Platelet classification method and system based on platelet alpha granule super-resolution image

Abstract

The application discloses a platelet classification method and system based on a platelet alpha granule super-resolution image, the classification method classifies the platelets according to the distribution form of the alpha granules, and specifically classifies into three categories of "regular distribution", "aggregated distribution" and "scattered distribution", the platelets are classified according to the distribution form of the alpha granules in a single platelet, and different samples can be distinguished. The platelet classification system based on the platelet alpha granule super-resolution image provided by the application acquires the platelet alpha granule super-resolution image based on a super-resolution microscopic imaging technology, extracts a single platelet alpha granule super-resolution image in the platelet alpha granule super-resolution image, classifies the single platelet alpha granule according to the distribution characteristics of the single platelet alpha granule, and has important significance for the research and detection of the physiological function of the platelets.

Description

Technical Field

[0001] This invention belongs to the field of biotechnology, and more specifically, relates to a platelet classification method and system based on super-resolution images of platelet α-particles. Background Technology

[0002] Platelets play a vital role in the human body, promoting hemostasis, accelerating blood clotting, and repairing blood vessels. They also function crucially in physiological and pathological processes such as immune defense, tumor growth and metastasis, and inflammation regulation. Platelets have a complex structure and composition. Although they lack a nucleus, they contain scattered granular components. Upon contact with a wound or non-vascular endothelial surface such as glass, platelets rapidly expand, with granules converging towards the center and extending multiple pseudopodia, transforming into dendritic platelets. Most granules are then released, and platelets fuse together, becoming sticky, deformable platelets. Dendritic platelets can revert to circulating platelets if the irritant is promptly eliminated, while the change in sticky, deformable platelets is irreversible.

[0003] Platelet microparticles are ultramicroscopic membrane vesicles released by platelets during activation. Their diameter is less than 0.5 μm, and their fine morphology cannot be observed under a regular microscope, making them undetectable by conventional platelet counting methods, including platelet counters. Two types of granules are scattered within platelets: alpha granules and dense granules. Alpha granules contain medium electron density content, and some granules have a core with higher electron density. Alpha granules contain fibrinogen, platelet factor 4, cathepsin A, cathepsin D, and acid hydrolases, among other things.

[0004] The size of alpha particles within platelets is on the nanometer scale. Traditional microscopes, limited by the diffraction limit, cannot perform precise detection of alpha particles within a single platelet, thus failing to directly provide information about the alpha particles of a single platelet. Currently, the commonly used methods for detecting changes in alpha particles in clinical practice are ELISA and flow cytometry. These two methods indirectly detect changes in alpha particles by measuring the amount of alpha particle protein released from a large number of platelets, and cannot directly reflect changes in alpha particles, which may indicate changes in platelet physiology or function.

[0005] Furthermore, while existing technologies can detect changes in individual platelet alpha particles using electron microscopy, such as detecting morphological parameters like the number, size, and diameter of alpha particles within a single platelet, electron microscopy cannot specifically provide the distribution characteristics of alpha particles within an individual platelet. This makes it difficult to classify platelets based on different alpha particle distributions. Additionally, electron microscopy suffers from drawbacks such as operational complexity, small imaging range, low efficiency, and long processing time, making it difficult to meet the needs of rapid testing of large numbers of blood samples. Moreover, the information obtained from detecting individual platelets in clinical practice is still relatively limited. Therefore, researching how to classify individual platelets is of great significance for the study and detection of platelet physiological functions. Summary of the Invention

[0006] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a platelet classification method and system based on super-resolution images of platelet alpha particles. The aim is to extract the imaging features of individual platelet alpha particles from the obtained super-resolution images using super-resolution microscopy. Statistical analysis of these imaging features reveals that platelet alpha particles can be classified into multiple categories based on their distribution characteristics. These categories have a clear correlation with the physiological functions of platelets and even blood function, thus possessing biological significance.

[0007] Individual platelets are classified according to different distribution patterns of α particles, enriching the information obtained from platelet and α particle detection. In particular, they are classified into three categories according to the "regular distribution", "aggregated distribution" and "scattered distribution" of α particles within an individual platelet. This solves the technical problem that existing technologies lack classification of individual platelets and currently provide limited information from platelet α particle detection.

[0008] To achieve the above objectives, according to one aspect of the present invention, a platelet classification method based on platelet α-particle super-resolution images is provided, comprising the following steps:

[0009] Obtain super-resolution images of immunofluorescently labeled α-particles of a single platelet and binarize them to obtain an image of intraplatelet α-particles composed of fluorescent and non-fluorescent pixels.

[0010] For the image of α particles within a single platelet, an image segmentation algorithm is used to segment the fluorescent signal pixel region into pixel regions of independent α particles to obtain the distribution image of α particles in a single platelet.

[0011] The platelets were classified according to the distribution pattern of α-particles, as follows:

[0012] (1) Search for the ellipse with the smallest area that satisfies condition A within the α-particle image of a single platelet, and use it as the outer circle: if the area of ​​the outer circle is above a preset regular distribution ellipse area threshold and the ratio of the minor axis to the major axis is within the preset regular distribution ellipse range, then the platelets involved in the single platelet α-particle distribution image are determined to be "regularly distributed"; the regular distribution ellipse area threshold is within 0.15 μm. 2 -20.0μm 2 The range of the regularly distributed ellipse is [0.2, 1].

[0013] (2) Search for the ellipse with the smallest area within the α-particle image of a single platelet that satisfies condition A, and use it as the outer circle: if the area of ​​the outer circle is below the area threshold of the aggregation distribution ellipse, then the platelets involved in the single platelet α-particle distribution image are determined to be "aggregated"; the area threshold of the aggregation distribution ellipse is within 0.02 μm. 2 -7.0μm 2 Between; the area threshold of the regularly distributed ellipse is greater than or equal to the area threshold of the clustered distributed ellipse;

[0014] (3) If the α-particle image of a single platelet does not belong to "regular distribution" and "aggregated distribution", then the platelets involved in the single platelet α-particle distribution image are judged to be "scattered distribution".

[0015] Condition A: α particles exceeding a preset outer ring proportion threshold are distributed inside the ellipse; the preset outer ring proportion threshold is above 95%.

[0016] Preferably, the platelet classification method based on platelet α-particle super-resolution images classifies platelets into subclasses according to the distribution pattern of α-particles within a single platelet that exhibits a "regular distribution," as follows:

[0017] The ellipse with the largest area satisfying condition B within the outer ring is selected as the inner ring; if the ratio of the area of ​​the inner ring to that of the outer ring exceeds a preset annular distribution threshold, then the platelets involved in the single platelet α-particle distribution image are determined to be in the "annular distribution" subcategory; otherwise, the platelets involved in the single platelet α-particle distribution image are determined to be in the "particle distribution" subcategory; the annular distribution threshold ranges from [0.5, 1).

[0018] Condition B: α particles exceeding the preset inner circle proportion threshold are distributed outside the ellipse; the preset inner circle proportion threshold is above 95%.

[0019] Preferably, the platelet classification method based on platelet α-particle super-resolution images uses the number of fluorescent pixel blocks as the number N of α-particles for a "scattered" distribution image of individual platelet α-particles.

[0020] Based on the relationship between the number of α particles and a preset quantity threshold, the images of α particles within a single platelet are classified into subcategories for "particle distribution".

[0021] Preferably, in the platelet classification method based on platelet α-particle super-resolution images, the preset quantity threshold is between 2 and 70.

[0022] Preferably, in the platelet classification method based on platelet α-particle super-resolution images, the area threshold of the regularly distributed ellipse is 3.0 μm.2 -13.0μm 2 The range of the regular distribution ellipse is [0.5, 1]; the area threshold of the clustered distribution ellipse is 2.5 μm. 2 -4.5μm 2 between.

[0023] According to another aspect of the present invention, a platelet classification system based on platelet α-particle super-resolution images is provided, including a platelet α-particle super-resolution image acquisition module, a single platelet α-particle distribution image extraction and annotation module, and a classification module;

[0024] The platelet α-particle super-resolution image acquisition module is used to perform binarization to obtain an image of a single platelet α-particle composed of fluorescent signal pixels and non-fluorescent signal pixels; and submit it to the single platelet α-particle distribution image extraction and annotation module.

[0025] The single platelet α-particle distribution image extraction and annotation module is used to segment the fluorescent signal pixel region into the pixel region of independent α-particles in the single platelet α-particle image using an image segmentation algorithm, obtain the single platelet α-particle distribution image, and submit it to the classification module.

[0026] The classification module is used to classify the platelets according to the distribution pattern of α particles, as follows:

[0027] (1) Search for the ellipse with the smallest area that satisfies condition A within the α-particle image of a single platelet, and use it as the outer circle: if the area of ​​the outer circle is above a preset regular distribution ellipse area threshold and the ratio of the minor axis to the major axis is within the preset regular distribution ellipse range, then the platelets involved in the single platelet α-particle distribution image are determined to be "regularly distributed"; the regular distribution ellipse area threshold is within 0.15 μm. 2 -20.0μm 2 The range of the regularly distributed ellipse is [0.2, 1].

[0028] (2) Search for the ellipse with the smallest area within the α-particle image of a single platelet that satisfies condition A, and use it as the outer circle: if the area of ​​the outer circle is below the area threshold of the aggregation distribution ellipse, then the platelets involved in the single platelet α-particle distribution image are determined to be "aggregated"; the area threshold of the aggregation distribution ellipse is within 0.02 μm. 2 -7.0μm 2 Between; the area threshold of the regularly distributed ellipse is greater than or equal to the area threshold of the clustered distributed ellipse;

[0029] (3) If the α-particle image of a single platelet does not belong to "regular distribution" and "aggregated distribution", then the platelets involved in the single platelet α-particle distribution image are judged to be "scattered distribution".

[0030] Condition A: α particles exceeding a preset outer ring proportion threshold are distributed inside the ellipse; the preset outer ring proportion threshold is above 95%.

[0031] Preferably, the platelet classification system based on platelet α-particle super-resolution images further includes a regular distribution sub-classification module; used to classify platelets according to the distribution pattern of α-particles in a single platelet with a "regular distribution", as follows:

[0032] The ellipse with the largest area satisfying condition B within the outer ring is selected as the inner ring; if the ratio of the area of ​​the inner ring to that of the outer ring exceeds a preset annular distribution threshold, then the platelets involved in the single platelet α-particle distribution image are determined to be in the "annular distribution" subcategory; otherwise, the platelets involved in the single platelet α-particle distribution image are determined to be in the "particle distribution" subcategory; the annular distribution threshold ranges from [0.5, 1).

[0033] Condition B: α particles exceeding the preset inner circle proportion threshold are distributed outside the ellipse; the preset inner circle proportion threshold is above 95%.

[0034] Preferably, the platelet classification system based on platelet α-particle super-resolution images further includes a scattered distribution sub-classification module; used to count the number of fluorescent pixel blocks as the number N of α-particles for a single platelet α-particle distribution image that is "scattered".

[0035] Based on the relationship between the number of α particles and a preset quantity threshold, the images of α particles within a single platelet are classified into subcategories for "particle distribution".

[0036] Preferably, in the platelet classification system based on platelet α-particle super-resolution images, the preset quantity threshold is between 2 and 70.

[0037] Preferably, in the platelet classification system based on platelet α-particle super-resolution images, the area threshold of the regularly distributed ellipse is 3.0 μm. 2 -13.0μm 2 The range of the regular distribution ellipse is [0.5, 1]; the area threshold of the clustered distribution ellipse is 2.5 μm. 2 -4.5μm 2 between.

[0038] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects:

[0039] The present invention provides a platelet classification method and system based on super-resolution images of platelet α particles. The method acquires super-resolution images of platelet α particles using super-resolution microscopy and classifies them according to the distribution characteristics of individual platelet α particles, which are divided into "regular distribution", "scattered distribution" or "aggregated distribution". This classification method can classify individual platelets and is of great significance for the study and detection of platelet physiological functions.

[0040] Furthermore, the "scattered distribution" is further subdivided into categories of "N < preset threshold" or "N ≥ preset threshold" based on the number of α particles within a single platelet. That is, single platelets are divided into four categories: "regular distribution", "aggregated distribution", "N < preset threshold" or "N ≥ preset threshold". The preset threshold for the number of α particles is set according to the actual situation of scientific or clinical research, so as to be able to distinguish the distribution of α particles within platelets.

[0041] In addition, the platelet classification system based on super-resolution images of platelet α particles provided by the present invention can automatically classify the platelets according to the distribution of α particles in the sample. Compared with existing methods for detecting α particle protein content, number, size and diameter, the present invention can provide richer information on the distribution location and quantity of α particles in platelets. Attached Figure Description

[0042] Figure 1 It is a super-resolution image of platelet alpha particles;

[0043] Figure 2 This is a diagram illustrating the classification of individual platelets;

[0044] Figure 3 This is a schematic diagram illustrating the process of establishing a platelet classification system based on super-resolution images of platelet alpha particles. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0046] Alpha granules, the most abundant secretory granules in platelets, generally exhibit different distribution patterns within the platelet. These different distribution patterns of alpha granules may lead to different functions among platelets. Classifying platelets into different categories based on the distribution pattern of alpha granules within the platelet facilitates researchers' in-depth study of platelet physiological functions. However, currently, changes in platelet alpha granules are mainly detected by ELISA and flow cytometry to detect alpha granule protein content, indirectly reflecting changes in alpha granules. Existing detection technologies cannot directly detect the distribution of alpha granules within a single platelet, and cannot reveal information such as platelet changes from changes in the distribution or content of alpha granules in a sample.

[0047] With the development of super-resolution microscopy, imaging and detection of biological samples at the nanoscale have been achieved. In particular, imaging techniques with a resolution below 100 nm can perform fine detection of a large number of platelet α particles. This experiment compares and analyzes super-resolution images of platelet α particles in a large number of samples and observes that α particles in individual platelets in different samples generally exhibit different distributions. For example, some α particles are distributed in rings within the platelet, and some of these rings have no α particles inside, while others have α particles distributed inside and / or outside. In some cases, multiple α particles aggregate together to form one or more clusters, and in others, different numbers of particles are distributed.

[0048] First, the platelet α particle super-resolution image is processed: using an image segmentation algorithm, the platelet α particle super-resolution image is cropped into a single platelet α particle immunofluorescence-labeled super-resolution image, and then binarized to obtain a single platelet α particle image composed of fluorescent signal pixels and non-fluorescent signal pixels.

[0049] For the image of α particles within a single platelet, an image segmentation algorithm is used to segment the fluorescent signal pixel region into pixel regions of independent α particles to obtain the distribution image of α particles in a single platelet.

[0050] Furthermore, it was found that classifying individual platelets according to the distribution pattern of their internal α-granules can distinguish different samples. The platelets were classified into three main categories based on the distribution pattern of their α-granules: "regular distribution," "aggregated distribution," and "scattered distribution."

[0051] "Regular distribution" refers to the distribution of alpha particles in platelets to form rings. These rings include those without alpha particles inside and those with alpha particles distributed inside and / or outside. Platelets with "regular distribution" of alpha particles can be further divided into platelets with alpha particles distributed in a ring or those with alpha particles distributed in a ring but with alpha particles scattered in a scattered manner inside. These are respectively called the "ring distribution" subcategory and the "particle distribution" subcategory.

[0052] "Aggregation distribution" refers to the aggregation of multiple α particles to form one or more clusters, where the size of a single cluster is less than 0.02 μm. 2 -7.0μm 2 ;

[0053] "Scattered distribution" refers to distributions other than "regular distribution" and "clustered distribution", and can be further divided into different subcategories based on the number of α particles using thresholds.

[0054] The obtained platelet categories and subcategories show significant differences in physiological behavior and other aspects. Classifying platelets according to the distribution of α-granules may help to understand the blood condition, thus facilitating research and physiological judgment.

[0055] Based on this discovery, the present invention provides a platelet classification method based on super-resolution images of platelet α-particles, which includes the following steps:

[0056] Obtain super-resolution images of immunofluorescently labeled α-particles of a single platelet and binarize them to obtain an image of intraplatelet α-particles composed of fluorescent and non-fluorescent pixels.

[0057] For the image of α particles within a single platelet, an image segmentation algorithm is used to segment the fluorescent signal pixel region into pixel regions of independent α particles to obtain the distribution image of α particles in a single platelet.

[0058] The platelets were classified according to the distribution pattern of α-particles, as follows:

[0059] (1) Search for the ellipse with the smallest area that satisfies condition A within the α-particle image of a single platelet, and use it as the outer circle: if the area of ​​the outer circle is above a preset regular distribution ellipse area threshold and the ratio of the minor axis to the major axis is within the preset regular distribution ellipse range, then the platelets involved in the single platelet α-particle distribution image are determined to be "regularly distributed"; the regular distribution ellipse area threshold is within 0.15 μm. 2 -20.0μm 2 Between, preferably 3.0μm 2 -13.0μm 2 The range of the regularly distributed ellipse is [0.2, 1], preferably [0.5, 1].

[0060] For images of α-particles within a single platelet with a "regular distribution," the platelets are subclassified according to the distribution pattern of the α-particles, as follows:

[0061] The ellipse with the largest area satisfying condition B within the outer ring is selected as the inner ring; if the ratio of the area of ​​the inner ring to that of the outer ring exceeds a preset annular distribution threshold, then the platelets involved in the single platelet α-particle distribution image are determined to be in the "annular distribution" subcategory; otherwise, the platelets involved in the single platelet α-particle distribution image are determined to be in the "particle distribution" subcategory; the annular distribution threshold ranges from [0.5, 1).

[0062] (2) Search for the ellipse with the smallest area within the α-particle image of a single platelet that satisfies condition A, and use it as the outer circle: if the area of ​​the outer circle is below the area threshold of the aggregation distribution ellipse, then the platelets involved in the single platelet α-particle distribution image are determined to be "aggregated"; the area threshold of the aggregation distribution ellipse is within 0.02 μm. 2 -7.0μm 2 Between, preferably 2.5μm 2 -4.5μm 2 Between; the area threshold of the regularly distributed ellipse is greater than or equal to the area threshold of the clustered distributed ellipse;

[0063] (3) If the α-particle image of a single platelet does not belong to "regular distribution" and "aggregated distribution", then the platelets involved in the single platelet α-particle distribution image are judged to be "scattered distribution".

[0064] For images of scattered individual platelet α-particles, the number of fluorescent pixel blocks is counted as the number N of α-particles;

[0065] The number of α particles is divided into subcategories based on the relationship between the number of α particles and a preset threshold. The preset threshold is set according to the actual situation of scientific or clinical research, preferably capable of distinguishing between experimental and control groups; in some embodiments, the preset threshold is taken between 2 and 70.

[0066] Condition A: α particles exceeding a preset outer ring proportion threshold are distributed inside the ellipse; the preset outer ring proportion threshold is above 95%.

[0067] Condition B: α particles exceeding the preset inner circle proportion threshold are distributed outside the ellipse; the preset inner circle proportion threshold is above 95%.

[0068] According to the classification method for platelet α-particle distribution described in this invention, a large number of images of individual platelet α-particle distributions are classified and labeled, and these images of different α-particle distributions are trained, verified and tested to achieve automatic classification of different α-particle distribution types, and to construct a platelet classification system based on platelet α-particle super-resolution images.

[0069] Based on this, the present invention also provides a platelet classification system based on platelet α particle super-resolution images, which includes a platelet α particle super-resolution image acquisition module, a single platelet α particle distribution image extraction and annotation module, and a classification module.

[0070] The platelet α-particle super-resolution image acquisition module is used to binarize and obtain an image of a single platelet α-particle composed of fluorescent signal pixels and non-fluorescent signal pixels; and submits it to the single platelet α-particle distribution image extraction and annotation module; the resolution of the super-resolution image is below 100nm, which enables fine detection of platelet α-particles and can directly detect changes in platelet α-particles; any super-resolution microscopy imaging technique such as SIM, STED, and STORM can be used to acquire the image;

[0071] The structure of platelet α-granules labeled with fluorescence is specific and does not require identification by professionals. Moreover, by using appropriate α-granule protein antibodies, such as VWF protein, VAMP-8 protein, or SPARC protein antibody, the α-granules in the platelets can be immunofluorescently labeled. By adjusting the fluorescence intensity and exposure time during imaging of each sample, it can be ensured that each super-resolution image reconstructed by the super-resolution algorithm is an image with a high signal-to-noise ratio, i.e., the ratio of fluorescence signal to background signal of the original image is ≥7:1. This is beneficial for the system to identify and extract the distribution of α-granules in individual platelets.

[0072] The single platelet α-particle distribution image extraction and annotation module is used to segment the fluorescent signal pixel region into the pixel region of independent α-particles in the single platelet α-particle image using an image segmentation algorithm, obtain the single platelet α-particle distribution image, and submit it to the classification module.

[0073] The classification module is used to classify the platelets according to the distribution pattern of α particles, as follows:

[0074] (1) Search for the ellipse with the smallest area that satisfies condition A within the α-particle image of a single platelet, and use it as the outer circle: if the area of ​​the outer circle is above a preset regular distribution ellipse area threshold and the ratio of the minor axis to the major axis is within the preset regular distribution ellipse range, then the platelets involved in the single platelet α-particle distribution image are determined to be "regularly distributed"; the regular distribution ellipse area threshold is within 0.15 μm. 2 -20.0μm 2 Between, preferably 3.0μm 2 -13.0μm 2 The range of the regularly distributed ellipse is [0.2, 1], preferably [0.5, 1].

[0075] For images of α-particles within a single platelet with a "regular distribution," the platelets are subclassified according to the distribution pattern of the α-particles, as follows:

[0076] The ellipse with the largest area satisfying condition B within the outer ring is selected as the inner ring; if the ratio of the area of ​​the inner ring to that of the outer ring exceeds a preset annular distribution threshold, then the platelets involved in the single platelet α-particle distribution image are determined to be in the "annular distribution" subcategory; otherwise, the platelets involved in the single platelet α-particle distribution image are determined to be in the "particle distribution" subcategory; the annular distribution threshold ranges from [0.5, 1).

[0077] (2) Search for the ellipse with the smallest area within the α-particle image of a single platelet that satisfies condition A, and use it as the outer circle: if the area of ​​the outer circle is below the area threshold of the aggregation distribution ellipse, then the platelets involved in the single platelet α-particle distribution image are determined to be "aggregated"; the area threshold of the aggregation distribution ellipse is within 0.02 μm. 2 -7.0μm 2 Between, preferably 2.5μm 2 -4.5μm 2 Between; the area threshold of the regularly distributed ellipse is greater than or equal to the area threshold of the clustered distributed ellipse;

[0078] (3) If the α-particle image of a single platelet does not belong to "regular distribution" and "aggregated distribution", then the platelets involved in the single platelet α-particle distribution image are judged to be "scattered distribution".

[0079] For images of scattered individual platelet α-particles, the number of fluorescent pixel blocks is counted as the number N of α-particles;

[0080] The subcategories of α-particle images within a single platelet are classified based on the relationship between the number of α-particles and a preset threshold. The preset threshold is set according to the actual situation of scientific or clinical research, preferably capable of distinguishing between experimental and control groups; in some embodiments, the preset threshold is taken between 2 and 70.

[0081] Condition A: α particles exceeding a preset outer ring proportion threshold are distributed inside the ellipse; the preset outer ring proportion threshold is above 95%.

[0082] Condition B: α particles exceeding the preset inner circle proportion threshold are distributed outside the ellipse; the preset inner circle proportion threshold is above 95%.

[0083] The following is an example:

[0084] Example 1: Platelet Classification Method Based on Platelet α-Particle Super-Resolution Images

[0085] Platelet classification methods based on super-resolution images of platelet alpha particles, such as Figure 2 As shown, it includes the following steps:

[0086] Obtaining super-resolution images of platelet α-granules: VWF protein antibody was used to immunofluorescence label α-granules within platelets. Super-resolution microscopy (SIM) was then used to image a large number of individual platelet α-granules, obtaining multiple super-resolution images of platelet α-granules. One of these super-resolution images is shown below. Figure 1 As shown;

[0087] (2) Classification: Based on the distribution pattern of α-granules within a single platelet, classification is carried out according to the following principles:

[0088] (1) Search for the ellipse with the smallest area in the single platelet α-particle image where more than 95% of the particles are distributed within it, and use it as the outer circle: if the area of ​​the outer circle is above a preset regular distribution ellipse area threshold and the ratio of the minor axis to the major axis is within the preset regular distribution ellipse range, then the platelets involved in the single platelet α-particle distribution image are determined to be "regularly distributed"; the regular distribution ellipse area threshold is 5.8 μm. 2 The range of the regularly distributed ellipse is [0.2, 1].

[0089] For images of α-particles within a single platelet with a "regular distribution," the platelets are subclassified according to the distribution pattern of the α-particles, as follows:

[0090] The inner circle is defined as the ellipse with the largest area where more than 95% of the alpha particles are distributed outside the outer circle. If the ratio of the area of ​​the inner circle to the area of ​​the outer circle exceeds a preset annular distribution threshold, the platelets involved in the single platelet alpha particle distribution image are determined to be in the "annular distribution" subcategory; otherwise, the platelets involved in the single platelet alpha particle distribution image are determined to be in the "particle distribution" subcategory. The annular distribution threshold is set to 0.8.

[0091] (2) Search for the ellipse with the smallest area within more than 95% of the α-particles in the single platelet α-particle image, and define it as the outer circle: if the area of ​​the outer circle is below the area threshold of the aggregation distribution ellipse, then the platelets involved in the single platelet α-particle distribution image are determined to be "aggregated"; the area threshold of the aggregation distribution ellipse is 2.8 μm. 2 between.

[0092] (3) If the α-particle image of a single platelet does not belong to "regular distribution" and "aggregated distribution", then the platelets involved in the single platelet α-particle distribution image are judged to be "scattered distribution".

[0093] For images of scattered individual platelet α-particles, the number of fluorescent pixel blocks is counted as the number N of α-particles;

[0094] The number of α particles is divided into subcategories based on their relationship with a preset threshold number, representing the "particle distribution" within a single platelet α particle image. The preset threshold number is set according to the actual situation of scientific or clinical research, preferably one that can distinguish between the experimental group and the control group; in this embodiment, the preset threshold is set to 10, i.e.

[0095] If the number N of α particles satisfies "N<10" or "N≥10", then the single platelet subclass is "N<10" or "N≥10".

[0096] Example 2: Platelet Classification System Based on Platelet α-Particle Super-Resolution Images

[0097] The process of establishing a platelet classification system based on super-resolution images of platelet alpha particles is as follows: Figure 3 As shown, the details are as follows:

[0098] (1) Obtaining super-resolution images of platelet α-particles: VWF protein antibody was used to immunofluorescence label α-particles within platelets. Super-resolution microscopy (SIM) was used to image a large number of individual platelet α-particles, obtaining multiple super-resolution images of platelet α-particles. One of these super-resolution images is shown below. Figure 1 As shown;

[0099] (2) Extracting and labeling super-resolution images of individual platelet α particles: The obtained super-resolution images of platelet α particles are cropped to obtain a large number of fluorescence images of individual platelet α particles. The distribution types of individual platelet α particles from a large number of different samples are labeled as "regular distribution," "aggregated distribution," or "scattered distribution" according to the method provided in Example 1. The distribution types are as follows: Figure 2 As shown.

[0100] We used algorithms such as ResNet50 and SVM to train, validate, and test a large number of manually labeled alpha particle fluorescence images of different types, and built a classification system that can autonomously crop, extract, label, and classify them.

[0101] This invention utilizes super-resolution microscopy to detect platelet alpha particles, discovering three distribution patterns: "regular distribution," "aggregated distribution," and "scattered distribution," potentially representing different physiological functions or reaction mechanisms of platelets. Upon activation, platelets release alpha particles beyond the cell membrane. This invention reveals that the "regular distribution" of single resting platelet alpha particles may indicate that the alpha particles within the resting platelet have tended towards the cell membrane and are released extracellularly, thereby releasing immune factors or coagulation factors to enhance the body's immune and coagulation responses. Therefore, the "regular distribution" of alpha particles is beneficial for studying the coagulation or immune response mechanisms of platelets. After platelet activation, multiple platelet alpha particles aggregate. This invention reveals that the "aggregated distribution" of single resting platelet alpha particles may indicate that the resting platelet is in a hypercoagulable state, making it highly sensitive to activators. Therefore, the "aggregated distribution" of alpha particles is beneficial for studying the mechanisms of thrombosis or hemorrhage under pathological conditions. The "scattered distribution" of α-granules within a single platelet, and the quantity of α-granules, can indirectly reflect the status of megakaryocytes in producing α-granule precursors, which is beneficial for in-depth research on the mechanism of megakaryocytes in producing α-granule precursors.

[0102] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A platelet classification method based on platelet alpha granule super-resolution images, characterized by, The method comprises the following steps: obtaining an alpha granule immunofluorescence labeled super-resolution image of a single platelet and binarizing the image to obtain an alpha granule image of the single platelet composed of fluorescent signal pixels and non-fluorescent signal pixels; for the alpha granule image of the single platelet, using an image segmentation algorithm to segment the fluorescent signal pixel region into independent alpha granule pixel regions, and obtaining an alpha granule distribution image of the single platelet; classifying the platelet according to the distribution form of the alpha granule, specifically as follows: (1) searching for an ellipse in the image of the single platelet alpha granules that satisfies condition A and has the smallest area, as the outer ring: if the area of the outer ring is above a preset regular distribution ellipse area threshold and the ratio of the minor axis to the major axis is within a preset regular distribution ellipse range, then the platelet involved in the image of the single platelet alpha granules is determined to be "regularly distributed"; the regular distribution ellipse area threshold is between 0.15 μm 2 - 20.0 μm 2 ; the regular distribution ellipse range is [0.2, 1]; the "regular distribution" refers to the distribution of alpha granules in the platelet to form a ring shape, and the ring shape includes a ring shape without alpha granules in the interior and a ring shape with alpha granules distributed in the interior and / or the exterior. (2) search the ellipse in the single platelet alpha granule image that meets condition A and has the minimum area as the outer ring: if the area of the outer ring is below the aggregation distribution ellipse area threshold, then determine that the platelet involved in the single platelet alpha granule distribution image is "aggregation distribution"; the aggregation distribution ellipse area threshold is between 0.02 μm 2 -7.5 μm 2 ; the regular distribution ellipse area threshold is greater than or equal to the aggregation distribution ellipse area threshold; the "aggregation distribution" refers to multiple alpha granules gathered together to form one or more clusters, wherein the size of a single cluster is between 0.02 μm 2 -7.0 μm 2 ; (3) if the alpha granule image of the single platelet does not belong to "regular distribution" and "aggregated distribution", judging that the platelet involved in the alpha granule distribution image of the single platelet is "scattered distribution"; condition A: more than a preset outer ring proportion threshold value of alpha granules are distributed in the interior of the ellipse; the preset outer ring proportion threshold value is above 95%.

2. The platelet classification method based on platelet alpha granule super-resolution images of claim 1, wherein, for the alpha granule image of the single platelet of "regular distribution", classifying the platelet according to the distribution form of the alpha granule, specifically as follows: searching for an ellipse in the outer ring that satisfies condition B and has the largest area as an inner ring; if the area ratio of the inner ring to the outer ring exceeds a preset ring distribution threshold value, judging that the platelet involved in the alpha granule distribution image of the single platelet is a "ring distribution" subclass, otherwise judging that the platelet involved in the alpha granule distribution image of the single platelet is a "granular distribution" subclass; the ring distribution threshold value is in the range of [0.5, 1); condition B: more than a preset inner ring proportion threshold value of alpha granules are distributed in the exterior of the ellipse; the preset inner ring proportion threshold value is above 95%.

3. The platelet classification method based on platelet alpha granule super-resolution images of claim 1, wherein, for the alpha granule distribution image of the single platelet of "scattered distribution", counting the number of fluorescent pixel blocks as the number N of alpha granules; dividing the subclass of the alpha granule image of the single platelet of "granular distribution" according to the number relationship between the number of alpha granules and a preset number threshold value.

4. The platelet classification method based on platelet alpha granule super-resolution images according to claim 3, wherein, The preset number threshold value is in the range of 2-70.

5. The platelet classification method based on platelet alpha granule super resolution images of claim 1, wherein, The regular distribution ellipse area threshold is between 3.0 pm 2 - 13.0 pm 2 ; the regular distribution ellipse range is [0.5, 1]; the aggregated distribution ellipse area threshold is between 2.5 pm 2 - 4.5 pm 2 .

6. A platelet classification system based on platelet alpha granule super-resolution images, characterized by, The method comprises a platelet alpha granule super-resolution image acquisition module, a single platelet alpha granule distribution image extraction and labeling module, and a classification module; The platelet alpha granule super-resolution image acquisition module is used to obtain an alpha granule image of a single platelet composed of fluorescent signal pixels and non-fluorescent signal pixels by binarization, and submit the image to the single platelet alpha granule distribution image extraction and labeling module; The single platelet alpha granule distribution image extraction and labeling module is used to, for the alpha granule image of the single platelet, segment the fluorescent signal pixel region into independent alpha granule pixel regions using an image segmentation algorithm, obtain an alpha granule distribution image of the single platelet, and submit the image to the classification module; The classification module is used to classify the platelet according to the distribution form of the alpha granule, specifically as follows: (1) searching for an ellipse in the image of the single platelet alpha granules that satisfies condition A and has the smallest area, as the outer ring: if the area of the outer ring is above a preset regular distribution ellipse area threshold and the ratio of the minor axis to the major axis is within a preset regular distribution ellipse range, then the platelet involved in the image of the single platelet alpha granules is determined to be "regularly distributed"; the regular distribution ellipse area threshold is between 0.15 μm 2 - 20.0 μm 2 ; the regular distribution ellipse range is [0.2, 1]; the "regular distribution" refers to the distribution of alpha granules in the platelet to form a ring shape, and the ring shape includes a ring shape without alpha granules in the interior and a ring shape with alpha granules distributed in the interior and / or the exterior. (2) search the ellipse in the single platelet alpha granule image that meets condition A and has the minimum area as the outer ring: if the area of the outer ring is below the aggregation distribution ellipse area threshold, then determine that the platelet involved in the single platelet alpha granule distribution image is "aggregation distribution"; the aggregation distribution ellipse area threshold is between 0.02 μm 2 -7.5 μm 2 ; the regular distribution ellipse area threshold is greater than or equal to the aggregation distribution ellipse area threshold; the "aggregation distribution" refers to multiple alpha granules gathered together to form one or more clusters, wherein the size of a single cluster is between 0.02 μm 2 -7.0 μm 2 ; (3) if the alpha granule image of the single platelet does not belong to "regular distribution" and "aggregated distribution", judging that the platelet involved in the alpha granule distribution image of the single platelet is "scattered distribution"; condition A: more than a preset outer ring proportion threshold value of alpha granules are distributed in the interior of the ellipse; the preset outer ring proportion threshold value is above 95%.

7. The platelet classification system based on platelet alpha-granule super-resolution images of claim 6, wherein, Also included is a rule distribution sub-classification module; for a single platelet alpha granule image of "rule distribution", the platelet is sub-classified according to the distribution form of alpha granules, specifically as follows: Search for the largest ellipse in the outer circle that meets condition B as the inner circle; if the area ratio of the inner circle to the outer circle exceeds the preset ring distribution threshold, it is judged that the platelet involved in the single platelet alpha granule distribution image is of the "ring distribution" subcategory, otherwise it is judged that the platelet involved in the single platelet alpha granule distribution image is of the "granular distribution" subcategory; the ring distribution threshold value range is [0.5, 1); Condition B: more than 95% of the alpha granules in the outer part of the ellipse exceed the preset inner circle proportion threshold value.

8. The platelet classification system based on platelet alpha-granule super-resolution images of claim 6, wherein, Also included is a scattered distribution sub-classification module; for a single platelet alpha granule distribution image of "scattered distribution", the number of fluorescent pixel blocks is counted as the number of alpha granules N; The number of alpha granules is divided into subcategories according to the number relationship with the preset number threshold.

9. The platelet classification system based on platelet alpha-granule super-resolution images of claim 8, wherein, The preset number threshold is between 2-70.

10. The platelet classification system based on platelet alpha-granule super-resolution images of claim 9, wherein, The regular distribution ellipse area threshold is between 3.0 pm 2 - 13.0 pm 2 ; the regular distribution ellipse range is [0.5, 1]; the aggregated distribution ellipse area threshold is between 2.5 pm 2 - 4.5 pm 2 .

Images