Oral periodontal macrophage in-situ typing and functional state identification method
By injecting low-concentration metabolic disturbances into periodontal pockets and combining surface-enhanced Raman spectroscopy with an unsupervised learning model, the problem of accurately identifying macrophage functional status in traditional methods has been solved, enabling dynamic monitoring and precise typing in vivo.
Patent Information
- Application Number
- CN202511524521.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-24
- Publication Date
- 2026-01-09
AI Technical Summary
Traditional methods for identifying macrophage subtypes rely on in vitro analysis, which disrupts the cell's native microenvironment and fails to accurately reflect its true state in vivo. Furthermore, typing based on static markers cannot capture the functional activity of macrophages.
A combination of low-concentration metabolic disturbance agents was injected into the periodontal pocket using a temperature-sensitive gel. Surface-enhanced Raman spectroscopy was used to scan at multiple time points to obtain fingerprint spectra at single-cell resolution. Combined with a self-organizing mapping unsupervised learning model, functional labels were automatically identified and assigned.
It accurately captures the functional state, heterogeneity, and plasticity of macrophages in complex pathological environments, providing a more realistic tool for identifying functional state.
Smart Images
Figure CN121294599A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the technical field of periodontal macrophage identification, and more particularly relates to a method for in-situ typing and functional state identification of periodontal macrophages in oral cavity. BACKGROUND
[0002] As a key immune regulatory cell, macrophages play a double-edged sword role in various diseases. They can clear pathogens and promote tissue repair, but also release inflammatory factors and active oxygen to cause tissue damage. Therefore, accurately analyzing the subtype composition and functional state of macrophages is of great significance for revealing the pathogenesis of diseases, evaluating disease activity, and developing targeted treatment strategies.
[0003] Traditional tests and identification of macrophage subtypes mainly rely on ex vivo analysis techniques such as flow cytometry, which requires cell separation or sectioning from periodontal tissue. This destroys the native microenvironment of cells and cannot reflect the true state of cells in vivo. At the same time, it is highly dependent on pre-set surface or intracellular molecular markers. However, macrophages have high plasticity and heterogeneity, and the expression of markers is not absolutely exclusive, and dynamically changes during inflammation, resulting in that the typing based on static markers cannot accurately capture the functional activity. SUMMARY
[0004] The application provides a method for in-situ typing and functional state identification of periodontal macrophages in oral cavity, which aims to solve the technical problem that the current typing based on static markers cannot accurately capture the functional activity of macrophages.
[0005] A method for in-situ typing and functional state identification of periodontal macrophages in oral cavity, comprising the following steps: S1. Injecting a warm gel solution containing calcium ions into the periodontal pocket to form a gel-covered area; S2. Using a microsyringe to deliver a low-concentration metabolic disturbance agent combination to the gel-covered area to stimulate metabolic disturbance of macrophages; S3. Using a SERS device to perform Raman spectrum scanning on the disturbed area at a plurality of predetermined time points to obtain time-series spectrum data of each cell; S4. Collecting spectrum data of non-cell areas as background templates while scanning the SERS device, and performing background signal subtraction based on the collected background templates and the obtained time-series spectrum data of each cell to obtain corrected time-series spectrum data; S5. Extracting intensity values of characteristic wave bands from the corrected time-series spectrum data for each cell, drawing a time-intensity curve based on the extracted intensity values, and constructing a ROS index and an arginine metabolism index based on the time-intensity curve; S6. Inputting the ROS index and arginine metabolism index of all cells into the SOM unsupervised learning model, automatically identifying the clustering pattern, outputting the classification label of each cell based on the clustering pattern and the predefined rules, and associating to the corresponding functional activity.
[0006] In the present application, the low-concentration metabolic regulator combination is injected into the periodontal pocket by using the temperature-sensitive gel to actively stimulate the core function of macrophages to reveal their real-time activity state; then, by means of surface-enhanced Raman spectroscopy, scanning is performed at multiple preset time points under the premise of keeping the periodontal microenvironment intact without separation, and single-cell resolution fingerprint spectrum is obtained, and the ROS index and arginine metabolism index are accurately quantified in the fingerprint spectrum, and then the two indexes are input into the self-organizing mapping unsupervised learning model for clustering, and the model automatically identifies the population according to the actual dynamic response of the cells to the stimulus; and then a functional label is given according to the biological rules, and the functional state, heterogeneity and plasticity of macrophages in a complex pathological environment are accurately captured.
[0007] Preferably, the step S1 comprises the following steps: Under sterile conditions, 30 grams of Pluronic F-127 powder is dissolved in 100 milliliters of ionized water, and then 5mM Ca²⁺ solution is added to the ionized water and stirred uniformly to obtain a prepared temperature-sensitive gel solution; The temperature-sensitive gel solution is injected into the periodontal pocket by using a syringe, and the temperature-sensitive gel solution undergoes phase transition in a body temperature environment to form a gel covering area.
[0008] Preferably, the low-concentration metabolic disturbance agent combination comprises an M1 pathway stimulant and an M2 pathway stimulant. The MI pathway stimulant comprises adenosine triphosphate and lipopolysaccharide, wherein the concentration of adenosine triphosphate is 10 micromole / liter, and the concentration of lipopolysaccharide is 0.1 microgram / ml, and wherein the adenosine triphosphate and the lipopolysaccharide are both dissolved in sterile phosphate buffer. The M2 pathway stimulant comprises L-arginine and interleukin-10, wherein the concentration of L-arginine is 1 millimole / liter, and the concentration of interleukin-10 is 10 nanomole / liter, and wherein the L-arginine and the interleukin-10 are both dissolved in sterile phosphate buffer.
[0009] Preferably, the multiple time points for Raman spectrum scanning comprise scanning immediately after stimulation, scanning two minutes after stimulation, scanning five minutes after stimulation, and scanning eight minutes after stimulation; and multiple scans are performed in the same field of view during scanning.
[0010] Preferably, data acquisition is performed on three key wavebands during Raman spectrum scanning, wherein the three wavebands comprise an oxidative stress waveband, a lipid peroxidation waveband, and a metabolite waveband. The wavelength range of the oxidative stress band is 500-600 cm⁻¹, the wavelength range of the lipid peroxidation band is 1500-1600 cm⁻¹, and the wavelength range of the metabolite band is 1000-1100 cm⁻¹. Based on this, time-series spectral data is obtained, including cell ID, time point, and spectral data.
[0011] Preferably, the background signal subtraction step is as follows: Based on the collected spectral data of non-cellular regions, the average value of the spectral data of non-cellular regions is calculated to obtain the background template; The spectral intensities of cellular and non-cellular regions at 1200 cm⁻¹ are obtained, and the ratio of the spectral intensities of cellular regions at 1200 cm⁻¹ to those of non-cellular regions at 1200 cm⁻¹ is used as a dynamic weight. The corrected time-series spectral data is obtained by subtracting the product of the background template and the dynamic weight from the acquired raw time-series spectral data.
[0012] Preferably, the steps for extracting the ROS index are as follows: The spectral data of the two bands, 500-600 cm⁻¹ and 1500-1600 cm⁻¹, are integrated separately. The integrated results of the two bands are then added together to obtain the cumulative intensity. The cumulative intensity is then integrated over a period of 0 to 8 minutes, and the average value is calculated to obtain the ROS index.
[0013] Preferably, the extraction steps for the arginine metabolic index are as follows: The spectral intensity at 1030 cm⁻¹ was extracted, and the difference between the spectral intensity at 1030 cm⁻¹ and at 8 minutes and 0 minutes after stimulation was calculated to obtain the arginine metabolism index.
[0014] Preferably, the SOM unsupervised learning model performs clustering based on the input data to obtain four cluster centers. For each cluster center, the average value is calculated. Based on the calculated average value, each cluster center is divided into corresponding functional states using predefined rules, and the cells belonging to the cluster center are assigned corresponding functional state labels.
[0015] Preferably, the predefined rules are as follows: If the average ROS index in any cluster is greater than the average arginine metabolism index multiplied by a predetermined first coefficient, it is determined to be in the M1 functional state. If the average value of the arginine metabolism index in any cluster is greater than the average value of the ROS index multiplied by a predetermined second coefficient, it is determined to be in the M2 functional state. If the above two conditions are not met, it is marked as a Transitional state.
[0016] The beneficial effects of this invention include: In this invention, a low-concentration combination of metabolic regulators is injected into the periodontal pocket using a temperature-sensitive gel to actively stimulate the core functions of macrophages and reveal their real-time activity state. Subsequently, surface-enhanced Raman spectroscopy is used to scan at multiple preset time points without separation and while maintaining the integrity of the periodontal microenvironment, to obtain single-cell resolution fingerprint spectra. The ROS index and arginine metabolism index are precisely quantified in the fingerprint spectra. Then, the two indices are input into a self-organizing map unsupervised learning model for clustering. The model automatically identifies the population based on the actual dynamic response of cells to stimuli. Finally, functional labels are assigned according to biological rules, accurately capturing the functional state, heterogeneity, and plasticity of macrophages in complex pathological environments. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This is an overall step diagram provided for an embodiment of the present invention. Detailed Implementation
[0019] To make the technical problems, technical solutions, and beneficial effects to be solved by this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and are not intended to limit the scope of this application.
[0020] See Figure 1 As shown, a method for in situ typing and functional status identification of oral periodontal macrophages includes the following steps: S1. Inject a temperature-sensitive gel solution containing calcium ions into the periodontal pocket to form a gel-covered area; Step S1 includes the following steps: Under aseptic conditions, 30 g of Pluronic F-127 powder was dissolved in 100 mL of deionized water, and then 5 mM Ca²⁺ solution was added to the deionized water and stirred until homogeneous to obtain the prepared thermosensitive gel solution. The thermosensitive gel solution is injected into the periodontal pocket using a syringe. The thermosensitive gel solution undergoes a phase change in the body temperature environment, forming a semi-solid isolation layer, i.e., the gel-covered area, within about 10 seconds. The semi-solid structure formed by the gel provides a fixed substrate for macrophages, ensuring the stability of the cells in subsequent processing and effectively blocking interference from saliva and bacteria, thus providing a stable environment for subsequent metabolic stimulation.
[0021] S2. Use a microinjector to deliver a low-concentration combination of metabolic disturbance agents into the gel-covered area to stimulate macrophages to metabolize. The low-concentration metabolic disturbance combination includes M1 pathway stimulants and M2 pathway stimulants; The MI pathway stimulant includes adenosine triphosphate (ATP) and lipopolysaccharide (LPS), wherein the concentration of ATP is 10 μmol / L and the concentration of LPS is 0.1 μg / mL, and both ATP and LPS are dissolved in sterile phosphate buffer. In this embodiment, 10 μmol / L ATP can effectively trigger ROS burst without causing cytotoxicity; and 0.1 μg / mL LPS can effectively activate the M1 pathway without inducing an excessive inflammatory response. The M2 pathway stimulant includes L-arginine and interleukin-10, wherein the concentration of L-arginine is 1 mmol / L and the concentration of interleukin-10 is 10 nanomol / L, wherein both L-arginine and interleukin-10 are dissolved in sterile phosphate buffer. In this embodiment, the use of 1 mmol / L L-arginine can significantly enhance urea metabolism, and the use of 10 nanomol / L interleukin-10 can effectively inhibit the activation of the M1 pathway while promoting the polarization of the M2 pathway. In this embodiment, the M1 pathway stimulant and the M2 pathway stimulant can be injected simultaneously or in batches.
[0022] S3. Using a SERS device, Rapman spectral scanning is performed on the disturbed region at multiple predetermined time points to obtain temporal spectral data for each cell; The multiple time points for Rapman spectral scanning include scanning immediately after stimulation, two minutes after stimulation, five minutes after stimulation, and eight minutes after stimulation; multiple scans are performed within the same field of view during the scanning process.
[0023] The Rapman spectroscopy scan acquires data in three key bands, including the oxidative stress band, the lipid peroxidation band, and the metabolite band. The oxidative stress band has a wavelength range of 500-600 cm⁻¹, which corresponds to the vibration of the Fe-S bond and can reflect the generation of reactive oxygen species (ROS). The lipid peroxidation band has a wavelength range of 1500-1600 cm⁻¹, which corresponds to the C=C stretching vibration and can reflect the degree of lipid peroxidation. The metabolite band has a wavelength range of 1000-1100 cm⁻¹, which corresponds to the characteristic peak of urea and can reflect changes in arginine metabolism. Based on this, time-series spectral data, including cell ID, time point, and spectral data, are obtained.
[0024] In this embodiment, the selection of the specific wavelength bands is to detect biomarkers related to macrophage metabolic dynamics. Each wavelength band corresponds to a different molecular vibrational mode, which can reflect the dynamic response of cells in terms of oxidative stress, lipid peroxidation, and metabolite changes. Based on the analysis of the corresponding wavelength bands, a systematic study of the kinetic energy changes of macrophages after metabolic perturbation can be achieved.
[0025] S4. During the SERS scan, spectral data of non-cellular regions are simultaneously acquired as a background template. Based on the acquired background template and the time-series spectral data of each cell, background signal subtraction is performed to obtain corrected time-series spectral data. The background signal subtraction steps are as follows: Based on the collected spectral data of non-cellular regions, the average value of the spectral data of non-cellular regions is calculated to obtain the background template; The spectral intensities of cellular and non-cellular regions at 1200 cm⁻¹ are obtained, and the ratio of the spectral intensities of the cellular region to those of the non-cellular region at 1200 cm⁻¹ is used as a dynamic weight; the specific expression is as follows: ; In the formula: Indicates dynamic weights; Indicates the cellular region in Spectral intensity at; This represents the spectral intensity of the non-cellular region at 1200 cm⁻¹; The corrected time-series spectral data is obtained by subtracting the product of the background template and the dynamic weights from the acquired raw time-series spectral data, as shown in the following expression: ; In the formula: This represents the corrected time-series spectral data; Represents the raw time-series spectral data; Indicates the background template; In this embodiment, the following is adopted: The spectral intensity at a given location is used to calculate the dynamic weights because, The surrounding region is primarily associated with COC asymmetric stretching vibrations, a vibrational mode that is extremely sensitive to changes in the local environment of the molecular chain, including hydrogen bonding interactions, chain segment motion, and alterations in solvation state; thus... The peak at the point becomes an ideal probe for characterizing the dynamic behavior of polymer systems. Secondly, it is used as the core of dynamic weight design, mainly because it has a high sensitivity to key dynamic processes of the system. It can not only improve the accuracy of background subtraction, but also provide high-quality spectral data for subsequent metabolic response kinetic modeling.
[0026] S5. For each cell, extract the intensity values of the characteristic bands from the corrected time-series spectral data, plot the time-intensity curve based on the extracted intensity values, and construct the ROS index and arginine metabolism index based on the time-intensity curve; The steps for extracting the ROS index are as follows: The spectral data in the 500-600 cm⁻¹ and 1500-1600 cm⁻¹ bands were integrated separately, and then the integrated results of the two bands were added together to obtain the cumulative intensity. The cumulative intensity was then integrated over a period of 0 to 8 minutes, and the average value was calculated to obtain the ROS index. The specific expression is as follows: ; In the formula: Indicates wave number; Indicates a point in time; This represents the integral value of spectral data in the 500-600 cm⁻¹ range; This represents the integral value of spectral data from 1500 to 1600 cm⁻¹. Indicates cumulative intensity; ; In the formula: Indicates the time step; Represents the ROS index; The extraction steps for the arginine metabolic index are as follows: The spectral intensity at 1030 cm⁻¹ was extracted, and the difference between the spectral intensity at 1030 cm⁻¹ and at 8 minutes and 0 minutes after stimulation was calculated to obtain the arginine metabolism index, the specific expression of which is as follows: ; In the formula: This represents the spectral intensity at time t; ; In the formula: Indicates time Spectral intensity at; Indicates time Spectral intensity at; Indicates a time point 8 minutes after stimulation; Indicates the time point immediately after the stimulus is collected; Indicates the arginine metabolism index; In this embodiment, analysis was performed using two wavelength bands: 500-600 cm⁻¹ and 1500-1600 cm⁻¹. This effectively reflects oxidative stress and lipid peroxidation. Furthermore, by integrating the values from both wavelengths and averaging them over 0 to 8 minutes, the ROS index was obtained, providing a comprehensive and quantifiable measure of cumulative oxidative damage throughout the stimulation process. Secondly, the spectral intensity at 1030 cm⁻¹ was extracted, corresponding to the relevant chemical bonds in arginine metabolism, allowing for a precise reflection of the cellular arginine metabolism level. This was further demonstrated through calculation... and The intensity interpolation at the point directly reflects the changes in arginine metabolism during stimulation, which is helpful for assessing the dynamic changes in metabolic flux.
[0027] S6. Input the ROS index and arginine metabolism index of all cells into the SOM unsupervised learning model, automatically identify the clustering pattern, output the classification label of each cell based on the clustering pattern and predefined rules, and associate it with the corresponding functional activity; The SOM unsupervised learning model clusters the input data to obtain four cluster centers. For each cluster center, an average value is calculated. Based on the calculated average value, predefined rules are used to classify each cluster center into corresponding functional states, and cells belonging to the same cluster center are assigned corresponding functional state labels. For example, the specific details are as follows: The SOM unsupervised learning model uses a 2×2 SOM network with a total of 4 neurons. The weight vector of each neuron is initialized with random values ranging from 0 to 1. During training, the number of iterations was set to 200, and the initial learning rate was 0.3, which was linearly decreased to 0.1 with each iteration. The formula for updating the learning rate is as follows: ; In the formula: This indicates an initial learning rate of 0.3. This indicates a final learning rate of 0.1. Indicates the current iteration number; This represents the total number of training iterations, which is 200. Furthermore, during training, the neighborhood radius decreases linearly from its initial value to zero. In this embodiment, the initial value of the neighborhood radius is set to the maximum distance of the network, that is, the maximum Euclidean distance in a 2×2 network. The adjustment formula for the neighborhood radius is as follows: ; In the formula: This indicates that the initial neighborhood radius is 1.414; Training is performed based on the parameters set above: Calculate the Euclidean distance between the input sample and all output layer neurons, and find the neuron with the smallest distance, i.e., BMU. Based on the current iteration number, and using the above formulas for calculating the learning rate and neighborhood radius, calculate the current learning rate and neighborhood radius. Use a Gaussian function to determine the degree of influence of neurons in the BMU neighborhood: ; In the formula: Indicates the neighborhood influence coefficient; This represents the Euclidean distance between neuron i and the BMU in the output grid coordinate system; Update the weight vectors of the BMU and its neighbors based on the neighborhood function and learning rate: ; In the formula: This represents the updated weight vector; This represents the current weight vector; Indicates the input sample; After 200 iterations, check if the weight vector has converged. If the weight vector is less than a preset threshold, training ends; otherwise, the number of iterations is extended. After training, save the final weight vector as the cluster center.
[0028] The SOM unsupervised learning model outputs four cluster centers, each containing multiple cells with the same or similar features. Based on this, functional classification is performed according to predefined rules. If the average ROS index in any cluster is greater than the average arginine metabolism index multiplied by a predetermined first coefficient, it is determined to be in the M1 functional state, indicating that the macrophages exhibit pro-inflammatory properties; in this embodiment, the value of the first coefficient is 2, and it can be adjusted according to the experience of domain experts. If the average value of the arginine metabolism index in any cluster is greater than the average value of the ROS index multiplied by a predetermined second coefficient, it is determined to be in the M2 functional state; this indicates that the macrophages exhibit anti-inflammatory and repair properties. In this embodiment, the second coefficient is set to 3, and it can be adjusted according to the experience of experts in the field.
[0029] If the above two conditions are not met, it is marked as a Transitional state, indicating that it is in the intermediate stage of functional transformation, that is, the transformation from M1 to M2 or from M2 to M1.
[0030] In this embodiment, a low-concentration metabolic disturbance agent combination is introduced in situ into the periodontal pocket using a temperature-sensitive gel carrier to actively stimulate the core functional metabolic pathways of macrophages, forcing cells to exhibit real-time and dynamic functional activity rather than relying on static molecular expression. Secondly, using surface-enhanced Raman spectroscopy, cells are scanned in situ at multiple predetermined time points under non-destructive, separation-free conditions while maintaining the integrity of the physiological microenvironment, obtaining fingerprint spectra with single-cell resolution. Background subtraction ensures signal specificity, and key functional activity indicators are quantified and extracted from the dynamic spectra. These indicators are then input into a SOM unsupervised learning model for cluster analysis. The model automatically identifies populations based entirely on the actual functional response patterns of cells and assigns functional labels according to predefined biological rules, abandoning preset static markers. By directly stimulating and dynamically monitoring the core metabolic function output of macrophages under in vivo micro-perturbation, the unsupervised model performs clustering and typing based on actual functional performance, thus accurately capturing the functional state, heterogeneity, and plasticity of macrophages in complex pathological environments. This provides a more realistic and reliable tool for understanding the immune imbalance mechanism of periodontitis and for precise immune intervention.
[0031] The above are merely preferred embodiments of this application and are not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the protection scope of this application.
Claims
1. A method for in situ typing and functional status identification of oral periodontal macrophages, characterized in that, Includes the following steps: S1. Inject a temperature-sensitive gel solution containing calcium ions into the periodontal pocket to form a gel-covered area; S2. Use a microinjector to deliver a low-concentration combination of metabolic disturbance agents into the gel-covered area to stimulate macrophages to metabolize. S3. Using a SERS device, Rapman spectral scanning is performed on the disturbed region at multiple predetermined time points to obtain temporal spectral data for each cell; S4. During the SERS scan, spectral data of non-cellular regions are simultaneously acquired as a background template. Based on the acquired background template and the time-series spectral data of each cell, background signal subtraction is performed to obtain corrected time-series spectral data. S5. For each cell, extract the intensity values of the characteristic bands from the corrected time-series spectral data, plot the time-intensity curve based on the extracted intensity values, and construct the ROS index and arginine metabolism index based on the time-intensity curve; S6. Input the ROS index and arginine metabolism index of all cells into the SOM unsupervised learning model, automatically identify the clustering pattern, output the classification label of each cell based on the clustering pattern and predefined rules, and associate it with the corresponding functional activity.
2. The method for in situ typing and functional status identification of oral periodontal macrophages according to claim 1, characterized in that, Step S1 includes the following steps: Under aseptic conditions, 30 g of Pluronic F-127 powder was dissolved in 100 mL of deionized water, and then 5 mM Ca²⁺ solution was added to the deionized water and stirred until homogeneous to obtain the prepared thermosensitive gel solution. The thermosensitive gel solution is injected into the periodontal pocket using a syringe. The thermosensitive gel solution undergoes a phase change in the body temperature environment, forming a gel-covered area.
3. The method for in situ typing and functional status identification of oral periodontal macrophages according to claim 1, characterized in that, The low-concentration metabolic disturbance combination includes M1 pathway stimulants and M2 pathway stimulants; The MI pathway stimulant comprises adenosine triphosphate (ATP) and lipopolysaccharide (LPS), wherein the concentration of ATP is 10 μmol / L and the concentration of LPS is 0.1 μg / mL, and both ATP and LPS are dissolved in sterile phosphate buffer. The M2 pathway stimulant comprises L-arginine and interleukin-10, wherein the concentration of L-arginine is 1 mmol / L and the concentration of interleukin-10 is 10 nanomol / L, wherein both L-arginine and interleukin-10 are dissolved in sterile phosphate buffer.
4. The method for in situ typing and functional status identification of oral periodontal macrophages according to claim 1, characterized in that, The multiple time points for Rapman spectral scanning include scanning immediately after stimulation, two minutes after stimulation, five minutes after stimulation, and eight minutes after stimulation; multiple scans are performed within the same field of view during the scanning process.
5. The method for in situ typing and functional status identification of oral periodontal macrophages according to claim 1, characterized in that, The Rapman spectroscopy scan acquires data in three key bands, including the oxidative stress band, the lipid peroxidation band, and the metabolite band. The wavelength range of the oxidative stress band is 500-600 cm⁻¹, the wavelength range of the lipid peroxidation band is 1500-1600 cm⁻¹, and the wavelength range of the metabolite band is 1000-1100 cm⁻¹. Based on this, time-series spectral data is obtained, including cell ID, time point, and spectral data.
6. The method for in situ typing and functional status identification of oral periodontal macrophages according to claim 1, characterized in that, The background signal subtraction steps are as follows: Based on the collected spectral data of non-cellular regions, the average value of the spectral data of non-cellular regions is calculated to obtain the background template; The spectral intensities of cellular and non-cellular regions at 1200 cm⁻¹ are obtained, and the ratio of the spectral intensities of cellular regions at 1200 cm⁻¹ to those of non-cellular regions at 1200 cm⁻¹ is used as a dynamic weight. The corrected time-series spectral data is obtained by subtracting the product of the background template and the dynamic weight from the acquired raw time-series spectral data.
7. The method for in situ typing and functional status identification of oral periodontal macrophages according to claim 1, characterized in that, The steps for extracting the ROS index are as follows: The spectral data of the two bands, 500-600 cm⁻¹ and 1500-1600 cm⁻¹, are integrated separately. The integrated results of the two bands are then added together to obtain the cumulative intensity. The cumulative intensity is then integrated over a period of 0 to 8 minutes, and the average value is calculated to obtain the ROS index.
8. The method for in situ typing and functional status identification of oral periodontal macrophages according to claim 1, characterized in that, The extraction steps for the arginine metabolic index are as follows: The spectral intensity at 1030 cm⁻¹ was extracted, and the difference between the spectral intensity at 1030 cm⁻¹ and at 8 minutes and 0 minutes after stimulation was calculated to obtain the arginine metabolism index.
9. The method for in situ typing and functional status identification of oral periodontal macrophages according to claim 1, characterized in that, The SOM unsupervised learning model clusters the input data to obtain four cluster centers. For each cluster center, the average value is calculated. Based on the calculated average value, each cluster center is divided into corresponding functional states using predefined rules, and the cells belonging to the cluster center are assigned the corresponding functional state labels.
10. The method for in situ typing and functional status identification of oral periodontal macrophages according to claim 1, characterized in that, The predefined rules are as follows: If the average ROS index in any cluster is greater than the average arginine metabolism index multiplied by a predetermined first coefficient, it is determined to be in the M1 functional state. If the average value of the arginine metabolism index in any cluster is greater than the average value of the ROS index multiplied by a predetermined second coefficient, it is determined to be in the M2 functional state. If the above two conditions are not met, it is marked as a Transitional state.