A nanoparticle magnetic flow detection method and system

Abstract

The application relates to the technical field of biological analysis detection, in particular to a nanoparticle magnetic flow detection method and system; the method comprises collecting a magnetization response signal generated by a magnetic nanoparticle in a biological liquid to be detected under a single-column running track excited by a double-frequency alternating magnetic field and a scattering light pulse signal and a fluorescence pulse signal generated by optical irradiation; extracting a magnetism characteristic parameter of the magnetization response signal and optical characteristic parameters of the scattering light pulse signal and the fluorescence pulse signal; associating the magnetism characteristic parameter and the optical characteristic parameters belonging to the same magnetic nanoparticle; performing state classification on the magnetic nanoparticle based on a comparison between the association result and a preset condition, counting the classification result, and outputting distribution characteristics of various states of the nanoparticle. The application has the effect of improving the accuracy of the biological analysis detection result.

Description

Technical Field

[0001] This application relates to the field of bioanalytical detection technology, and in particular to a magnetic flow cytometry method and system for detecting nanoparticles. Background Technology

[0002] Magnetic nanoparticles are a class of nanoscale materials with magnetic response properties, typically consisting of a magnetic core and a functionalized modification layer surrounding the core. The magnetic core can be made of magnetic materials such as iron oxides, while the outer modification layer can be made of materials such as silica, dextran, polyethylene glycol, antibodies, or nucleic acid aptamers. Because magnetic nanoparticles can generate detectable magnetic response signals under an applied magnetic field, and possess characteristics such as tunable size, easily modifiable surface, and the ability to achieve targeted biological binding, they are applied in fields such as bioseparation, immunoassay, magnetic resonance imaging, and in vitro diagnostics.

[0003] In practical applications of magnetic nanoparticles, it is often necessary to detect and analyze their particle size distribution, aggregation state, surface modification, and binding state with biological targets. For example, when magnetic nanoparticles are used for biodetection or targeted identification, they may need to enter complex biological fluid environments such as serum, plasma, and cell culture media, and bind to target cells, exosomes, or specific biomolecules through surface modification structures. However, complex biological fluids usually contain a large number of natural nano- and submicron-sized particles, including exosomes, microvesicles, lipoproteins, and protein aggregates. These particles share certain similarities with functionalized magnetic nanoparticles in terms of particle size range, surface charge, and colloidal behavior, which can interfere with the accurate identification and state analysis of magnetic nanoparticles.

[0004] In existing technologies, physicochemical characterization methods such as electron microscopy, dynamic light scattering, and zeta potential detection are commonly used to analyze nanoparticles. Among them, electron microscopy can obtain morphological information of nanoparticles, but it requires sample preparation, is easily affected by factors such as sample drying and aggregation during the detection process, and has a low detection throughput; dynamic light scattering can obtain overall particle size information, but its detection results reflect the average scattering characteristics of all particles in the sample, making it difficult to distinguish between magnetic nanoparticles and endogenous particles in biological fluids; zeta potential detection can analyze the surface charge characteristics of particles, but since magnetic nanoparticles tend to form a protein adsorption layer after entering biological fluids, their surface properties may change, making accurate identification based on a single physical parameter difficult.

[0005] Furthermore, existing magnetic detection methods, such as vibrating sample magnetometers, superconducting quantum interference devices, and AC magnetic susceptibility detection methods, while able to reduce non-magnetic background interference by utilizing the inherent magnetic response characteristics of magnetic nanoparticles, typically measure the magnetic signal of the entire sample. This yields the average magnetic response result of multiple magnetic nanoparticles superimposed, failing to capture the magnetic characteristics of individual nanoparticles. When magnetic nanoparticles in a sample are in a freely dispersed, aggregated, or bound state with biological particles, the differences in magnetic responses corresponding to these different states can easily superimpose during the overall measurement process, making it difficult to further determine the actual state of individual magnetic nanoparticles.

[0006] On the other hand, traditional flow cytometry mainly relies on scattered light or fluorescence signals to analyze the target particles, enabling particle event counting. However, for small nanoscale magnetic particles, the optical scattering signal is weak, which is easily limited by detection sensitivity. Furthermore, in complex biological fluids, non-target particles such as endogenous vesicles may produce optical responses similar to those of target particles, making it difficult for traditional optical flow cytometry to simultaneously meet the requirements for specific identification of magnetic nanoparticles and single-particle state analysis.

[0007] In summary, existing methods for detecting magnetic nanoparticles typically rely on a single magnetic or optical signal for analysis. While magnetic detection can identify magnetic particles, it struggles to determine whether the particles carry target molecules or whether non-specific adsorption has occurred. Optical detection, although capable of acquiring fluorescence or scattering information, is susceptible to interference from non-magnetic background particles in complex biological fluids, making it impossible to confirm whether the detected object is indeed the target magnetic nanoparticle. Therefore, in complex biological fluid environments, existing detection methods struggle to simultaneously identify magnetic nanoparticles and determine their binding state, easily leading to significant deviations in the detection results. Summary of the Invention

[0008] To address the issue of large deviations in detection results in related technologies, this application provides a magnetic flow cytometry detection method and system for nanoparticles.

[0009] In a first aspect, this application provides a method for detecting magnetic nanoparticles using flow cytometry, employing the following technical solution: A magnetic flow cytometry detection method for nanoparticles includes: collecting the magnetization response signal generated by magnetic nanoparticles in the biological liquid to be tested under a single-track trajectory when excited by a dual-frequency alternating magnetic field, as well as the scattered light pulse signal and fluorescence pulse signal generated by optical irradiation; Extract the magnetic characteristic parameters of the magnetization response signal, and the optical characteristic parameters of the scattered light pulse signal and the fluorescence pulse signal; correlate the magnetic characteristic parameters and the optical characteristic parameters belonging to the same magnetic nanoparticle; Based on the correlation results and the preset conditions, the magnetic nanoparticles are classified into states, the classification results are statistically analyzed, and the distribution characteristics of each state of the nanoparticles are output.

[0010] Magnetic nanoparticles were forcibly singled out and simultaneously subjected to dual-frequency alternating magnetic field excitation and optical irradiation. The corresponding magnetization response signals, scattered light pulse signals, and fluorescence pulse signals were collected. Magnetic and optical characteristic parameters were further extracted, and the two types of characteristic parameters belonging to the same magnetic nanoparticle were correlated. The state of the magnetic nanoparticles was classified according to the correlation results.

[0011] Compared to detection methods that rely solely on a single magnetic or optical signal, this approach leverages the complementary relationship between magnetic response and optical features. Magnetic features reflect the magnetic response state of the magnetic nanoparticles themselves, while optical features reflect the scattering and fluorescence labeling states of the particles. This single-particle-level information fusion enhances the ability to identify magnetic nanoparticles in complex biofluids. Furthermore, by classifying the associated high-dimensional features, it is possible to distinguish magnetic nanoparticles in different states and output their distribution characteristics. This reduces interference from non-magnetic background particles or particles in different binding states, thereby improving the accuracy of the detection results.

[0012] Optionally, the magnetic characteristic parameters include the magnetic signal amplitude and the magnetic signal phase angle; the optical characteristic parameters include the peak intensity of the scattered light pulse signal and the peak intensity of the fluorescence pulse signal.

[0013] By simultaneously acquiring the magnetic signal amplitude, magnetic signal phase angle, and the intensity of scattered light and fluorescence signals, the detection process can utilize the magnetic response information of the magnetic nanoparticles themselves, particle size-related optical information, and surface marker-related fluorescence information. Specifically, the magnetic signal amplitude reflects the strength of the magnetic response of the magnetic nanoparticles, the magnetic signal phase angle reflects the delay in magnetic response caused by changes in the surrounding environment, the scattered light pulse signal provides particle-related optical response information, and the fluorescence pulse signal provides target marker information. Compared to single-parameter detection methods, this approach increases the characteristic differences between magnetic nanoparticles in different states, improving the reliability of subsequent classification and judgment.

[0014] Optionally, the magnetic characteristic parameters of the magnetization response signal are extracted, including: The magnetization response signal is bandpass filtered; Using the high-order mixing frequency formed by the dual-frequency alternating magnetic field as the local reference mixing carrier, a dual-channel digital phase-locked amplification algorithm is used to orthogonally demodulate the filtered magnetization response signal, resulting in the first mixing component with the highest signal-to-noise ratio. The highest peak value of the voltage pulse envelope corresponding to the first mixing component is extracted as the magnetic signal amplitude, and the phase delay difference between the voltage pulse envelope and the local reference mixing carrier is calculated as the magnetic signal phase angle.

[0015] The first mixing component, which has the highest signal-to-noise ratio from the nonlinear magnetic response signal generated by magnetic nanoparticles, reduces the impact of low-frequency flow noise and other interference signals on the detection results. Simultaneously, by acquiring amplitude and phase information through lock-in amplification, the response intensity and response delay characteristics of the magnetic nanoparticles can be preserved. This allows the detection system not only to determine the presence of magnetic nanoparticles but also to further analyze the differences in magnetic response caused by changes in particle state, improving the ability to distinguish single particles in complex sample environments.

[0016] Optionally, the dual-frequency alternating magnetic field includes a frequency of The first alternating magnetic field and frequency are The second alternating magnetic field; the higher-order mixing frequency is configured as follows: or .

[0017] By employing a dual-frequency alternating magnetic field excitation method, magnetic nanoparticles generate a nonlinear mixing response under the combined action of magnetic fields of different frequencies, thereby forming a characteristic magnetic signal that is distinct from background environmental noise.

[0018] Optionally, the steps for obtaining the peak intensity of the scattered light pulse signal and the peak intensity of the fluorescence pulse signal in the optical characteristic parameters include: The scattered light pulse signal and the fluorescence pulse signal are processed using a moving average filtering algorithm to dynamically eliminate low-frequency drift; The peak intensity of the scattered light pulse signal and the peak intensity of the fluorescence pulse signal after baseline recovery are calculated using a peak search algorithm.

[0019] Since optical detection signals are easily affected by light source fluctuations, changes in the detection environment, and low-frequency drift factors during actual detection, directly using the original optical signals can easily lead to peak extraction errors. Therefore, a moving average filtering algorithm is used to process the scattered light pulse signal and the fluorescence pulse signal to dynamically eliminate low-frequency drift.

[0020] Optionally, the magnetic characteristic parameters belonging to the same magnetic nanoparticle are associated with the optical characteristic parameters, including: Set a time window for single-particle event extraction; If the voltage pulse envelope of the magnetization response signal and the pulse peak of the scattered light pulse signal or the fluorescence pulse signal are extracted simultaneously within any single-particle event extraction time window, then the magnetic signal amplitude, the magnetic signal phase angle, the peak intensity of the scattered light pulse signal and the peak intensity of the fluorescence pulse signal within the current single-particle event extraction time window are merged into a high-dimensional feature vector to complete the association of the two types of feature parameters.

[0021] Within the same time window, the magnetic signal amplitude, magnetic signal phase angle, scattered light peak intensity, and fluorescence peak intensity corresponding to the magnetic nanoparticles are acquired simultaneously, and the above parameters are merged to form a high-dimensional feature vector, thereby realizing the correlation between magnetic feature parameters and optical feature parameters.

[0022] Optionally, the magnetic nanoparticles are classified into states based on the correlation results compared with preset conditions, including a step for identifying the specific target binding state: If the amplitude of the magnetic signal in the high-dimensional feature vector is higher than a preset background noise threshold, and the peak intensity of the fluorescence pulse signal is higher than a preset fluorescence threshold, and the phase angle of the magnetic signal falls into a preset first offset interval, then it is determined to belong to the specific target binding state.

[0023] When the magnetic signal amplitude meets the conditions for the presence of magnetic particles, the fluorescence signal meets the conditions for target labeling, and the phase angle of the magnetic signal is within the corresponding offset range, the event is determined to be a specific target binding state. Combining information from three dimensions—magnetic response, fluorescence labeling, and phase change—can reduce misjudgments of non-target particles caused by relying solely on fluorescence signals, and also reduce the problem of failing to determine the target binding state based solely on magnetic signals, thereby improving the accuracy of identifying the state of target-bound magnetic nanoparticles.

[0024] Optionally, classifying the magnetic nanoparticles into states based on the correlation results and pre-defined conditions also includes steps for identifying free-floating events and non-specific adsorption events. If the amplitude of the magnetic signal is higher than the background noise threshold, and the peak intensity of the fluorescence pulse signal is equal to or lower than the fluorescence threshold, and the phase angle of the magnetic signal falls into the preset non-offset interval, then it is determined to belong to the state of free magnetic nanoparticles. If the amplitude of the magnetic signal is higher than the background noise threshold, and the peak intensity of the fluorescence pulse signal is equal to or lower than the fluorescence threshold, and the phase angle of the magnetic signal falls into the preset second offset interval, then it is determined to belong to the non-specific vesicle adsorption state. Wherein, any absolute value of phase shift within the second offset interval is greater than any absolute value of phase shift within the non-offset interval, and any absolute value of phase shift within the first offset interval is greater than any absolute value of phase shift within the second offset interval.

[0025] Magnetic nanoparticles in different states are classified by comparing the amplitude of the magnetic signal, the intensity of the fluorescence signal, and the range of the magnetic signal phase angle. In the free-floating state, particles lack external coating structures and exhibit small phase changes in the magnetic response. In the non-specific adsorption state, particles show more significant phase shifts due to changes in effective volume caused by the adsorption of biological components. By utilizing changes in the magnetic signal phase angle to distinguish different particle states, non-specific adsorption events can be identified without relying on additional labeling information. This reduces the impact of non-target binding states in complex biological fluids on detection results and improves the precision of analytical analysis.

[0026] Optionally, the magnetic nanoparticles are classified into states based on the correlation results compared with preset conditions, and the process also includes a background impurity removal step: If the amplitude of the magnetic signal is equal to or lower than the background noise threshold, and the peak intensity of the scattered light pulse signal is higher than the preset scattered light threshold, then it is determined to belong to the non-magnetic background impurity state.

[0027] Secondly, this application provides a magnetic flow cytometry detection system for nanoparticles, employing the following technical solution: A nanoparticle magnetic flow cytometry detection system includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a nanoparticle magnetic flow cytometry detection method as described above is implemented.

[0028] The above-mentioned method for detecting magnetic nanoparticles by magnetic flow cytometry is used to generate a computer program, which is then stored in a memory for loading and execution by a processor. This allows for the creation of a system based on the memory and processor, making it convenient to use.

[0029] This application has the following technical advantages: Magnetic nanoparticles were individually arranged using microfluidic technology and simultaneously subjected to dual-frequency nonlinear magnetic excitation and optical irradiation. The corresponding magnetic response signal, scattered light pulse signal, and fluorescence pulse signal were extracted. The two types of signals were correlated to fuse the magnetic and optical characteristic parameters of each magnetic nanoparticle in complex biofluids, eliminating interference from non-magnetic background impurities, and achieving multidimensional and precise identification of the free, non-specific adsorption, and targeted binding states of the particles. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of a method for detecting magnetic flow cytometry of nanoparticles according to an embodiment of this application. Detailed Implementation

[0031] This application discloses a method for detecting magnetic nanoparticles using flow cytometry. It employs microfluidic dynamics focusing technology to forcibly align discrete magnetic nanoparticles into a single-particle stream. Simultaneously, a dual-frequency alternating magnetic field is applied to excite the unique nonlinear magnetic response characteristics of the magnetic nanoparticles, and forward / lateral scattering and surface fluorescence characteristics are extracted by combining optical irradiation. By fusing the Brownian relaxation delay characteristics reflected by the magnetic signal and the specific luminescence characteristics of the optical signal, the method effectively eliminates background interference from numerous non-magnetic endogenous large biological particles in complex liquid matrices, accurately determining whether each magnetic nanoparticle is in a free state, a non-specifically encapsulated state of a protein crown, or a specific target-bound state, thus reducing detection result bias.

[0032] Reference Figure 1 A method for detecting magnetic nanoparticles by flow cytometry includes steps S1-S3.

[0033] Step S1: Collect the magnetization response signal generated by the magnetic nanoparticles in the biological liquid under test under a single-track trajectory when excited by a dual-frequency alternating magnetic field, as well as the scattered light pulse signal and fluorescence pulse signal generated by optical irradiation.

[0034] The flow cytometer used for detection comprises a sample cell, a sheath fluid cell, a microinjection pump, and a microfluidic chip. The microinjection pump drives the biological liquid and sheath fluid to be tested into the three-dimensional focusing channel inside the microfluidic chip, respectively.

[0035] For example, the cross-sectional dimensions of the three-dimensional focusing channel inside the microfluidic chip are designed at the micrometer level, with its typical width and depth controlled within a certain range. to Within a range, for example, in a preferred embodiment, it is set to width. ,depth This technique is a conventional technique in this field and will not be described in detail here.

[0036] The test biological fluid can be undiluted blood serum, plasma, or cell culture supernatant; endogenous biological vesicles can include exosomes, microvesicles, or low-density lipoproteins; the sheath fluid can be standard phosphate buffer. The test biological fluid flows in the central layer of the three-dimensional focusing channel. The sheath fluid exerts hydrodynamic compression on the test biological fluid from multiple dimensions, causing the cross-sectional area of ​​the test biological fluid stream to shrink. Because the physical size of the three-dimensional focusing channel is comparable to the particle scale, combined with the shrinkage effect of the stream cross-sectional area, the tiny physical space forces the sample flow containing magnetic nanoparticles to be compressed near the central axis of the three-dimensional focusing channel, ensuring that the magnetic nanoparticles can only flow forward in a single column.

[0037] In this embodiment, the preferred injection flow rate range for the biological liquid to be tested is as follows: to For example, it can be set to Understandably, when the injection flow rate is lower than... At this time, the laminar flow stability within the three-dimensional focusing channel is poor, leading to vibration at the hydrodynamic focusing interface; when the injection velocity is higher than... When the number of magnetic nanoparticles passing through the cross-section of the three-dimensional focusing channel per unit time is excessive, the probability of two or even multiple particles appearing simultaneously under the Poisson distribution increases, disrupting the single-row arrangement. Therefore, the injection flow rate should be controlled at... This ensures that individual magnetic nanoparticles pass sequentially through tiny three-dimensional focusing channels, guaranteeing the accuracy of subsequent single-particle event acquisition.

[0038] As the magnetic nanoparticles flow forward in a single column, they pass through the detection zone, where a dual-frequency alternating magnetic field and optical irradiation are applied to their single-column trajectory.

[0039] It is understandable that in the flow cytometer, in the detection zone downstream of the three-dimensional focusing channel, the first excitation coil and the second excitation coil arranged around the three-dimensional focusing channel generate a first alternating magnetic field and a second alternating magnetic field, respectively, to transiently magnetize the magnetic nanoparticles that pass through in sequence.

[0040] In this embodiment, the frequency of the first alternating magnetic field The range is to The amplitude is The frequency of the second alternating magnetic field The range is to The amplitude is .

[0041] When the frequency configuration is below the lower limit mentioned above, the time for a single magnetic nanoparticle to pass through the detection area is too short to cover a sufficient number of magnetic field reversal cycles; when the frequency configuration is above the upper limit mentioned above, the skin effect of the inductive element will cause severe baseline temperature drift.

[0042] To describe the transient magnetization behavior of magnetic nanoparticles under the superposition of a first alternating magnetic field and a second alternating magnetic field, the Langevin function equation is followed.

[0043] The magnetization process of magnetic nanoparticles in the superparamagnetic state exhibits typical nonlinear saturation characteristics, requiring the Langevin distribution from statistical thermodynamics to quantitatively describe the macroscopic behavior of the magnetic moment under the interplay of an external magnetic field and thermal perturbations. The corresponding calculation formula is as follows: ; In the formula, This represents the instantaneous magnetization intensity of the magnetic nanoparticles; This represents the saturation magnetization of the magnetic nanoparticles. Represents the vacuum permeability constant; This represents the equivalent magnetic moment inside a single magnetic nanoparticle. This represents the instantaneous external magnetic field strength after the superposition of the first and second alternating magnetic fields; Represents the Boltzmann constant; This indicates the absolute temperature of the biological liquid being tested.

[0044] Based on the above formula, it can be seen that the instantaneous magnetization intensity increases nonlinearly with the increase of the instantaneous external magnetic field intensity. When the instantaneous external magnetic field intensity is small, thermal disturbance dominates and the magnetic moment orientation tends to be random; as the instantaneous external magnetic field intensity gradually increases, the equivalent magnetic moment of the magnetic nanoparticles overcomes the thermal disturbance and gradually aligns with the direction of the external magnetic field, the slope gradually decreases and approaches the saturation magnetization intensity.

[0045] In practical flow cytometry scenarios, a high-intensity first alternating magnetic field is responsible for driving the magnetic nanoparticles to this nonlinear saturation region, while a low-intensity second alternating magnetic field is responsible for high-frequency sampling of this nonlinear region. The combined effect of the two inevitably leads to the generation of high-order harmonics and nonlinear mixing components in the final response signal of the magnetic nanoparticles.

[0046] Simultaneously, within the same detection area, a semiconductor laser source emits a wavelength of... A focused laser beam is vertically irradiated onto a single-track moving fluid path, causing any solid particle passing through it to undergo light scattering and potential fluorescence excitation.

[0047] At this point, when magnetic nanoparticles flow through the detection area and generate magnetic field disturbances, the TMR (Tunneling Magnetoresistance) sensor attached to the bottom of the three-dimensional focusing channel can capture the local magnetic flux change and convert it into a continuous voltage analog signal, thereby obtaining a nonlinear magnetization response signal.

[0048] Simultaneously, on the optical collection side, the photon flux at different deflection angles is detected by a PMT (Photomultiplier Tube) array.

[0049] Specifically, after isolating the excitation light background using a pre-filter, the first PMT acquires the scattered light pulse signal reflecting the outer diameter profile of the particle hydration dynamics; the second PMT acquires the fluorescence pulse signal reflecting the intensity of a specific biochemical label.

[0050] S2: Extract the magnetic characteristic parameters of the magnetization response signal, and the optical characteristic parameters of the scattered light pulse signal and the fluorescence pulse signal; correlate the magnetic characteristic parameters and the optical characteristic parameters belonging to the same magnetic nanoparticle.

[0051] For any particle queuing through the detection zone, its corresponding magnetization response signal is subjected to analog pre-amplification and bandpass filtering.

[0052] Specifically, the center frequency is set as A high-order Butterworth bandpass filter is used to filter out low-frequency flow noise and power frequency interference. Subsequently, a dual-channel digital lock-in amplification algorithm is employed to... (For example: )or As a local reference mixer carrier, the filtered signal is quadrature demodulated to extract the first mixer component with the highest signal-to-noise ratio.

[0053] From the first mixing component, for each discrete voltage pulse envelope, the highest peak value of its envelope is calculated and defined as the magnetic signal amplitude corresponding to the magnetic nanoparticle. Simultaneously, the phase delay difference between the voltage pulse signal and the local reference mixing carrier is calculated and defined as the magnetic signal phase angle corresponding to the magnetic nanoparticle.

[0054] To explain the physical mechanism of the phase angle shift of the magnetic signal, the Brownian relaxation time formula is followed.

[0055] When magnetic nanoparticles physically rotate within a viscous biological matrix to align with an external magnetic field, a physical delay occurs due to hydrodynamic resistance. Therefore, a quantitative assessment of the hysteresis contribution of the particle's encapsulation volume to its relaxation behavior is necessary. The corresponding calculation formula is as follows: ; In the formula, This represents the Brownian relaxation time caused by the overall rotation of magnetic nanoparticles in a fluid. Indicates the dynamic viscosity coefficient of the biological fluid being tested; The total hydration kinetic volume of the magnetic nanoparticles, including the surface coating layer, can be obtained based on the intensity of scattered light, particle size calibration relationship, or a preset calibration curve. Represents the Boltzmann constant; This indicates the absolute temperature of the biological liquid being tested.

[0056] Based on the above formula, it can be seen that the Brownian relaxation time and the hydration kinetic volume exhibit a strictly proportional linear increasing relationship. In a real-world scenario where the viscosity and temperature of the biological fluid under test remain constant, when a large amount of serum protein is adsorbed onto the surface of the free magnetic nanoparticles to form a protein crown, or when large biological vesicles such as exosomes are specifically bound, their effective hydration kinetic volume will expand geometrically. This expansion directly causes the particles to "rotate slowly" in the alternating magnetic field, manifested as an increase in the Brownian relaxation time. This is reflected in the first mixing component extracted by lock-in amplification, which exhibits a more negative magnetic signal phase angle shift.

[0057] For example, assuming absolute temperature In the environment, the dynamic viscosity coefficient of a certain biological liquid to be tested The initial hydration kinetic volume of a free magnetic nanoparticle. Substituting into the formula, we calculate the Brownian relaxation time in the free state. When the magnetic nanoparticle binds to a large endogenous biological vesicle, its effective hydration kinetic volume increases to [value missing]. Substituting back into the formula, the combined Brownian relaxation time... This extension of relaxation time will manifest as a significant phase lag in the first mixing component.

[0058] To address the amplitude dependence of the first mixing component, a Neel relaxation time formula is constructed, and the Neel relaxation theory is used to explain the correlation between the magnetic signal amplitude and the magnetic core structure.

[0059] Understandably, the magnetic moment reversal of the lattice within magnetic nanoparticles is also constrained by the material's anisotropic barrier, requiring the analytical extraction of signal characteristics related to the particle's rigid physical magnetic core using the NieR equation. The corresponding calculation formula is as follows: ; In the formula, This represents the Niehr relaxation time caused by the rotation of the magnetic moment inside the magnetic nanoparticle relative to the crystal lattice. This represents the characteristic time constant that reflects the inherent frequency of flipping attempts of the material; This represents the magnetic anisotropy constant of magnetic nanomaterials; This represents the volume of the purely physical magnetic core of the magnetic nanoparticles. Represents the Boltzmann constant; This indicates the absolute temperature of the biological liquid being tested.

[0060] Based on the above formula, it can be seen that the Niehr relaxation time increases exponentially with the increase of the physical magnetic core volume. In flow cytometry, since the surface biochemical modifications of magnetic nanoparticles (such as conjugated antibodies, nucleic acids, etc.) do not change their core lattice size, the magnetic signal amplitude can specifically reflect the inherent mass distribution of magnetic nanoparticles on a macroscopic scale, regardless of whether their surface is bare or bound to exosomes.

[0061] For example, suppose a magnetic nanoparticle has a characteristic time constant. The ratio of the material's anisotropic energy barrier to its thermal energy. Substituting into the formula, we calculate the NieR relaxation time. .

[0062] Furthermore, in the optical signal branch, high-speed transimpedance amplifiers are used to perform current-to-voltage conversion on the discrete scattered light pulse signals and fluorescence pulse signals. A moving average filtering algorithm is used for baseline restoration to dynamically eliminate low-frequency drift, and a peak search algorithm is used to calculate the peak intensities of the scattered light pulse signal and the fluorescence pulse signal, respectively.

[0063] After obtaining the magnetic characteristic parameters (magnetic signal amplitude and magnetic signal phase angle) and optical characteristic parameters (peak intensity of scattered light pulse signal and peak intensity of fluorescence pulse signal), the two types of characteristic parameters belonging to the same magnetic nanoparticle are correlated.

[0064] Specifically, a single-particle event extraction time window is set. For the co-located photomagnetic detection area, the single-particle event extraction time window is set to [value missing]. .

[0065] Specifically, the same internal global clock of the multi-channel synchronous data acquisition card is used to timestamp all discrete signals. The peak value of the voltage pulse in the extracted magnetization response signal is used as the timestamp. This serves as an absolute time reference. Based on this absolute time reference, fixed time tolerances are extended forward and backward (e.g., set to the pulse width through which particles flow). ), thus generating the extraction interval as The single-particle event extraction time window.

[0066] If the pulse envelope of the first mixing component and the pulse peak of the optical sensor are detected simultaneously within any single-particle event extraction time window, the magnetic signal amplitude, magnetic signal phase angle, scattered light pulse signal intensity and fluorescence pulse signal intensity within the current time window are combined into a high-dimensional feature vector.

[0067] S3: Based on the correlation results and the preset conditions, classify the magnetic nanoparticles into states, statistically analyze the classification results, and output the distribution characteristics of each state of the nanoparticles.

[0068] Specifically, the preset conditions mainly include the following types, and different preset conditions will be identified as different status categories: 1. If the magnetic signal amplitude is higher than the preset background noise threshold, and the fluorescence pulse signal intensity is lower than the preset fluorescence threshold, and the magnetic signal phase angle is within the unoffset range (greater than or equal to...), and less than or equal to If the magnetic nanoparticle is within a certain range, it is determined that the magnetic nanoparticle belongs to a "free-floating magnetic nanoparticle state." A free-floating event represents a pristine probe in the sample solution that has not undergone any bioadhesion.

[0069] 2. If the magnetic signal amplitude is higher than the preset background noise threshold, and the fluorescence pulse signal intensity is lower than the preset fluorescence threshold, but the magnetic signal phase angle falls into the preset second offset interval (greater than or equal to...), and less than or equal to If the event is within the range of ), then the event is determined to be in a "non-specific vesicle adsorption state".

[0070] For example, non-specific adsorption, i.e., magnetic nanoparticles physically adsorb non-targeted albumin or ordinary detached microvesicles in serum through electrostatic or hydrophobic interactions. Such particles expand in volume and produce a phase shift, but do not have targeted fluorescent labeling.

[0071] 3. If the magnetic signal amplitude is higher than the preset background noise threshold, and the fluorescence pulse signal intensity is higher than the preset fluorescence threshold, and the magnetic signal phase angle is within the first offset interval (greater than or equal to...), and less than If the fluorescence intensity falls within a certain range, the event is determined to be in a "specific target binding state." The fluorescence threshold is determined based on the optical background signal calibration of the negative control sample. Before actual detection, a negative control sample is pre-collected, such as magnetic nanoparticles or blank biological matrix that have not bound to the target fluorescent probe. The fluorescence channel continuously outputs waveforms when the sample passes through the detection area. The average intensity value and standard deviation of the fluorescence background signal corresponding to the negative control sample are extracted, and the average intensity value is added to a predetermined factor, preferably 3 times the standard deviation, and set as the fluorescence intensity threshold.

[0072] For example, a magnetic nanoparticle coupled with an anti-PD-L1 fluorescent antibody successfully recognized and bound to tumor exosomes rich in the corresponding antigen on its surface, thereby simultaneously triggering magnetic phase shift and specific fluorescence.

[0073] 4. If the magnetic signal amplitude is equal to or lower than the background noise threshold, but the intensity of the scattered light pulse signal shows an obvious peak, then the event is determined to be a "non-magnetic background impurity state" caused by endogenous biological vesicles, and it is removed from the counting statistics.

[0074] It is understandable that for detection events where the magnetic signal amplitude, fluorescence pulse signal intensity, and magnetic signal phase angle do not simultaneously meet any of the above state classification conditions, the corresponding magnetic nanoparticle state cannot be determined. Therefore, such detection events are marked as invalid detection events and removed in subsequent state distribution statistics to avoid abnormal detection signals affecting the statistical results.

[0075] It is understood that the above-mentioned unoffset interval, first offset interval, and second offset interval range are only examples, and those skilled in the art can adjust them according to the experimental calibration results during the specific implementation process.

[0076] In other embodiments, considering that the fused high-dimensional feature vector contains at least four independent features such as magnetic signal amplitude, magnetic signal phase angle, scattered light pulse signal intensity and fluorescence pulse signal intensity, a pre-trained machine learning classification model can be used to replace the fixed logical threshold discriminant.

[0077] For example, the high-dimensional feature vector can be input into a machine learning classification model, which includes, but is not limited to, a random forest algorithm or a convolutional neural network (CNN) model.

[0078] By using nonlinear mapping and feature cross-analysis within the model, the predicted probability of whether the particle event belongs to a free state, a non-specific vesicle adsorption state, or a specific target binding state is output.

[0079] In constructing and optimizing the machine learning classification model, the system collects high-dimensional feature samples of magnetic nanoparticles in known states as a training set. When receiving user feedback or introducing expert calibration data for model iteration, the system strictly uses the posterior probability as the core variable for adjusting model parameters. That is, based on the prior distribution of the new input feedback samples, the system calculates and updates the posterior probability distribution of each state category in real time, thereby driving the dynamic fine-tuning of the classification decision boundary. This ensures that the system maintains a sharp and precise particle classification accuracy when dealing with complex biological liquid matrices from different batches.

[0080] Finally, the count values ​​of various events in the current biological liquid under test are accumulated, and the distribution characteristics of various states of nanoparticles are output.

[0081] Based on the classification results, the target binding double positivity rate of the output samples is calculated.

[0082] For any given batch of tests, the double positive rate of targeted binding is equal to the count value of "specific target binding state" divided by the total number of magnetic events that combine "free magnetic nanoparticle state", "non-specific vesicle adsorption state" and "specific target binding state".

[0083] Simultaneously, by integrating the injection flow rate and sampling time in step S1, the total flow volume of the analytical liquid is calculated, and the absolute volume concentration data of the target complex is output. A histogram of hydration kinetic size distribution with magnetic signal amplitude as the mapping correlation is generated, as well as a two-dimensional scatter plot matrix with magnetic signal phase angle as the horizontal axis and fluorescence pulse signal intensity as the vertical axis, which intuitively shows the true distribution state of target particles in complex biological matrix.

[0084] It is understandable that this embodiment fully leverages the complementary discrimination capability of magneto-optical detection at the single-particle level, reducing the false positive defects of optical methods and the heterogeneity masking problem of macroscopic magnetic spectroscopy methods.

[0085] This application also discloses a nanoparticle magnetic flow cytometry detection system, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a nanoparticle magnetic flow cytometry detection method according to this application.

[0086] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0087] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for detecting the magnetic properties of nanoparticles by flow cytometry, characterized in that, The magnetization response signal generated by magnetic nanoparticles in the biological liquid under test under a single-track trajectory and excited by a dual-frequency alternating magnetic field, as well as the scattered light pulse signal and fluorescence pulse signal generated by optical irradiation, were collected. Extract the magnetic characteristic parameters of the magnetization response signal, and the optical characteristic parameters of the scattered light pulse signal and the fluorescence pulse signal; The magnetic characteristic parameters belonging to the same magnetic nanoparticle are correlated with the optical characteristic parameters; Based on the correlation results and the preset conditions, the magnetic nanoparticles are classified into states, the classification results are statistically analyzed, and the distribution characteristics of each state of the nanoparticles are output.

2. The method for detecting magnetic nanoparticles by flow cytometry according to claim 1, characterized in that, The magnetic characteristic parameters include the magnetic signal amplitude and the magnetic signal phase angle; the optical characteristic parameters include the peak intensity of the scattered light pulse signal and the peak intensity of the fluorescence pulse signal.

3. The method for detecting magnetic nanoparticles by flow cytometry according to claim 2, characterized in that, Extracting the magnetic characteristic parameters of the magnetization response signal includes: The magnetization response signal is bandpass filtered; Using the high-order mixing frequency formed by the dual-frequency alternating magnetic field as the local reference mixing carrier, a dual-channel digital phase-locked amplification algorithm is used to orthogonally demodulate the filtered magnetization response signal, resulting in the first mixing component with the highest signal-to-noise ratio. The highest peak value of the voltage pulse envelope corresponding to the first mixing component is extracted as the magnetic signal amplitude, and the phase delay difference between the voltage pulse envelope and the local reference mixing carrier is calculated as the magnetic signal phase angle.

4. The method for detecting magnetic nanoparticles by flow cytometry according to claim 3, characterized in that, The dual-frequency alternating magnetic field includes a frequency of The first alternating magnetic field and frequency are The second alternating magnetic field; the higher-order mixing frequency is configured as follows: or .

5. The method for detecting magnetic nanoparticles by flow cytometry according to claim 2, characterized in that, The steps for obtaining the peak intensity of the scattered light pulse signal and the peak intensity of the fluorescence pulse signal in the optical characteristic parameters include: The scattered light pulse signal and the fluorescence pulse signal are processed using a moving average filtering algorithm to dynamically eliminate low-frequency drift; The peak intensity of the scattered light pulse signal and the peak intensity of the fluorescence pulse signal after baseline recovery are calculated using a peak search algorithm.

6. The method for detecting magnetic nanoparticles by magnetic flow cytometry according to claim 2, characterized in that, Associating the magnetic characteristic parameters with the optical characteristic parameters belonging to the same magnetic nanoparticle includes: Set a time window for single-particle event extraction; If the voltage pulse envelope of the magnetization response signal and the pulse peak of the scattered light pulse signal or the fluorescence pulse signal are extracted simultaneously within any single-particle event extraction time window, then the magnetic signal amplitude, the magnetic signal phase angle, the peak intensity of the scattered light pulse signal and the peak intensity of the fluorescence pulse signal within the current single-particle event extraction time window are merged into a high-dimensional feature vector to complete the association of the two types of feature parameters.

7. The method for detecting magnetic nanoparticles by flow cytometry according to claim 2, characterized in that, Based on the correlation results and pre-defined conditions, the magnetic nanoparticles are classified into states, including the identification step of specific target binding states: If the amplitude of the magnetic signal in the high-dimensional feature vector is higher than a preset background noise threshold, and the peak intensity of the fluorescence pulse signal is higher than a preset fluorescence threshold, and the phase angle of the magnetic signal falls into a preset first offset interval, then it is determined to belong to the specific target binding state.

8. The method for detecting magnetic nanoparticles by flow cytometry according to claim 7, characterized in that, The magnetic nanoparticles are classified into states based on the correlation results and preset conditions. The classification also includes steps for identifying free-floating events and non-specific adsorption events. If the amplitude of the magnetic signal is higher than the background noise threshold, and the peak intensity of the fluorescence pulse signal is equal to or lower than the fluorescence threshold, and the phase angle of the magnetic signal falls into the preset non-offset interval, then it is determined to belong to the state of free magnetic nanoparticles. If the amplitude of the magnetic signal is higher than the background noise threshold, and the peak intensity of the fluorescence pulse signal is equal to or lower than the fluorescence threshold, and the phase angle of the magnetic signal falls into the preset second offset interval, then it is determined to belong to the non-specific vesicle adsorption state. Wherein, any absolute value of phase shift within the second offset interval is greater than any absolute value of phase shift within the non-offset interval, and any absolute value of phase shift within the first offset interval is greater than any absolute value of phase shift within the second offset interval.

9. The method for detecting magnetic nanoparticles by flow cytometry according to claim 8, characterized in that, The magnetic nanoparticles are classified into states based on the correlation results and preset conditions, and the process also includes a background impurity removal step. If the amplitude of the magnetic signal is equal to or lower than the background noise threshold, and the peak intensity of the scattered light pulse signal is higher than the preset scattered light threshold, then it is determined to belong to the non-magnetic background impurity state.

10. A magnetic flow cytometry detection system for nanoparticles, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement a method for detecting magnetic nanoparticles according to any one of claims 1-9.

Images