Melatonin detection method

Abstract

The invention discloses a melatonin detection method which comprises the following steps: S1, coupling an anti-melatonin antibody with a PE fluorescent label with carboxyl fluorescent magnetic beads to construct antibody coupled magnetic beads MBs (at) Ab; s2, preparing a melatonin analogue FITC-MT by using the 5-MT and the FITC-NHS (N-hydroxysuccinimide); s3, preparing MBs (at) Ab and FITC-MT into a reaction system, reacting with melatonin standard substances with different concentrations or a to-be-detected sample, detecting by using a flow cytometer, processing a detection result of the flow cytometer by using a machine learning algorithm, and calculating the melatonin concentration in the to-be-detected sample. MT in a to-be-detected sample can compete for FITC-MT combined with MBs (at) Ab in a short time to form a stable reaction system, the multi-dimensional features of MBs (at) Ab are subjected to flow detection to directly remove impurities, agglomeration and the like in MBs (at) Ab, different fluorescence features of an optimal magnetic bead group in the system are searched and analyzed in combination with a machine learning algorithm, and the content of MT can be accurately detected; the reaction system constructed by the invention is simple to operate and extremely low in cost, and meanwhile, the result accuracy can be ensured.

Description

Technical Field

[0001] The present invention relates to the field of melatonin detection, and in particular to a melatonin detection method. Background Art

[0002] Circadian rhythms are regular oscillations of physiological, behavioral, and metabolic activities that occur within an organism over a roughly 24-hour period. They are driven by an endogenous biological clock and regulated by external timing factors such as the environmental light-dark cycle. The core of their molecular mechanism is a transcriptional-translational negative feedback loop. This rhythmic system is divided into central and local biological clocks in peripheral tissues, forming a multi-organ coordinated oscillation network. Circadian rhythm disturbances are closely associated with a variety of pathophysiological processes. Studies have shown that long-term circadian rhythm disturbances can significantly increase the incidence and development of cardiovascular disease, malignant tumors, and neurodegenerative diseases. Conversely, correcting rhythm disturbances plays an important role in the treatment of these diseases. Furthermore, with the increasing understanding of circadian rhythm processes, chronotherapy is considered one of the most promising directions in personalized medicine.

[0003] The detection of circadian rhythms is crucial in both basic and clinical research. Melatonin (MT), with its well-defined diurnal secretion pattern and sensitivity to light signals, has become the gold standard for assessing circadian clock phase and function. MT synthesis is regulated by the pineal gland in a suprachiasmatic nucleus-dependent manner, with its nocturnal peak tightly controlled by the endogenous circadian clock and rapidly inhibited by ambient light intensity. Dynamic monitoring of plasma or salivary MT levels allows for precise quantification of circadian clock phase shifts and amplitudes, a method essential for the diagnosis of jet lag syndrome, a non-24-hour sleep-wake disorder. Therefore, efficient detection of MT is crucial for in-depth research into its biological functions and clinical translation, particularly in the diagnosis and precision treatment of diseases associated with circadian rhythm disorders.

[0004] In recent years, MT detection techniques have been continuously optimized, from traditional radioimmunoassay (RIA) to highly sensitive liquid chromatography-tandem mass spectrometry (LC-MS / MS), with detection limits reaching the pg / mL level. The urinary excretion rhythm of the MT metabolite 6-sulfaminomethyl melatonin (aMT6s) provides a noninvasive, long-term monitoring method, particularly suitable for epidemiological studies. However, current methods for MT detection, including high-performance liquid chromatography, liquid chromatography-tandem mass spectrometry, radioimmunoassay, and enzyme-linked immunosorbent assay, all have limitations that restrict their clinical application.

[0005] High-performance liquid chromatography (HPLC) requires stringent chromatographic separation and cumbersome sample pretreatment steps; liquid-phase tandem mass spectrometry (LC-MS / MS) requires expensive equipment and specialized technicians; radioimmunoassays (RIAs) are subject to radioisotope radiation and contamination, resulting in large batch-to-batch coefficients of variation; and enzyme-linked immunosorbent assays (ELISAs) often experience cross-reactions, long incubation times, and poor accuracy. Therefore, due to current technological limitations, existing detection methods are still unable to meet the needs of clinical testing: they require both high sensitivity and high accuracy, simple and fast operational procedures without reliance on expensive testing equipment, and low-cost, rapid and accurate testing that can be performed by non-professionals. Summary of the Invention

[0006] The purpose of the present invention is to provide a melatonin detection method, which on the one hand addresses the shortcomings of current MT detection and on the other hand meets the detection needs of clinical practice.

[0007] To achieve the above object, the present invention provides a melatonin detection method comprising the following steps:

[0008] S1. Conjugate the PE fluorescently labeled anti-melatonin antibody to carboxyl fluorescent magnetic beads to construct antibody-coupled magnetic beads MBs@Ab;

[0009] S2. Use 5-MT and FITC-NHS to prepare the melatonin analog FITC-MT;

[0010] S3. MBs@Ab and FITC-MT are prepared into a reaction system, reacted with different concentrations of melatonin standards or test samples, and detected by flow cytometry. A machine learning algorithm is used to correct the flow cytometry detection results and calculate the melatonin concentration in the test sample.

[0011] Preferably, in S1, the preparation of the PE fluorescently labeled anti-melatonin antibody comprises:

[0012] 1 mg / mL anti-melatonin antibody was added to the heterobifunctional cross-linker and incubated at room temperature for 1 hour to obtain a mixture, which was washed with PBS, ultrafiltered, centrifuged, and resuspended in PBS to 1 mg / mL;

[0013] The resuspension was incubated with modified PE at room temperature for 4 hours, and PBS buffer containing a quencher, BSA, and sodium azide was added and incubated at room temperature for 1 hour to obtain PE-labeled anti-MT antibodies.

[0014] Preferably, in S1, the method for coupling the PE fluorescently labeled anti-melatonin antibody to the carboxyl fluorescent magnetic beads comprises:

[0015] The carboxyl fluorescent magnetic beads were washed twice with 20 times the volume of activation buffer of the carboxyl fluorescent magnetic beads. Activation buffer, 50 mg / mL sulfo-NHS and 50 mg / mL EDAC were added at a volume ratio of 10:200:2:1 of carboxyl fluorescent magnetic beads: activation buffer: sulfo-NHS: EDAC, and incubated at room temperature for 30 minutes. PE fluorescent-labeled anti-melatonin antibody and labeling buffer were added at a volume ratio of 10:1:40 of carboxyl fluorescent magnetic beads: PE fluorescent-labeled anti-melatonin antibody: labeling buffer, and incubated at room temperature for 2 hours. The complex was blocked by adding 40 times the volume of 1% BSA solution of the carboxyl fluorescent magnetic beads and incubated for 2 hours. The complex was washed 2-3 times with 20 times the volume of TBST of the carboxyl fluorescent magnetic beads to obtain antibody-coupled magnetic beads MBs@Ab.

[0016] Preferably, in S2, the preparation of the melatonin analog FITC-MT comprises:

[0017] 3.8 mg of 5-methoxytryptamine and 11.8 mg of FITC-NHS were dissolved in 2 mL of DMSO, triethylamine was added to a final volume of 10 mL, and the mixture was incubated at room temperature for 12 hours to obtain the melatonin analog FITC-MT.

[0018] Preferably, in S3, the reaction system is constructed by mixing 125 μL of melatonin standard or test sample with 10 μL of 3.5 nmol / L FITC-MT, adding 5 μL of MBs@Ab, and incubating at room temperature in the dark for 3 hours.

[0019] Preferably, in S3, the machine learning algorithm is an XGBoost algorithm, and the application method includes:

[0020] S31. Normalize the raw flow cytometric signals to eliminate scale differences, screen and separate magnetic bead signals using the intrinsic fluorescence characteristic parameters of the magnetic beads, divide the signal data into training and test sets in a ratio of 7:3, and divide the magnetic beads into different subsets based on K-means clustering;

[0021] S32. Perform feature engineering on 15 original parameters extracted from five channels of magnetic beads, including detection time, detection peak, and detection signal area. Combine the extracted key feature set with the melatonin concentration label to construct a training dataset, and use the training dataset for training.

[0022] S33. During the training process, grid search combined with 5-fold cross validation is applied to evaluate the generalization performance of each set of parameters and determine the optimal parameter configuration;

[0023] The XGBoost model uses mean square error as the loss function, iteratively optimizes the prediction results through gradient boosting, and gradually improves the determination coefficient of the fitting curve.

[0024] The specific application principle of the present invention is that MT and MT mimetic (FITC-MT) can compete for binding to specific binding sites on antibody-coupled magnetic beads (MBs@Ab). The concentrations of FITC-MT and MT in the reaction system can be reflected by detecting the multidimensional fluorescence signal characteristics of MBs@Ab using a flow cytometer. This multidimensional fluorescence information is then analyzed and processed by a machine learning algorithm to obtain the MT concentration in the sample to be tested.

[0025] Therefore, the melatonin detection method of the present invention has the following beneficial effects:

[0026] (1) The detection process of the present invention utilizes the principle that the specificity of MT binding to antibodies is greater than the binding efficiency of FITC-MT, and the specificity of FITC-MT binding to antibodies is greater than that of 5-MT. Therefore, purification of FITC-MT is not required during the preparation process, and MT in the sample to be tested can quickly compete with the bound FITC-MT, forming a stable reaction system. Impurities, agglomerations, and other conditions can be directly eliminated by flow cytometry detection of the multidimensional characteristics of MBs@Ab. Combined with a machine learning algorithm, the different fluorescence characteristics of the optimal microsphere population in the system are found and analyzed, enabling accurate detection of MT content. The reaction system constructed by the present invention is simple to operate and extremely low in cost, while also ensuring accurate results.

[0027] (2) Under oscillating reaction conditions, the MBs@Ab magnetic beads will perform irregular Brownian motion in the reaction solution, giving the coupled antibodies more opportunities to bind to the MT and FITC-MT to be tested. The antigen-antibody specific binding ability is extremely strong, enabling rapid reaction, completing the reaction process in a short period of time and enabling detection. Unlike some existing test kits on the market that require batch testing, this detection method is not affected by sample size, and even single-tube testing does not increase costs.

[0028] (3) Since a single magnetic bead has a large number of carboxyl binding sites, and after binding to the antibody, it further forms corresponding antigen binding sites; therefore, there are a large number of antigen binding sites in the system, and its detection ability can be adjusted according to the MBs@Ab and FITC-MT in the competitive reaction system, and the detection range of MT is wide.

[0029] (4) The present invention has strong scalability. The magnetic beads selected in the present invention can carry other different types of antibodies (all antibodies are evenly bound to one microsphere) in a highly uniform system, and can be expanded to achieve the purpose of detecting the content of different substances in one system.

[0030] (5) The present invention can effectively solve the problems of uneven dispersion of magnetic beads in the reaction system and uneven reaction conditions by analyzing the flow single-particle detection and multi-dimensional data feature analysis mode. By dynamically adjusting the weights of different characteristic fluorescence in the analysis process, the present invention focuses on finding and performing analysis and calculation based on the most uniform group of magnetic beads in the reaction system, thereby reducing random errors. In combination with machine learning algorithms, the precise detection and analysis required for clinical practice can be achieved.

[0031] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0033] Figure 1 Flowchart for antibody labeling and antibody coupling to magnetic beads;

[0034] Figure 2 Laser confocal images before and after magnetic beads were coupled to antibodies;

[0035] Figure 3 The flow cytometry results before and after the antibody is coupled to the magnetic beads, where A is the flow cytometry result before the antibody is coupled to the magnetic beads, and B is the flow cytometry result after the antibody is coupled to the magnetic beads;

[0036] Figure 4 is the absorption spectrum of the antibody solution with gradient concentration;

[0037] Figure 5 is the linear relationship between the absorbance of the antibody solution at 280 nm and the antibody concentration;

[0038] Figure 6 The amount of antibody consumed during the process of magnetic bead coupling;

[0039] Figure 7 is the PE fluorescence positive rate after magnetic beads are coupled with antibodies;

[0040] Figure 8 is the PE positive rate of MBs@Ab stored in a dark environment at 4°C;

[0041] Figure 9 Comparison of PE positive rates of MBs@Ab at 4°C and room temperature;

[0042] Figure 10 Schematic diagram of FITC-MT synthesis;

[0043] Figure 11 This is the laser confocal image after FITC-MT and MBs@Ab were mixed and incubated;

[0044] Figure 12 The flow cytometry results after incubation of 5-MT or FITC-MT with MBs@Ab are shown, where A is the flow cytometry result after incubation of 5-MT with MBs@Ab, and B is the flow cytometry result after incubation of FITC-MT with MBs@Ab;

[0045] Figure 13 Figure 5 shows the flow cytometry results of different reaction ratios of 5-MT and FITC-NHS, where A is a 5-MT:FITC-NHS volume ratio of 1:1, B is a 5-MT:FITC-NHS volume ratio of 1:1.25, C is a 5-MT:FITC-NHS volume ratio of 1:1.50, D is a 5-MT:FITC-NHS volume ratio of 1:1.75, and E is a FITC-NHS volume ratio of 1:2.

[0046] Figure 14 is the FITC positive rate at different reaction ratios of 5-MT and FITC-NHS;

[0047] Figure 15 The FITC positive rate after incubation of FITC-MT with different concentrations and MBs@Ab for 0.5 h;

[0048] Figure 16 The FITC positive rate after incubation of different concentrations of FITC-MT and MBs@Ab for 3 hours;

[0049] Figure 17 The flow cytometry results after incubation of 1.0 nmol / mL FITC-MT and MBs@Ab for 1, 3, and 12 hours, where A represents 1 hour, B represents 3 hours, and C represents 12 hours.

[0050] Figure 18 FITC-positive rates after 1, 3, and 12 h of incubation of 0.5 nmol / mL and 1.0 nmol / mL FITC-MT with MBs@Ab;

[0051] Figure 19 The FITC-positive rate after 300 pg / mL melatonin was mixed with 0.25 nmol / mL, 0.5 nmol / mL, and 0.75 nmol / mL FITC-MT and then incubated with MBs@Ab for 3 hours;

[0052] Figure 20 Flowchart of data processing for machine learning algorithms;

[0053] Figure 21 Figure 3 shows the screening results of magnetic beads based on their characteristic parameters after testing melatonin standards at different concentrations. A represents MT = 0 pg / mL, B represents MT = 9.68 pg / mL, C represents MT = 18.75 pg / mL, D represents MT = 37.50 pg / mL, E represents MT = 75 pg / mL, F represents MT = 150 pg / mL, and G represents MT = 300 pg / mL.

[0054] Figure 22 The clustering of magnetic beads under different numbers of clusters, where A is the cluster grouping with 3 clusters, B is the cluster grouping with 4 clusters, C is the cluster grouping with 5 clusters, D is the cluster grouping with 6 clusters, E is the cluster grouping with 7 clusters, F is the cluster grouping with 8 clusters, and G is the cluster grouping with 9 clusters;

[0055] Figure 23 is the fitting result during the machine learning algorithm training process, where A, B, and C are the fitting conditions of the standard curve as the training process progresses;

[0056] Figure 24 Statistics on R2 changes during machine learning algorithm training;

[0057] Figure 25 This is the standard curve processed by the machine learning algorithm;

[0058] Figure 26 The following are the flow cytometry test results after processing by the machine learning algorithm, where A represents a melatonin standard concentration of 0 pg / mL, B represents a melatonin standard concentration of 9.68 pg / mL, C represents a melatonin standard concentration of 18.75 pg / mL, D represents a melatonin standard concentration of 37.5 pg / mL, E represents a melatonin standard concentration of 75 pg / mL, F represents a melatonin standard concentration of 150 pg / mL, and G represents a melatonin standard concentration of 300 pg / mL.

[0059] Figure 27 To demonstrate the detection performance of MT content using MBs@Ab;

[0060] Figure 28 Comparison of melatonin concentrations in human serum determined by MBs@Ab and ELISA;

[0061] Figure 29 Statistics of serum melatonin concentrations in healthy people and patients with heart failure;

[0062] Figure 30 The scatter plot shows the correlation between age and melatonin concentration. DETAILED DESCRIPTION

[0063] The technical solution of the present invention is further described below with reference to the accompanying drawings and embodiments.

[0064] In order to make the purpose, technical solutions and advantages of the present application clearer, more thorough and more complete, the technical solutions of the present invention are clearly and completely described below through the accompanying drawings and Examples. The following detailed description is an explanation of the embodiments and is intended to provide further details of the present invention. Unless otherwise specified, all technical terms used in the present invention have the same meaning as those generally understood by those skilled in the art to which the application belongs.

[0065] The instruments, equipment, reagents and materials used in the examples were obtained from commercial sources.

[0066] Example 1

[0067] Preparation of PE fluorescently labeled anti-MT antibody and coupling of anti-MT antibody with carboxyl fluorescent magnetic beads, the process is as follows Figure 1 shown.

[0068] Fluorescently label the anti-MT antibody using a commercial PE antibody labeling kit (Cat: EX7035, Solebro) according to the kit instructions. Add the anti-MT antibody (1 mg / mL, Ulsen) to a heterobifunctional crosslinker and incubate at room temperature for 1 hour. Wash the mixture with PBS buffer, then ultrafilter and centrifuge, and resuspend it in PBS buffer to 1 mg / mL. Transfer the resuspension to a tube containing modified PE and incubate at room temperature for 4 hours. Then, add PBS buffer containing a quencher, BSA, and sodium azide and incubate at room temperature for 1 hour to obtain the PE-labeled anti-MT antibody.

[0069] The carboxyl fluorescent magnetic beads (MBs) used were purchased from Bixin Biotechnology (Hainan Province) Co., Ltd. 10 μL of beads (1×10 7 Beads / mL) were washed twice. Subsequently, 200 μL activation buffer, 2 μL 50 mg / mL sulfo-NHS and 1 μL 50 mg / mL EDAC were added and incubated at room temperature for 30 minutes. For antibody conjugation, 1 μL antibody (1 mg / mL) and 40 μL labeling buffer (0.1 M MEST, pH 5.0, 0.05% Tween 20) were added and incubated at room temperature for 2 hours. The complex was then blocked with 400 μL 1% BSA solution and incubated for 2 hours. After washing 2-3 times with 200 μL TBST, the prepared antibody-coupled magnetic beads (MBs@Ab) were stored in the dark at 2-8°C for later use.

[0070] like Figure 2As shown in the figure, laser confocal microscopy revealed that compared with uncoupled magnetic beads, antibody-coupled magnetic beads showed obvious fluorescence around them under the PE fluorescence field.

[0071] like Figure 3 As shown in Figure 3, flow cytometry analysis showed that the antibody-coupled magnetic beads exhibited significantly enhanced PE fluorescence intensity compared to the uncoupled magnetic beads.

[0072] These findings confirmed the successful coupling of anti-MT antibodies to magnetic beads.

[0073] Antibody consumption during the binding process was quantified using a micro-volume spectrophotometer. The absorbance of the antibody solution at varying concentrations was measured and an absorption spectrum was generated, such as Figure 4 As shown. A standard curve is generated based on the absorbance of the solution at 280nm, as shown Figure 5 As shown in the figure, the consumption of antibodies in the process of magnetic bead coupling antibodies is calculated based on the standard curve and absorbance, as shown in the figure. Figure 6 The results showed that in each coupling process, an average of 61.1% of the antibody in the antibody solution with an initial concentration of 1 mg / mL was successfully coupled to the magnetic beads.

[0074] The threshold of PE fluorescence positive magnetic beads was set to 1000 (the average PE fluorescence intensity of magnetic beads before coupling was 225 ± 77.3), and the PE fluorescence intensity of magnetic beads after coupling was analyzed by flow cytometry. Figure 7 The results showed that an average of over 99% of the magnetic beads were PE-positive, indicating that under the established experimental conditions and procedures, the magnetic beads obtained were almost all magnetic beads with melatonin antibodies.

[0075] The MBs@Ab magnetic beads were placed at 4°C in a dark environment for stability test. Figure 8 As shown in the figure, the PE positive rate of the magnetic beads can be maintained above 99% within 7 days.

[0076] The MBs@Ab magnetic beads were placed at room temperature and protected from light for stability test. Figure 9 As shown in the figure, the PE positive rate of magnetic beads is equivalent to that at 4°C.

[0077] These results indicate that the melatonin antibody is stably and reliably coupled to the magnetic beads, can be stored for a long time at 4°C in a dark environment, and maintains very good stability in a room temperature detection environment.

[0078] Example 2

[0079] Preparation of MT mimetic (FITC-MT) is as follows Figure 10 shown.

[0080] 3.8 mg of 5-methoxytryptamine (5-MT) (Cat: HY-W015169, purity ≥99%, MCE) and 11.8 mg of FITC-NHS (Cat: A498967, purity ≥95%, Ambeed) were dissolved in 2 mL of DMSO (Cat: D6258, purity ≥99%, McLean), and then triethylamine (Cat: T103285, purity ≥99%, Aladdin) was added to a final volume of 10 mL. The mixture was incubated at room temperature for 12 hours to obtain the melatonin analog FITC-MT.

[0081] Test Example 1

[0082] After FITC-MT and MBs@Ab were mixed and incubated for 1 hour, the mixture was observed using a laser confocal microscope. Figure 11 The results showed that under the FITC fluorescence field, obvious fluorescence appeared around MBs@Ab, confirming that the synthesized FITC-MT could be captured by the anti-melatonin antibody on MBs@Ab.

[0083] Test Example 2

[0084] Analyze whether the possible impurities in the synthesized product affect the binding of the antibody-coupled magnetic beads to FITC-MT.

[0085] First, 5-MT or FITC-MT was mixed with MBs@Ab and incubated for 1 hour, and then detected by flow cytometry. Figure 12 The results showed that the flow cytometry results of MBs@Ab mixed with FITC-MT showed a significant increase in FITC fluorescence (positive threshold: >10 4 ), while the flow cytometry results of MBs@Ab mixed with 5-MT showed that no obvious changes in FITC were observed.

[0086] Then, different ratios of 5-MT and FITC-NHS were used for chemical reaction, and the reaction products were mixed with MBs@Ab and incubated for a period of time before flow cytometry detection. Figure 13 And the FITC positive rate under different reaction ratios of 5-MT and FITC-NHS was determined, as shown in Figure 14 The results showed that when the amount of FITC-NHS was equal to or greater than that of 5-MT, the reaction product would cause >99% of the magnetic beads to show enhanced FITC fluorescence.

[0087] The above results show that the purity of FITC-MT in the synthesized product is very high, and the possible impurities have almost no effect on the binding of FITC-MT and MBs@Ab.

[0088] Test Example 3

[0089] Different concentrations of FITC-MT were mixed with MBs@Ab and incubated for 0.5 h and 3 h before flow cytometry detection. The proportions of FITC-positive magnetic beads were as follows: Figure 15 and 16 shown.

[0090] Depend on Figure 15 、 Figure 16 It can be seen that when the FITC-MT concentration reaches above 1 nmol / mL, the magnetic beads are basically FITC-positive.

[0091] 1.0 nmol / mL FITC-MT was mixed with MBs@Ab and incubated for 1, 3, and 12 hours before flow cytometry detection. Figure 17 shown.

[0092] 0.5 nmol / mL and 1.0 nmol / mL FITC-MT were mixed with MBs@Ab and incubated for 1, 3, and 12 h, respectively, and then flow cytometry was performed. The FITC-positive magnetic beads were more than Figure 18 shown.

[0093] Depend on Figure 17 、 Figure 18 It can be seen that during the 3-12 h incubation of FITC-MT and MBs@Ab, the proportion of FITC-positive magnetic beads remained basically stable.

[0094] The above results indicate that the ideal incubation time of FITC-MT and MBs@Ab is 3 h, and the appropriate concentration of FITC-MT for co-incubation with MBs@Ab may be around 0.5 nmol / mL.

[0095] 300 pg / mL melatonin was mixed with 0.25 nmol / mL, 0.5 nmol / mL, and 0.75 nmol / mL FITC-MT concentrations, and then MBs@Ab was added and incubated for 3 hours. Flow cytometry was then performed. The results are shown in Figure 2. Figure 19 The results showed that the positive rates of FITC magnetic beads were 32.5%, 63.2%, and 99.7%, respectively. To ensure a wide range of fluorescence discrimination within the melatonin concentration range of 0 to 300 pg / mL, a FITC-MT concentration of 0.25 nmol / mL was selected for subsequent MT detection.

[0096] Test Example 4

[0097] (1) Construction of reaction system and detection of MT using MBs@Ab

[0098] Melatonin (Cat: HY-B0075, purity ≥99%, MCE) was added to antibody-depleted plasma from healthy individuals to prepare melatonin standard solutions of varying concentrations (0, 9.68, 18.75, 37.5, 75, 150, and 300 pg / mL).

[0099] Mix 125 μL of standard or sample with 10 μL of 3.5 nmol / mL FITC-MT, then add 5 μL of the prepared MBs@Ab. Incubate at room temperature in the dark for 3 hours. Adjust the voltage of each flow cytometer channel, set the detection volume to 100 μL, and perform the assay.

[0100] The flow cytometry data of the same batch of standards and samples were imported into the machine learning algorithm for analysis and processing to obtain the melatonin concentration of the sample to be tested.

[0101] (2) Use the XGBoost algorithm to optimize the detection performance of magnetic beads (MBs@Ab).

[0102] In order to solve the problem of MBs@Ab homogeneity changing during the detection process and affecting the detection performance, a machine learning algorithm is introduced to improve the detection accuracy. The machine learning algorithm is mainly used to complete the screening, calibration and regression calculation of MBs@Ab. The processing flow is as follows: Figure 20 As shown, the specific implementation process is as follows:

[0103] A large amount of flow cytometry data from the detection of melatonin standards at different concentrations was annotated for the training and evaluation of machine learning models. A total of approximately 54,000 MBs@Ab post-reaction data and approximately 8,000 baseline MBs@Ab magnetic beads were analyzed. The extracted parameters included the detection time limit (W), detection peak intensity (H), and detection signal area (A) of the five channels FSC, SSC, PE, FITC, and APC, totaling 15 raw parameters, which were then processed through feature engineering. The entire dataset was divided into 70% training data and 30% test data, and trained using the XGBoost algorithm. To ensure robust performance, grid search hyperparameter adjustment was used, and 5-fold cross-validation techniques were applied to analyze model performance.

[0104] The test data of melatonin standards of different concentrations were processed by the following process:

[0105] (1) Data preprocessing: First, the original streaming detection signal is normalized to eliminate the scale differences between different samples and channels.

[0106] (2) Magnetic bead identification and screening: Based on the intrinsic fluorescence characteristic parameters of MBs@Ab magnetic beads, a screening algorithm is applied to identify and separate the target magnetic beads, effectively removing background noise and impurity particles (such as Figure 21 shown).

[0107] (3) K-means-based magnetic bead clustering: The post-reaction magnetic bead population screened by S2 is unsupervisedly grouped using the K-means clustering algorithm. The magnetic beads with heterogeneity are divided into different subsets (such as Figure 22 shown).

[0108] (4) Feature Engineering: Based on the cluster analysis, further feature engineering is performed. Combining the clustering results with prior knowledge of the melatonin response mechanism, key feature parameters that can effectively characterize the relationship between magnetic beads and melatonin concentration response are screened, extracted, and constructed. Feature combinations are optimized to enhance the effectiveness of model input information.

[0109] (5) Model training and evaluation: The extracted key feature set is combined with the corresponding melatonin standard concentration label to construct a training dataset, and the dataset is used to train the XGBoost model.

[0110] During the training of the machine learning algorithm, the coefficient of determination of the fitting curve continues to improve, e.g. Figure 23 and Figure 24 The final fitting curve generated after model processing is shown as Figure 25 As shown, R 2 Reached 0.9946.

[0111] The flow detection results processed by the machine learning algorithm are as follows Figure 26 These findings indicate that the developed flow cytometry data algorithm effectively addresses the impact of changes in MBs@Ab homogeneity and that this method can effectively detect melatonin content in solution.

[0112] Test Example 5

[0113] According to the ELISA kit instructions (Cat: ELK9164, ELK (Wuhan) Biological), melatonin standards or samples are added to the microplate wells and incubated in the dark to promote antigen-antibody binding. After continuous washing, biotinylated antibodies are introduced to bind to the antigen. Unbound components are removed by repeated washing, and then the HRP-coupled enzyme is applied to amplify the signal. TMB substrate is then added to initiate the enzymatic reaction and terminated with an acidic stop solution. The absorbance at 450 nm is measured by a microplate reader, and the melatonin concentration is calculated by comparing the standard curve.

[0114] The test method of Test Example 4 and the established XGBoost model were used for testing.

[0115] The ELISA kit was used to compare with the method in Test Example 4. The results are shown in Table 1:

[0116] Table 1 Comparison of ELISA kit and test example 4

[0117]

[0118] The detection performance of the method in test case 4 is as follows Figure 27 As shown, the detection results are basically consistent with the actual MT concentration.

[0119] The ELISA kit and XGBoost model were further used to detect the melatonin concentration in human serum. The test results are shown in Table 2:

[0120] Table 2 Comparison of serum test results between ELISA kit and test example 4

[0121]

[0122]

[0123] like Figure 28 As shown in Figure 3, the MT level derived by the XGBoost model is very close to the MT level determined by the ELISA kit.

[0124] These experiments showed that the detection results of the test example 4 method were comparable to those of the ELISA kit and could effectively quantify the concentration of melatonin in serum.

[0125] Test Example 6

[0126] The MT detection method constructed by the present invention is used to detect and analyze the MT level in the serum of patients with heart failure.

[0127] Further validation was performed using sera from 44 healthy individuals and 44 patients with heart failure. The healthy individuals were matched to the heart failure patients based on age and gender. Serum sampling was performed between 6:00 AM and 7:00 AM.

[0128] Using the method of Test Example 4, the serum melatonin concentration of healthy individuals was 153.8±73.8pg / mL, and the serum melatonin concentration of heart failure patients was 99.8±72.6pg / mL (as shown in Figure 4). Figure 29 shown).

[0129] Further correlation analysis was performed, such as Figure 30 The results showed that serum melatonin levels were negatively correlated with age. This finding is similar to those reported in the literature: compared with healthy individuals, patients with heart failure have lower serum melatonin levels; and serum melatonin levels in humans tend to decrease with age.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solutions of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for detecting melatonin, characterized in that: The steps include: S1. Conjugate the PE fluorescently labeled anti-melatonin antibody to carboxyl fluorescent magnetic beads to construct antibody-coupled magnetic beads MBs@Ab; S2. Use 5-MT and FITC-NHS to prepare the melatonin analog FITC-MT; S3. MBs@Ab and FITC-MT are prepared into a reaction system, reacted with different concentrations of melatonin standards or test samples, and detected by flow cytometry. A machine learning algorithm is used to process the flow cytometry detection results and calculate the melatonin concentration in the test sample.

2. A melatonin detection method according to claim 1, characterized in that: In S1, the preparation of PE fluorescently labeled anti-melatonin antibody includes: 1 mg / mL anti-melatonin antibody was added to the heterobifunctional cross-linker and incubated at room temperature for 1 hour to obtain a mixture, which was washed with PBS, ultrafiltered, centrifuged, and resuspended in PBS to 1 mg / mL; The resuspension was incubated with modified PE at room temperature for 4 hours, and PBS buffer containing a quencher, BSA, and sodium azide was added and incubated at room temperature for 1 hour to obtain PE-labeled anti-MT antibodies.

3. The melatonin detection method according to claim 1, wherein: In S1, the method for coupling the PE fluorescently labeled anti-melatonin antibody to the carboxyl fluorescent magnetic beads includes: The carboxyl fluorescent magnetic beads were washed twice with activation buffer (20 times the volume of carboxyl fluorescent magnetic beads). Activation buffer, 50 mg / mL sulfo-NHS, and 50 mg / mL EDAC were added at a volume ratio of 10:200:2:1 for carboxyl fluorescent magnetic beads: activation buffer: sulfo-NHS: EDAC, and the mixture was incubated at room temperature for 30 minutes. PE-labeled anti-melatonin antibody and labeling buffer were added at a volume ratio of 10:1:40 for carboxyl fluorescent magnetic beads: PE-labeled anti-melatonin antibody: labeling buffer, and the mixture was incubated at room temperature for 2 hours. The complex was blocked by adding 1% BSA solution (40 times the volume of carboxyl fluorescent magnetic beads) and incubated for 2 hours. The complex was then washed 2-3 times with 20 times the volume of TBST to obtain antibody-coupled magnetic beads MBs@Ab.

4. The melatonin detection method according to claim 1, wherein: In S2, the preparation of the melatonin analog FITC-MT includes: 3.8 mg of 5-methoxytryptamine and 11.8 mg of FITC-NHS were dissolved in 2 mL of DMSO, triethylamine was added to a final volume of 10 mL, and the mixture was incubated at room temperature for 12 hours to obtain the melatonin analog FITC-MT.

5. The melatonin detection method according to claim 1, wherein: In S3, the reaction system was constructed by mixing 125 μL of melatonin standard or test sample with 10 μL of 3.5 nmol / L FITC-MT, adding 5 μL of MBs@Ab, and incubating at room temperature in the dark for 3 h.

6. The melatonin detection method according to claim 1, wherein: In S3, the machine learning algorithm is the XGBoost algorithm, and the application methods include: S31. Normalize the raw flow cytometric signals to eliminate scale differences, screen and separate magnetic bead signals using the intrinsic fluorescence characteristic parameters of the magnetic beads, divide the signal data into training and test sets in a ratio of 7:3, and divide the magnetic beads into different subsets based on K-means clustering; S32. Perform feature engineering on 15 original parameters extracted from five channels of magnetic beads, including detection time, detection peak, and detection signal area. Combine the extracted key feature set with the melatonin concentration label to construct a training dataset, and use the training dataset for training. S33. During the training process, grid search combined with 5-fold cross validation is applied to evaluate the generalization performance of each set of parameters and determine the optimal parameter configuration; The XGBoost model uses mean square error as the loss function, iteratively optimizes the prediction results through gradient boosting, and gradually improves the determination coefficient of the fitting curve.

Images