes a region division unit and a TD adjustment unit;

[0028] The region segmentation unit is used to analyze the comprehensive droplet image and segment it into a marker region and a background region;

[0029] The TD adjustment unit is used to analyze the threshold of the comprehensive droplet image based on the segmentation results, obtain the TD threshold, and then adjust the threshold of the marker droplet image to the TD threshold to obtain the marker adjusted image.

[0030] Furthermore, the region partitioning unit is configured with a flag partitioning strategy, which includes:

[0031] Acquire a composite droplet image, name the region where the fluorescent marker is located in the composite droplet image as the fluorescent region, and convert the composite droplet image into a grayscale image and name it a grayscale droplet image;

[0032] A template image is provided by the tester, which contains a composite microdroplet image of the fluorescent region circled by the tester. The template image is then converted into a grayscale image and named a grayscale template image.

[0033] Pixels belonging to the fluorescent region in the grayscale template image are named fluorescent template points, and the remaining pixels are named background template points. The grayscale value of the fluorescent template points is obtained and named the fluorescent template grayscale. The fluorescent regions are numbered using the symbol FR. i Let be an expression, where i is a positive integer and i is the index of FR, and FR is calculated. i The average gray value of the fluorescent template in the sample is denoted as HF. i ;

[0034] When analyzing any fluorescent template point, it is named a regional analysis point, and the fluorescence template gray value of the regional analysis point is marked as RAP. Fluorescent template points within eight neighborhoods of the regional analysis point are obtained and named neighborhood points, and the fluorescence template gray value of the neighborhood points is marked as NG. |RAP-NG| is calculated, and the calculation result is marked as VD. Each neighborhood point has a VD.

[0035] Statistics FR i VD in FR i All VDs in the set are grouped into a set named the fluctuation difference set, and represented by the symbol GA. i express;

[0036] Set the first gradient value, labeled Q. Divide the range from 0 to 255 into 256 / Q ranges, named the difference ranges. The span of each difference range is Q. Number the difference ranges in ascending order, using the symbol DR.j This indicates that j is a positive integer and j is the index of DR;

[0037] For each GA i Perform independent analysis and statistical analysis of GA. i In DR j The number of VDs, denoted as NR j , with DR j For the X-axis, NR j Create a histogram for the Y-axis and name it the Fluctuation Analysis Chart. Name the tallest histogram the baseline histogram and name the histogram that is to the right of the baseline histogram and furthest from it the fluctuation histogram.

[0038] Obtain the DR of the reference column j The median value is named the baseline value, and the DR of the fluctuation bars is obtained. j The maximum value is named the volatility value, and the base value and volatility value are represented by the symbols RV and RR, respectively.

[0039] Calculate RR / RV, name the result the volatility limit value, and set GA... i The fluctuation limit value is marked as FLV i ;

[0040] With HF i FLV is the horizontal axis. i Establish a two-dimensional coordinate system for the vertical axis, named the Regional Gray-Level Fluctuation Relationship Map, and convert the FLV... i According to HF i Enter the regional grayscale fluctuation relationship map, perform function regression analysis on the regional grayscale fluctuation relationship map, and obtain the regional grayscale fluctuation relationship function;

[0041] Name the pixels in the grayscale droplet image as grayscale droplet points, name the grayscale value of the grayscale droplet points as grayscale droplet grayscale, find the grayscale droplet point with the largest grayscale grayscale and name it the target point;

[0042] Define a target analysis region and name the pixels in the target analysis region as target analysis points. Initially, only target points exist in the target analysis region. Calculate the average gray value of the target analysis points in the target analysis region and name it as the target average gray value. Substitute the target average gray value into the region gray value fluctuation relationship function and name the fluctuation limit value obtained by the solution as the target fluctuation limit. Analyze the benchmark value of the target analysis region based on the fluctuation analysis map and name it as the target benchmark. If the target benchmark is zero, then assign the target benchmark to the preset benchmark. Calculate the product of the target benchmark and the target fluctuation limit to obtain the same domain threshold.

[0043] Pixels that are within the eight-neighborhood of the target analysis point (excluding the target analysis point) are named as outward analysis points. The VD of the outward analysis points and the target analysis point is calculated, and the calculation result is marked as VFD. It is determined whether VFD is less than or equal to the same-domain threshold. If it is, the outward analysis points are included in the target analysis area. If not, the outward analysis points do not belong to the target analysis area. Whenever a new target analysis point is included in the target analysis area, the outward analysis points are updated and the analysis is performed again. If all outward analysis points do not belong to the target analysis area, the target analysis area is named as the marker area.

[0044] The target point is re-identified and the marker region is analyzed, but the pixels within the marker region are no longer included in the analysis. This results in different marker regions. If two marker regions are adjacent, the HF values ​​in both regions are set. i The smaller logo area is changed to the background area.

[0045] Furthermore, the TD adjustment unit is configured with a TD adjustment strategy, which includes:

[0046] Adjust the Threshold of the composite droplet image from smallest to largest, and name the non-black parts in the adjusted composite droplet image as the effective region;

[0047] The graphic composed of the marked areas is named the standard area of ​​the marks. A similarity analysis is performed on the effective area of ​​the marks and the standard area of ​​the marks. The effective area of ​​the marks with the highest similarity is named the accurate mark area. The Threshold value corresponding to the accurate mark area is marked as the TD threshold.

[0048] The threshold of all the marker droplet images is adjusted to the TD threshold to obtain the marker-adjusted image.

[0049] Furthermore, the synchronous microfluidic analysis module is configured with a synchronous microfluidic analysis strategy, which includes:

[0050] The fluorescence intensity of different markers was extracted by adjusting the image using the markers;

[0051] Obtain a standard curve, and convert the fluorescence intensity into concentration based on the standard curve to obtain the index concentration.

[0052] The beneficial effects of this invention are as follows: This invention uses markers to label biological samples, cuts mixed samples into different droplets, and then uses lasers to excite the markers in the mixed samples, causing the markers to fluoresce and capturing a composite droplet image. The laser light is then filtered through a filter before capturing another image of the labeled droplets. The composite droplet image is then analyzed to obtain a regional grayscale fluctuation function. The advantage lies in the ability to simultaneously analyze multiple indicators based on markers with different excitation wavelengths and filters. Furthermore, in analyzing the regional grayscale fluctuation function, it considers that when pixels in a region are grouped into the same region, the applicable judgment threshold varies based on the intensity of color. That is, the magnitude of the grayscale value affects the region division result. Therefore, the applicable judgment threshold is dynamically analyzed based on the grayscale value of each region. Simultaneously, a baseline value for each region is calculated. For regions with the same average grayscale value, the baseline value may be larger or smaller. Calculating the judgment threshold based on the baseline value allows for more precise judgment. This judgment threshold is the same-domain threshold in this invention, improving the accuracy and convenience of microfluidic analysis of biological sample indicators.

[0053] This invention divides the marker region and background region based on a regional grayscale fluctuation relationship function. Then, based on the division results, it analyzes the threshold of the microdroplet image and determines the TD threshold. The threshold of the marker microdroplet image is then adjusted to the TD threshold to obtain the marker-adjusted image. Finally, the fluorescence intensity of different markers is extracted from the marker-adjusted image, and the concentration of different indicators in the biological sample is calculated based on the fluorescence intensity of the markers. The advantage is that it can automatically, efficiently, and accurately adjust the threshold of all marker microdroplet images, avoiding the low efficiency and low precision of manual adjustment, and improving the accuracy and efficiency of microfluidic analysis of biological sample indicators. Attached Figure Description

[0054] Figure 1 This is a schematic diagram of the system of the present invention;

[0055] Figure 2 This is a schematic diagram of the template image of the present invention;

[0056] Figure 3 This is a schematic diagram of the wave analysis graph of the present invention;

[0057] Figure 4 This is a schematic diagram of the effective area of ​​the mark in this invention. Detailed Implementation

[0058] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0059] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0060] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0061] Example 1, please refer to Figure 1 As shown, this application provides a machine learning-based multi-index synchronous microfluidic analysis system for biological samples, including a sample labeling module, a microfluidic segmentation module, a spectrophotometric range analysis module, a spectrophotometric extraction module, and a synchronous microfluidic analysis module; the sample labeling module, microfluidic segmentation module, spectrophotometric range analysis module, and spectrophotometric extraction module are respectively connected to the synchronous microfluidic analysis module for data transmission.

[0062] The sample labeling module is used to label biological samples using markers;

[0063] The sample labeling module is configured with a sample labeling strategy, which includes:

[0064] Biomarkers are prepared for indicators that need to be detected in biological samples, and each biomarker has a different excitation wavelength;

[0065] The biomarker is thoroughly mixed with the biological sample to obtain a mixed sample;

[0066] In practical applications, the processing and mixing of markers are existing technologies, and will not be described in detail in this embodiment. When selecting markers, it is necessary to ensure that the excitation wavelength of the markers corresponding to each indicator is different. For example, if three indicators need to be measured, three markers with different excitation wavelengths should be selected and marked respectively. When using lasers for irradiation in the future, lasers of the corresponding wavelengths should also be used for irradiation.

[0067] The microfluidic segmentation module is used to segment biological samples into different droplets. Based on microfluidic technology, different lasers are used to simultaneously irradiate the droplets and capture composite droplet images. The microfluidic segmentation module includes a microfluidic cutting unit and a marker excitation unit.

[0068] The microfluidic cutting unit is used to cut mixed samples into different droplets;

[0069] The microfluidic dicing unit is configured with a microfluidic dicing strategy, which includes:

[0070] At the microscale, the surface tension of the mixed sample is disrupted by the fluid shear force of the continuous phase, and the mixed sample is cut into different microdroplets.

[0071] Each droplet has the same volume;

[0072] In practical applications, the basic technical principles of microfluidics are all existing technologies, and will not be specifically explained in this embodiment.

[0073] The marker excitation unit is used to excite markers in the mixed sample using a laser, causing the markers to fluoresce and to capture a composite microdroplet image;

[0074] The flag activation unit is configured with a flag activation strategy, which includes:

[0075] The markers are numbered using the symbol MR. n This indicates that n is a positive integer and n is the index of MR;

[0076] Will have MR n The laser excitation wavelength is labeled LR. n Use all LR n Simultaneously, the droplets are irradiated;

[0077] Images of the droplets were captured and named composite droplet images;

[0078] In practical applications, for example, if there are three different excitation wavelengths, then three laser beams need to be used simultaneously to irradiate the microdroplets, and the three laser beams are three different excitation wavelengths.

[0079] The spectrophotometer extraction module is used to capture images of the marked microdroplets after filtering the laser light through a filter.

[0080] The spectroscopic extraction module is configured with a spectroscopic extraction strategy, which includes:

[0081] Starting with n=1, a filter is used to filter the laser, so that only LR is visible. n It can irradiate microdroplets through a filter, capture images of the microdroplets, and obtain the nth marker image;

[0082] By incrementing n and repeating the process, different nth label images are obtained. All nth label images are collectively referred to as label droplet images.

[0083] In practical applications, filtering laser light with filters is an existing technical means, which will not be described in detail in this embodiment. The focus of this embodiment is on the subsequent analysis process of the integrated microdroplet image.

[0084] The spectroscopic range analysis module is used to analyze the composite droplet image, divide the marker region and background region, and then analyze the threshold of the composite droplet image based on the division results to obtain the TD threshold. Finally, the threshold of the marker droplet image is adjusted to the TD threshold to obtain the marker-adjusted image. The spectroscopic range analysis module includes a region division unit and a TD adjustment unit.

[0085] The region segmentation unit is used to analyze the comprehensive droplet image and divide it into marker regions and background regions;

[0086] The region division unit is configured with a marker division strategy, which includes:

[0087] Acquire a composite droplet image, name the region where the fluorescent marker is located in the composite droplet image as the fluorescent region, and convert the composite droplet image into a grayscale image and name it a grayscale droplet image;

[0088] Please see Figure 2 As shown, the template image is provided by the tester. The template image contains a composite microdroplet image of the fluorescent area circled by the tester. The template image is converted into a grayscale image and named grayscale template image.

[0089] In practical applications, the grayscale template image is obtained as follows: Figure 2 As shown, Figure 2 The gray area represents the fluorescent region, and the white outline of the fluorescent region is the contour drawn by the tester. In the initial analysis process, the tester only needs to analyze and calculate a single composite droplet image to provide the most accurate template image. Subsequent analyses can then be performed automatically without further manual intervention. Moreover, the template image only needs to be analyzed once; the analysis of the template image is essentially a calibration of the analysis model. After calibration, it can be used long-term, meaning the regional grayscale fluctuation relationship function can be used indefinitely without additional analysis. Figure 2 Only a portion of the grayscale template image is shown; the purpose is to zoom in to see the transition around the fluorescent areas. Figure 2 It can be clearly seen that there are some faint colors when the fluorescent area transitions to the background area. These parts do not belong to the fluorescent area, but they are very likely to interfere with the analysis results in the actual analysis process. Therefore, they need to be removed to find the most accurate fluorescent area, that is, the marker area.

[0090] Pixels belonging to the fluorescent region in the grayscale template image are named fluorescent template points, and the remaining pixels are named background template points. The grayscale value of the fluorescent template points is obtained and named the fluorescent template grayscale. The fluorescent regions are numbered using the symbol FR. i Let be an expression, where i is a positive integer and i is the index of FR, and FR is calculated. i The average gray value of the fluorescent template in the sample is denoted as HF. i ;

[0091] When analyzing any fluorescent template point, it is named a regional analysis point, and the fluorescence template gray value of the regional analysis point is marked as RAP. Fluorescent template points within eight neighborhoods of the regional analysis point are obtained and named neighborhood points, and the fluorescence template gray value of the neighborhood points is marked as NG. |RAP-NG| is calculated, and the calculation result is marked as VD. Each neighborhood point has a VD.

[0092] Statistics FR i VD in FR i All VDs in the set are grouped into a set named the fluctuation difference set, and represented by the symbol GA. i express;

[0093] In practical applications, Figure 2 The pixels located within the pure white outline are the fluorescent template pixels, and the remaining pixels are the background template pixels. Figure 2 There are 4 fluorescent regions in total, numbered as FR i , 1≤i≤4, with Figure 2 Taking the fluorescent region in the upper left corner as an example, its number is FR1. At this time, i=1. The eight-neighborhood is a pre-defined concept and will not be further explained in this embodiment. Calculating the difference between the gray values ​​of the neighboring points and the region analysis points is actually analyzing the fluctuation value of the gray values ​​between pixels in the same fluorescent region, which is VD. For example, the gray value of the region analysis point is 88, i.e., RAP=88, while the gray value of a certain neighboring point is 89, i.e., NG=89. Finally, VD=1 is calculated. A region analysis point has multiple neighboring points, and each neighboring point has a VD. All VDs in FR1 are counted and formed into a set to obtain GA1.

[0094] Set the first gradient value, labeled Q. Divide the range from 0 to 255 into 256 / Q ranges, named the difference ranges. The span of each difference range is Q. Number the difference ranges in ascending order, using the symbol DR. j This indicates that j is a positive integer and j is the index of DR;

[0095] In practical applications, a large amount of existing image data shows that the gray values ​​of multiple pixels of the same origin and color in most images still show slight differences. This is due to color value fluctuations during computer imaging to restore the realism of objects. The same applies to the transition parts around fluorescent areas. The difference in gray values ​​of multiple pixels of the same origin and color in most images is concentrated around 3. Therefore, in this embodiment, the first gradient value Q is set to 3, dividing 0 to 255 into 256 / Q ranges. In fact, starting from 0, every 3 gray values ​​form a difference range. It is not necessary to divide 256 / Q ranges because the difference in gray values ​​is very small in practical applications. In the analysis of histograms, if the height of a histogram is 0, it is not necessary to include it in the reference range, that is, a histogram with a height of 0 is meaningless. In this embodiment, the maximum value of VD is 11, so only 4 difference ranges need to be divided, namely [0,2], [3,5], [6,8] and [9,11], which are DR1 to DR4 respectively.

[0096] Please see Figure 3 As shown, for each GA i Perform independent analysis and statistical analysis of GA. i In DR j The number of VDs, denoted as NR j , with DR j For the X-axis, NR j Create a histogram for the Y-axis and name it the Fluctuation Analysis Chart. Name the tallest histogram the baseline histogram and name the histogram that is to the right of the baseline histogram and furthest from it the fluctuation histogram.

[0097] Obtain the DR of the reference column j The median value is named the baseline value, and the DR of the fluctuation bars is obtained. j The maximum value is named the volatility value, and the base value and volatility value are represented by the symbols RV and RR, respectively.

[0098] Calculate RR / RV, name the result the volatility limit value, and set GA... i The fluctuation limit value is marked as FLV i ;

[0099] With HF i FLV is the horizontal axis. i Establish a two-dimensional coordinate system for the vertical axis, named the Regional Gray-Level Fluctuation Relationship Map, and convert the FLV... i According to HF i Enter the regional grayscale fluctuation relationship map, perform function regression analysis on the regional grayscale fluctuation relationship map, and obtain the regional grayscale fluctuation relationship function;

[0100] In practice, the fluctuation analysis diagram is obtained by statistical analysis and construction, as shown below. Figure 3 As shown, Figure 3The horizontal axis is actually DR. j The sequence number j, where the highest histogram bar is the baseline bar, and the rightmost histogram bar is the fluctuation bar. The DR corresponding to the baseline value. j The range is [3,5], therefore the baseline value RV is obtained as 4, the fluctuation value RR is obtained as 11, and the fluctuation limit value FLV1 is calculated as 2.75. The FLV of all fluorescent regions is analyzed. i Then, a regional grayscale fluctuation relationship map is constructed, and finally, based on function regression analysis, the regional grayscale fluctuation relationship function is obtained as Y = -0.00009 × X. 2 +0.0238×X+1.491, where Y is the FLV. i X is HF i The baseline value represents the maximum VD value within the fluorescent region. The upper limit of grayscale fluctuation based on the baseline value is the fluctuation limit value. The regional grayscale fluctuation relationship function reflects the relationship between the average grayscale of the fluorescent region and the fluctuation limit value. Different average grayscale values ​​have different applicable fluctuation limit values, so it is necessary to analyze the relationship between them.

[0101] Name the pixels in the grayscale droplet image as grayscale droplet points, name the grayscale value of the grayscale droplet points as grayscale droplet grayscale, find the grayscale droplet point with the largest grayscale grayscale and name it the target point;

[0102] Define the target analysis region and name the pixels in the target analysis region as target analysis points. Initially, only target points exist in the target analysis region. Calculate the average gray value of the target analysis points in the target analysis region and name it as the target average gray value. Substitute the target average gray value into the region gray value fluctuation relationship function and name the fluctuation limit value obtained by the solution as the target fluctuation limit. Analyze the benchmark value of the target analysis region based on the fluctuation analysis map and name it as the target benchmark. If the target benchmark is zero, then assign the target benchmark to the preset benchmark. Calculate the product of the target benchmark and the target fluctuation limit to obtain the same domain threshold.

[0103] In practical applications, in grayscale droplet images, the marker is significantly brighter than the background color. Therefore, using the grayscale droplet with the highest grayscale value as the target point for analysis ensures that the selected area is either the brightest part of the marker region or the background region. Initially, there is only one pixel in the target analysis area, so the result when analyzing the target baseline value will be 0. If the target baseline value is 0, a valid intra-domain threshold cannot be calculated. Therefore, a value needs to be assigned to the target baseline value that is equal to 0. Usually, it is assigned the value Q, that is, if the target baseline value is equal to 0, it is set to 3. The average grayscale value is calculated to be 74. Substituting this into the calculation, the target fluctuation limit is found to be 2.75936. Further calculation yields an intra-domain threshold of 8.27808. Using the rounding up method to retain the integer, the final intra-domain threshold is 9.

[0104] Pixels that are within the eight-neighborhood of the target analysis point (excluding the target analysis point) are named as outward analysis points. The VD of the outward analysis points and the target analysis point is calculated, and the calculation result is marked as VFD. It is determined whether VFD is less than or equal to the same-domain threshold. If it is, the outward analysis points are included in the target analysis area. If not, the outward analysis points do not belong to the target analysis area. Whenever a new target analysis point is included in the target analysis area, the outward analysis points are updated and the analysis is performed again. If all outward analysis points do not belong to the target analysis area, the target analysis area is named as the marker area.

[0105] The target point is re-identified and the marker region is analyzed, but the pixels within the marker region are no longer included in the analysis. This results in different marker regions. If two marker regions are adjacent, the HF values ​​in both regions are set. i The smaller logo area is changed to a background area;

[0106] In practical applications, if VFD is less than or equal to the same-domain threshold, it means that the extended analysis points and the target analysis points are from the same source and have similar colors, belonging to the same region. Therefore, the extended analysis points with VFD less than or equal to the same-domain threshold are included in the target analysis area, resulting in new target analysis points. After obtaining new target analysis points, the target analysis area expands. At this time, the target analysis area will come into contact with new extended analysis points, so analysis needs to be performed again until all extended analysis points no longer belong to the target analysis area, finally obtaining a marker region. Repeat the analysis of the marker region to obtain different numbers of marker regions. Usually, if two marker regions are adjacent, it means that one marker region is the part that transitions from the other marker region to the background region, and it also belongs to the background region, so it needs to be removed. The gray value of this part is usually smaller than that of the marker region, so the HF value of the marker region is removed. i The smaller logo area is changed into the background area, resulting in a logo area and a background area.

[0107] The TD adjustment unit is used to analyze the threshold of the comprehensive droplet image based on the segmentation results and obtain the TD threshold. Then, the threshold of the marker droplet image is adjusted to the TD threshold to obtain the marker adjusted image.

[0108] The TD adjustment unit is configured with a TD adjustment strategy, which includes:

[0109] Adjust the Threshold of the composite droplet image from smallest to largest, and name the non-black parts in the adjusted composite droplet image as the effective region;

[0110] Please see Figure 4As shown, the graphic composed of the marked areas is named the standard area of ​​the marks. A similarity analysis is performed on the effective area of ​​the marks and the standard area of ​​the marks. The effective area of ​​the marks with the highest similarity is named the accurate mark area. The Threshold value corresponding to the accurate mark area is marked as the TD threshold.

[0111] The threshold of all labeled droplet images is adjusted to the TD threshold to obtain the labeled adjusted image;

[0112] In practical applications, Threshold is a built-in parameter of the image. It needs to be adjusted in fluorescence intensity analysis to obtain accurate imaging of the marker. By adjusting Threshold, the effective area of ​​the marker is made most similar to the standard area of ​​the marker; this effective area is the region that most closely resembles the actual fluorescent region of the marker. Figure 4 As shown, Figure 4 The black area represents the background color, and the gray area represents the effective area of ​​the marker. The Threshold adjustment result is optimal when the effective area of ​​the marker is most similar to the standard area of ​​the marker. This yields the TD threshold. Then, the Threshold of all marker droplet images is adjusted to the TD threshold to obtain the marker-adjusted image, thereby obtaining the fluorescence intensity of different markers.

[0113] The synchronous microfluidic analysis module is used to extract the fluorescence intensity of different markers by adjusting the image and to calculate the concentration of different indicators in biological samples based on the fluorescence intensity of the markers;

[0114] The synchronous microfluidic analysis module is configured with synchronous microfluidic analysis strategies, which include:

[0115] The fluorescence intensity of different markers was extracted by adjusting the image using the markers;

[0116] Obtain a standard curve, and convert the fluorescence intensity into concentration based on the standard curve to obtain the index concentration;

[0117] In practical applications, extracting fluorescence intensity from images and converting fluorescence intensity into concentration based on standard curves are existing fluorescence intensity processing techniques, which will not be specifically described in this embodiment. The focus of this embodiment is on the intelligent and precise adjustment of Threshold.

[0118] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0119] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings or direct couplings or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

Claims

1. A microfluidic analysis system for multiple indicators of biological samples simultaneously, characterized in that, It includes a sample labeling module, a microfluidic segmentation module, a spectroscopic range analysis module, a spectroscopic extraction module, and a synchronous microfluidic analysis module; the sample labeling module, microfluidic segmentation module, spectroscopic range analysis module, and spectroscopic extraction module are respectively connected to the synchronous microfluidic analysis module for data transmission. The sample labeling module is used to label biological samples using markers; The microfluidic segmentation module is used to segment biological samples into different droplets, and based on microfluidic technology, different lasers are used to simultaneously irradiate the droplets and capture composite droplet images; The spectral extraction module is used to capture images of the marked microdroplets after filtering the laser light through a filter; The spectral range analysis module is used to analyze the composite droplet image, divide the marker region and background region, and then analyze the threshold of the composite droplet image based on the division results to obtain the TD threshold. Then, the threshold of the marker droplet image is adjusted to the TD threshold to obtain the marker adjusted image. The synchronous microfluidic analysis module is used to extract the fluorescence intensity of different markers from the image by adjusting the markers and to calculate the concentration of different indicators in the biological sample based on the fluorescence intensity of the markers.

2. The multi-index simultaneous microfluidic analysis system for biological samples according to claim 1, characterized in that, The sample labeling module is configured with a sample labeling strategy, which includes: Biomarkers are prepared for indicators that need to be detected in biological samples, and each biomarker has a different excitation wavelength; The biomarker is thoroughly mixed with the biological sample to obtain a mixed sample.

3. The multi-index simultaneous microfluidic analysis system for biological samples according to claim 2, characterized in that, The microfluidic segmentation module includes a microfluidic cutting unit and a marker excitation unit; The microfluidic cutting unit is used to cut the mixed sample into different droplets; The marker excitation unit is used to excite the markers in the mixed sample with a laser, causing the markers to fluoresce and to capture a composite microdroplet image.

4. The multi-index simultaneous microfluidic analysis system for biological samples according to claim 3, characterized in that, The microfluidic cutting unit is configured with a microfluidic cutting strategy, which includes: At the microscale, the surface tension of the mixed sample is disrupted by the fluid shear force of the continuous phase, and the mixed sample is cut into different microdroplets. Each droplet has the same volume.

5. The multi-index simultaneous microfluidic analysis system for biological samples according to claim 4, characterized in that, The flag activation unit is configured with a flag activation strategy, which includes: The markers are numbered using the symbol MR. n This indicates that n is a positive integer and n is the index of MR; Will have MR n The laser excitation wavelength is labeled LR. n Use all LR n Simultaneously, the droplets are irradiated; Images of the droplets were captured and named composite droplet images.

6. The multi-index simultaneous microfluidic analysis system for biological samples according to claim 5, characterized in that, The spectroscopic extraction module is configured with a spectroscopic extraction strategy, which includes: Starting with n=1, a filter is used to filter the laser, so that only LR... n It can irradiate microdroplets through a filter, capture images of the microdroplets, and obtain the nth marker image; By incrementing n and repeating the process, different nth label images are obtained. All nth label images are collectively referred to as label droplet images.

7. The multi-index simultaneous microfluidic analysis system for biological samples according to claim 6, characterized in that, The spectral range analysis module includes a region division unit and a TD adjustment unit; The region segmentation unit is used to analyze the comprehensive droplet image and segment it into a marker region and a background region; The TD adjustment unit is used to analyze the threshold of the comprehensive droplet image based on the segmentation results, obtain the TD threshold, and then adjust the threshold of the marker droplet image to the TD threshold to obtain the marker adjusted image.

8. The simultaneous microfluidic analysis system for multiple indicators of biological samples according to claim 7, characterized in that, The region partitioning unit is configured with a flag partitioning strategy, which includes: Acquire a composite droplet image, name the region where the fluorescent marker is located in the composite droplet image as the fluorescent region, and convert the composite droplet image into a grayscale image and name it a grayscale droplet image; A template image is provided by the tester, which contains a composite microdroplet image of the fluorescent region circled by the tester. The template image is then converted into a grayscale image and named a grayscale template image. Pixels belonging to the fluorescent region in the grayscale template image are named fluorescent template points, and the remaining pixels are named background template points. The grayscale value of the fluorescent template points is obtained and named the fluorescent template grayscale. The fluorescent regions are numbered using the symbol FR. i Let be an expression, where i is a positive integer and i is the index of FR, and FR is calculated. i The average gray value of the fluorescent template in the sample is denoted as HF. i ; When analyzing any fluorescent template point, it is named a regional analysis point, and the fluorescence template gray value of the regional analysis point is marked as RAP. Fluorescent template points within eight neighborhoods of the regional analysis point are obtained and named neighborhood points, and the fluorescence template gray value of the neighborhood points is marked as NG. |RAP-NG| is calculated, and the calculation result is marked as VD. Each neighborhood point has a VD. Statistics FR i VD in FR i All VDs in the set are grouped into a set named the fluctuation difference set, and represented by the symbol GA. i express; Set the first gradient value, labeled Q. Divide the range from 0 to 255 into 256 / Q ranges, named the difference ranges. The span of each difference range is Q. Number the difference ranges in ascending order, using the symbol DR. j This indicates that j is a positive integer and j is the index of DR; For each GA i Perform independent analysis and statistical analysis of GA. i In DR j The number of VDs, denoted as NR j , with DR j For the X-axis, NR j Create a histogram for the Y-axis and name it the Fluctuation Analysis Chart. Name the tallest histogram the baseline histogram and name the histogram that is to the right of the baseline histogram and furthest from it the fluctuation histogram. Obtain the DR of the reference column j The median value is named the baseline value, and the DR of the fluctuation bars is obtained. j The maximum value is named the volatility value, and the base value and volatility value are represented by the symbols RV and RR, respectively. Calculate RR / RV, name the result the volatility limit value, and set GA... i The fluctuation limit value is marked as FLV i ; With HF i FLV is the horizontal axis. i Establish a two-dimensional coordinate system for the vertical axis, named the Regional Gray-Level Fluctuation Relationship Map, and convert the FLV... i According to HF i Enter the regional grayscale fluctuation relationship map, perform function regression analysis on the regional grayscale fluctuation relationship map, and obtain the regional grayscale fluctuation relationship function; Name the pixels in the grayscale droplet image as grayscale droplet points, name the grayscale value of the grayscale droplet points as grayscale droplet grayscale, find the grayscale droplet point with the largest grayscale grayscale and name it the target point; Define a target analysis region and name the pixels in the target analysis region as target analysis points. Initially, only target points exist in the target analysis region. Calculate the average gray value of the target analysis points in the target analysis region and name it as the target average gray value. Substitute the target average gray value into the region gray value fluctuation relationship function and name the fluctuation limit value obtained by the solution as the target fluctuation limit. Analyze the benchmark value of the target analysis region based on the fluctuation analysis map and name it as the target benchmark. If the target benchmark is zero, then assign the target benchmark to the preset benchmark. Calculate the product of the target benchmark and the target fluctuation limit to obtain the same domain threshold. Pixels that are within the eight-neighborhood of the target analysis point (excluding the target analysis point) are named as outward analysis points. The VD of the outward analysis points and the target analysis point is calculated, and the calculation result is marked as VFD. It is determined whether VFD is less than or equal to the same-domain threshold. If it is, the outward analysis points are included in the target analysis area. If not, the outward analysis points do not belong to the target analysis area. Whenever a new target analysis point is included in the target analysis area, the outward analysis points are updated and the analysis is performed again. If all outward analysis points do not belong to the target analysis area, the target analysis area is named as the marker area. The target point is re-identified and the marker region is analyzed, but the pixels within the marker region are no longer included in the analysis. This results in different marker regions. If two marker regions are adjacent, the HF values ​​in both regions are set. i The smaller logo area is changed to the background area.

9. The multi-index simultaneous microfluidic analysis system for biological samples according to claim 8, characterized in that, The TD adjustment unit is configured with a TD adjustment strategy, which includes: Adjust the Threshold of the composite droplet image from smallest to largest, and name the non-black parts in the adjusted composite droplet image as the effective region; The graphic composed of the marked areas is named the standard area of ​​the marks. A similarity analysis is performed on the effective area of ​​the marks and the standard area of ​​the marks. The effective area of ​​the marks with the highest similarity is named the accurate mark area. The Threshold value corresponding to the accurate mark area is marked as the TD threshold. The threshold of all the marker droplet images is adjusted to the TD threshold to obtain the marker-adjusted image.

10. The multi-index simultaneous microfluidic analysis system for biological samples according to claim 9, characterized in that, The synchronous microfluidic analysis module is configured with a synchronous microfluidic analysis strategy, which includes: The fluorescence intensity of different markers was extracted by adjusting the image using the markers; Obtain a standard curve, and convert the fluorescence intensity into concentration based on the standard curve to obtain the index concentration.

Images