Biological activity detection method and device based on deep learning
By developing a deep learning-based bioactivity detection method and device, the problems of insufficient sensitivity and poor device adaptability of traditional fluorescence sensing detection have been solved, achieving high-sensitivity and high-specificity detection, which is applicable to fields such as environmental monitoring and food testing.
Patent Information
- Application Number
- CN202610124817.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-01-29
- Publication Date
- 2026-04-10
AI Technical Summary
Existing fluorescence sensing detection methods suffer from insufficient detection sensitivity, poor identification specificity, non-standard sample processing procedures, poor adaptability of detection devices, and weak anti-interference ability.
A deep learning-based bioactivity detection method is adopted, in which intermediates are prepared by Suzuki coupling reaction and active units are introduced to inhibit aggregation-induced quenching effect. Data analysis is performed by combining linear discriminant analysis algorithm. The detection device integrates a multi-functional linkage mechanism and a self-adjusting protection mechanism to achieve precise positioning, multi-angle adjustment and anti-interference.
It significantly improves detection sensitivity and specificity, enhances the standardization of detection procedures and data reliability, optimizes device adaptability and protection performance, and is suitable for fields such as environmental monitoring and food testing.
Smart Images
Figure CN121830610A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of biological detection, in particular to a biological activity detection method and device based on deep learning. BACKGROUND
[0002] In the field of biological detection, the demand for detecting pollutants such as perfluoro and polyfluoroalkyl substances (PFAS) and bioactive ingredients is increasingly urgent. Fluorescence sensing method has become one of the mainstream detection technologies in this field due to its advantages of simple operation, rapid response, high sensitivity, etc. In the prior art, fluorescence sensing detection schemes are mostly based on traditional small organic molecule probes or conventional conjugated polymer materials. The core principle is that the probe interacts specifically with the target and causes a change in the fluorescence signal, and then combined with spectral analysis methods to realize the qualitative and quantitative detection of the target. The detection device is designed with a single clamping positioning function as the core, which fixes the position of the sample pool to ensure the stability of the spectral acquisition process, and is widely used in environmental monitoring, food detection and other fields.
[0003] However, the existing fluorescence sensing detection method has significant technical limitations. First, traditional conjugated polymer probes are easily affected by aggregation-induced quenching (ACQ) effect, resulting in decreased fluorescence quantum yield, insufficient detection sensitivity, and difficulty in meeting the precise detection needs of nM-level perfluorooctane sulfonic acid (PFOS) and other low-concentration target pollutants. Second, the functional modification scheme of the probe is single, and the recognition specificity of different chain lengths and similar structures of the target is poor, which cannot effectively distinguish the specific components of similar pollutants, limiting the refinement degree of the detection results. Third, the sample processing procedure lacks standardized design, and problems such as incomplete impurity removal and poor consistency of parallel samples are common, further reducing the reliability of the detection data.
[0004] At the level of the supporting detection device, the existing equipment also has many defects. First, the clamping mechanism of the traditional device is mostly a rigid fixed structure, which cannot adapt to different specifications of the sample pool or the detection object, and lacks elastic buffer design, which easily causes damage to the sample pool or the detection object due to excessive clamping force. Second, the device does not have a posture adjustment function, and the sample can only be detected at a single angle, making it difficult to achieve full-range collection of fluorescence signals, resulting in signal blind areas in the detection results. Third, the device has weak anti-interference ability, and external mechanical shaking or temperature fluctuations can easily affect the positioning accuracy and signal stability of the sample, and the sealing and protection performance is insufficient, which cannot effectively isolate the interference of the external environment on the fluorescence signal, further reducing the detection accuracy.
[0005] The above problems jointly restrict the practicability and popularization value of the biological activity detection technology, therefore, an integrated solution with high-sensitivity detection method and multifunctional adaptive device is urgently needed. Based on this, the application provides a biological activity detection method and device based on deep learning. SUMMARY
[0006] 1. The technical problem to be solved by the application is: Therefore, the technical problem to be solved by the application is to provide a biological activity detection method and device based on deep learning, so as to solve the following technical problems existing in the prior art: (1) The existing fluorescence sensing detection method has limitations, the traditional conjugated polymer probe is easily affected by the aggregation-induced quenching (ACQ) effect, resulting in insufficient detection sensitivity, the functional modification scheme of the probe is single, the recognition specificity for different chain length and structure similar target objects is poor, and the sample processing procedure lacks standardized design, which affects the reliability of the detection data, and it is difficult to meet the accurate and fine detection requirements of low-concentration target pollutants.
[0007] (2) The matching detection device has defects, the clamping mechanism of the traditional device is a rigid fixed structure, which cannot adapt to different specifications of detection objects and is easy to cause damage, does not have the posture adjusting function, resulting in signal blind area in detection, and has weak anti-interference ability and insufficient sealing protection performance, so that the external environmental factors easily affect the sample positioning accuracy and signal stability, and reduce the detection precision.
[0008] 2. Technical scheme To achieve the above object, the application provides the following technical scheme: a biological activity detection method and device based on deep learning, comprising the following steps: S1, preparing for detection, including preparation of related reagents and materials and debugging of detection instruments; S2, collecting and processing the sample to be detected to obtain a sample liquid suitable for detection; S3, making the processed sample liquid react with the conjugated polymer related reagent, and collecting the related signals in the reaction process; S4, analyzing the collected signals, and determining the detection result combined with the analysis result.
[0009] As a preferred, the detection is fluorescence sensing detection, which is used for detecting pollutants, the reagent and material preparation in S1 includes synthesizing a target conjugated polymer probe, preparing an intermediate through Suzuki coupling reaction, and then introducing an active unit through functional modification to inhibit the aggregation-induced quenching (ACQ) effect; a series of concentration gradient standard detection liquid, buffer and blank control liquid are prepared.
[0010] As preferred, the active unit includes triphenylamine TPA, methyl imidazole; the standard detection liquid is PFOA / PFOS standard liquid, and the concentration range is 0-36 μM.
[0011] As preferred, the sample treatment in S2 includes removing impurities in the sample to be detected by centrifugation, the centrifugal speed is 8000-10000 r / min, the centrifugal time is 10 min, and the supernatant is taken as the detection mother liquor; according to the pre-experiment result, the detection mother liquor is diluted to an appropriate concentration range with a buffer, and the volume is adjusted to the standard sample pool scale, and three groups of parallel samples are set.
[0012] As preferred, the reaction and signal collection in S3 include adding a quantitative conjugated polymer probe to the treated sample liquid, placing it in a constant temperature oscillator for 10-15 min, the oscillation temperature is 25℃, and the rotation speed is 150 r / min; the sample pool after reaction is placed into a fluorescence spectrophotometer, 370 nm is used as the excitation wavelength, the fluorescence emission spectrum in the range of 400-700 nm is scanned, the characteristic peak intensity and peak shape change are recorded, and the fluorescence color of the sample under the ultraviolet lamp is collected by a portable device, which is converted into RGB digital information by an APP; the above operation is performed on the blank control liquid at the same time, and the fluorescence signal is collected as a reference value.
[0013] As preferred, the data analysis and result determination in S4 include calculating the fluorescence intensity normalization value of the sample, drawing a concentration-fluorescence intensity standard curve, determining the detection limit (LOD) by linear fitting, and the linear fitting correlation coefficient R²≥0.99; according to the type of sample fluorescence signal, the concentration of the target pollutant is determined combined with the standard curve, different chain length pollutants are distinguished by linear discriminant analysis (LDA) algorithm, and qualitative and semi-quantitative detection results are output.
[0014] A biological activity detection device based on deep learning, comprising a detection table, a sealing cover is rotatably installed on one side of the upper surface of the detection table, a power module, an X-ray emitting module, a heat dissipation module and a control module are respectively installed in the detection table, the power module is used for power supply of the X-ray emitting module, the heat dissipation module is used for heat dissipation inside the detection table, and a multifunctional linkage mechanism and a self-adjusting protection mechanism are further included. The multifunctional linkage mechanism is arranged on the upper surface of the detection table, and is used for positioning and protection of the detection object. The self-adjusting protection mechanism is arranged in the multifunctional linkage mechanism, and is used for linkage control of the multifunctional linkage mechanism.
[0015] As preferred, the multifunctional linkage mechanism comprises a sealing bin mounted on the upper surface of the detection table, which can form a sealed detection bin after being closed with the sealing cover, and an annular table is fixedly installed in the sealing bin, a sliding gear ring is slidingly installed on the inner surface of the annular table, a rotating gear is circumferentially and rotatably installed on one side of the sealing bin close to the sliding gear ring, the tooth surface of the rotating gear is engaged on the sliding gear ring, a connecting gear is fixedly installed on the upper surface of the rotating gear, the tooth surface of the connecting gear is engaged with a sliding gear rack, the sliding gear rack is slidingly installed in the annular table, and the sliding gear rack is arranged above the sliding gear ring.
[0016] As preferred, the self-adjusting protection mechanism comprises an arc-shaped guard plate, a guide rail gear belt is transmissionally installed in the middle of the arc-shaped guard plate, tooth grooves are uniformly formed in the middle inner surface of the guide rail gear belt, guide slides are symmetrically slidingly installed in the sliding gear rack, one end of the guide slide is fixedly installed on the arc-shaped guard plate, an auxiliary spring is sleeved on the outer surface of the other end of the guide slide, one end of the auxiliary spring is fixedly installed on the sliding gear rack, the other end of the auxiliary spring is fixedly installed on the guide slide, a guide groove is formed in the middle of one end of the sliding gear rack close to the arc-shaped guard plate, a driving gear is arranged in the guide groove, the tooth surface of the driving gear is engaged on the guide rail gear belt, connecting shafts are fixedly installed at both ends of the driving gear, the outer surfaces of both ends of the connecting shafts are movably installed in the guide groove, and both ends of the connecting shafts are rotatably installed on the guide slide.
[0017] Compared with the prior art, the biological activity detection method and device based on deep learning provided by the present application have the following beneficial effects: I. The detection sensitivity and specificity are significantly improved; The present scheme prepares intermediates through Suzuki coupling reaction and introduces triphenylamine TPA, methyl imidazole and other active units, effectively inhibits the aggregation-induced quenching (ACQ) effect of traditional conjugated polymer probes, greatly improves the fluorescence quantum yield, and can realize precise detection of nM-level low-concentration pollutants such as PFOS; combined with linear discriminant analysis (LDA) algorithm, it can accurately distinguish similar pollutants with different chain lengths and structures, solve the problem of insufficient recognition specificity of traditional detection methods, and realize fine detection of qualitative and semi-quantitative.
[0018] II. The detection process is standardized and the data reliability is enhanced; The sample processing link adopts the standardized impurity removal process of 8000-10000r / min centrifugation for 10 minutes, is matched with buffer dilution and constant volume, and 3 sets of parallel samples are set, which effectively reduces impurity interference and experimental error; in the data analysis stage, the detection limit is determined through fluorescence intensity normalization processing and linear fitting (R²≥0.99), a whole-process standardization system from sample preparation to result determination is formed, the reliability and repeatability of the detection data are significantly improved, and the defects of non-standard sample processing in the prior art are made up.
[0019] III. Detection device adaptability and protection performance optimization; The device integrates a multifunctional linkage mechanism and a self-adjusting protection mechanism. Through the meshing transmission of the sliding tooth ring, the rotating tooth, and the sliding tooth bar, precise positioning of different specifications of sample pools can be achieved. The elastic buffer structure formed by the arc-shaped guard plate and the auxiliary spring avoids sample damage caused by rigid clamping. At the same time, the cooperation of the guide rail tooth belt and the driving tooth supports multi-angle posture adjustment of the sample, eliminates the signal acquisition blind area, and solves the problems of poor adaptability, lack of protection, and posture adjustment function of traditional devices.
[0020] IV. Anti-interference ability and detection scene adaptability improvement; The sealed bin and the sealed cover form a closed detection bin, which can effectively isolate external mechanical shaking, temperature fluctuations, and environmental light interference on the fluorescence signal. The detection method combines the spectrum acquisition of the fluorescence spectrophotometer and the RGB signal conversion of the portable device, which not only meets the laboratory precision detection demand, but also has the potential for on-site rapid detection. The device structure with strong adaptability broadens the application scenarios in environmental monitoring, food detection, and other fields, and takes into account the detection accuracy and practicality. BRIEF DESCRIPTION OF DRAWINGS
[0021] Figure 1 is a schematic diagram of the three-dimensional structure of the present application; Figure 2 is an auxiliary schematic diagram of the three-dimensional structure of the present application; Figure 3 is a schematic diagram of the three-dimensional structure of the present application Figure 2 is an enlarged view of position A in the present application; Figure 4 is a schematic diagram of the three-dimensional structure of the present application Figure 5 is an auxiliary schematic diagram of the three-dimensional structure of the present application Figure 6 is a schematic diagram of the three-dimensional structure of the present application Figure 5 is an enlarged view of position B in the present application; Figure 7 is a flowchart of a biological activity detection method based on deep learning.
[0022] In the figure: 1, detection table; 11, sealed cover; 12, power module; 13, X-ray emission module; 14, heat dissipation module; 15, control module; 2, multifunctional linkage mechanism; 21, sealed bin; 22, ring table; 23, sliding tooth ring; 24, rotating tooth; 25, connecting tooth; 26, sliding tooth bar; 3, self-adjusting protection mechanism; 31, arc-shaped guard plate; 32, guide rail tooth belt; 33, driving tooth; 34, guide slide; 35, guide groove; 36, connecting shaft; 37, auxiliary spring. DETAILED DESCRIPTION
[0023] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of protection of the present application.
[0024] The present application will be further described in detail below according to the drawings and embodiments.
[0025] Embodiment 1, please refer to Figures 1 to 7 The present application will be further described in detail below according to the drawings and embodiments. To solve the problems mentioned in the technical solutions, the present application provides a biological activity detection method and device based on deep learning, including a detection table 1. A sealing cover 11 is rotatably installed on one side of the upper surface of the detection table 1, and a power supply module 12, an X-ray emitting module 13, a heat dissipation module 14 and a control module 15 are embedded inside the sealing cover 11. The power supply module 12 provides stable power supply for the X-ray emitting module 13, the heat dissipation module 14 is responsible for exporting the heat inside the detection table 1 to ensure stable operation of the equipment, and the control module 15 coordinates the collaborative work of various components. The device also integrates a multifunctional linkage mechanism 2 and a self-adjusting protection mechanism 3 to form an integrated detection and protection system.
[0026] The multifunctional linkage mechanism 2 is assembled on the upper surface of the detection table 1, and the core function is to realize accurate positioning and clamping protection of the detected object, and to ensure the stability of the object posture during the detection process. The self-adjusting protection mechanism 3 is integrated inside the multifunctional linkage mechanism 2, which can adapt to different specifications of the detected object through linkage control, and has the functions of buffering protection and posture adjustment.
[0027] Specifically, as shown in Figures 1 to 3 The multifunctional linkage mechanism 2 includes a sealed bin 21, which is fixedly installed on the upper surface of the detection table 1 and can form a sealed detection space after being closed with the sealing cover 11, effectively isolating external interference and ensuring detection safety. An annular table 22 is fixedly arranged inside the sealed bin 21, and a sliding gear ring 23 is slidingly embedded on the inner surface of the annular table 22. A rotating gear 24 is rotatably installed in the circumferential direction of one side of the sealed bin 21 close to the sliding gear ring 23, and the tooth surface of the rotating gear 24 is in meshing transmission with the sliding gear ring 23. A connecting gear 25 is fixedly connected to the upper surface of the rotating gear 24, and the connecting gear 25 is in meshing transmission with a sliding gear strip 26, which is slidingly installed inside the annular table 22 and located above the sliding gear ring 23.
[0028] Further, the middle of the gear 24 is equipped with a micro drive motor and a matching controller, and the micro drive motor is driven to rotate by the control module 15, so as to drive the gear 24 to rotate synchronously. By the meshing relationship of the gear 24, the sliding gear ring 23 and the connecting gear 25, the multiple sliding racks 26 can be driven to move towards or in the opposite direction synchronously, so as to realize the clamping and loosening of the detected object. The multifunctional linkage mechanism 2 can quickly position the detected object in the center, and significantly reduce the problem of detection precision reduction caused by the deviation of the object in the detection process.
[0029] Specifically, as shown in the figure, Figures 4 to 6 The self-adjusting protection mechanism 3 includes an arc-shaped guard plate 31, a guide rail gear belt 32 is transmissionally arranged in the middle of the arc-shaped guard plate 31, the inner surface of the middle of the guide rail gear belt 32 is uniformly provided with a meshing groove, which is used for cooperating with the transmission structure to realize the posture adjustment. The guide sliding 34 is symmetrically and slidingly arranged in the sliding rack 26, one end of the guide sliding 34 is fixedly connected with the arc-shaped guard plate 31, and the other end of the guide sliding 34 is externally sleeved with an auxiliary spring 37; the two ends of the auxiliary spring 37 are respectively fixed to the sliding rack 26 and the guide sliding 34, so as to form an elastic buffer structure. The middle of one end of the sliding rack 26 close to the arc-shaped guard plate 31 is provided with a guide groove 35, a driving gear 33 is arranged in the guide groove 35, the tooth surface of the driving gear 33 is meshed with the meshing groove of the guide rail gear belt 32, and the two ends of the driving gear 33 are fixedly connected with a connecting shaft 36; the two ends of the connecting shaft 36 are movably embedded in the guide groove 35, and the connecting shaft 36 is rotationally connected with the guide sliding 34.
[0030] Under the cooperative control of the multifunctional linkage mechanism 2 and the self-adjusting protection mechanism 3, when the multifunctional linkage mechanism 2 drives the sliding rack 26 to clamp and stretch, the arc-shaped guard plate 31 can quickly adhere to the surface of the detected object to realize clamping. At the same time, the guide rail gear belt 32 is provided with rubber pads at both ends of the structure, which can effectively reduce the damage of mechanical clamping to the detected object; the meshing groove in the middle of the guide rail gear belt 32 and the driving gear 33 form stable transmission, the guide rail gear belt 32 can be driven to transmit in the arc-shaped guard plate 31 by rotating the driving gear 33, so as to realize the multi-angle rotating adjustment of the detected object, and meet the demand of omnidirectional detection.
[0031] It should be emphasized that the scheme not only realizes the clamping positioning function of the detected object, but also completes the posture rotating adjustment through the cooperation of the guide rail gear belt 32 and the driving gear 33; under the adaptive action of the guide groove 35 and the driving gear 33, in combination with the limiting and guiding of the guide sliding 34, the dynamic shockproof effect can be realized, and the influence of external mechanical shaking on the detection precision can be effectively weakened. In addition, the linkage structure of the guide sliding 34 and the auxiliary spring 37 can provide elastic buffer for the clamping process, so as to avoid the damage of rigid clamping to the detected object. The rotation of the driving gear 33 is driven by the servo motor arranged on the guide sliding 34, and the sliding groove is arranged on the upper surface of the sliding rack 26 to provide guiding support for the movement of the servo motor with the guide sliding 34, so as to ensure the smoothness of the linkage of the mechanism.
[0032] Embodiment 2, based on embodiment 1 but with differences, a deep learning-based bioactivity detection method and device proposed by the present application is described below in conjunction with specific examples and drawings, and the specific content is as follows: The fluorescence sensing method detects conjugated polymers, taking PFAS detection as an example. The method is based on the aggregation-induced emission AIE and the Foster resonance energy transfer FRET mechanism of conjugated polymers, and is suitable for rapid, high-sensitivity on-site detection. The core process is as follows: 1. Preparation stage; reagent and material preparation: synthesis of target conjugated polymer probes such as FTD-MI and FTD-C8-MI, preparation of intermediates by Suzuki coupling reaction, and introduction of active units such as triphenylamine TPA and methyl imidazole through functional modification to inhibit the aggregation-induced quenching ACQ effect and strengthen the binding capacity with the detection target; preparation of a series of concentration gradient standard detection solutions such as PFOA / PFOS standard solution, concentration range 0-36 μM, buffer to maintain system stability, and blank control solution solvent system without target pollutants.
[0033] Instrument debugging: prepare the fluorescence spectrophotometer, 365 nm ultraviolet excitation light source, portable detection device containing sample cell, and smartphone APP acquisition module, calibrate the instrument wavelength accuracy absorption peak 380 nm, 465 nm, and emission peak 510-540 nm to ensure the accuracy of fluorescence signal acquisition; debug the APP RGB signal conversion function to establish the corresponding relationship between fluorescence intensity and digital signal.
[0034] 2. Sample processing stage; sample collection: select the sample to be detected such as water sample and food matrix extract, remove impurities by centrifugation at 8000-10000 r / min for 10 min, and take the supernatant as the detection mother liquor to avoid interference of particulate matter with the fluorescence signal.
[0035] Sample dilution and constant volume: according to the pre-experiment results, dilute the detection mother liquor with buffer to the appropriate concentration range to ensure that the target pollutant concentration is within the response range of the polymer probe, and set the constant volume to the standard sample cell scale, while setting 3 groups of parallel samples to reduce experimental error.
[0036] 3. Detection and signal acquisition stage; Probe and sample reaction: add a constant amount of conjugated polymer probe such as FTD-MI to the treated sample solution with a final concentration of 50 μM, place it in a constant temperature oscillator at 25°C, rotate at 150 r / min for 10-15 min, make the probe and target pollutants fully combine to form a complex through electrostatic and hydrophobic interaction, and trigger the fluorescence signal change.
[0037] Signal collection: Put the sample cell after reaction into the fluorescence spectrophotometer, take 370 nm as the excitation wavelength, scan the fluorescence emission spectrum in the range of 400-700 nm, record the characteristic peak intensity such as the 540 nm peak of FTD-MI, the 510 nm peak of FTD-C8-MI and the peak shape change; at the same time, use the device of Example 1 to collect the fluorescence color of the sample under the ultraviolet lamp, convert it into RGB digital information through APP, and keep the original data.
[0038] Blank control experiment: Perform the above operation on the blank control liquid synchronously, collect the fluorescence signal as the baseline value, and use it to deduct the background interference.
[0039] 4. Data analysis and result determination stage Data processing: Calculate the fluorescence intensity normalization value of the sample, draw the concentration-fluorescence intensity standard curve, and determine the detection limit by linear fitting. For example, the detection limit of FTD-MI for PFOS can be as low as 0.64 nM.
[0040] Result determination: According to the type of sample fluorescence signal, determine the concentration of target pollutants in combination with the standard curve; through linear discriminant analysis (LDA) algorithm, distinguish pollutants of different chain lengths, and output qualitative and semi-quantitative detection results.
[0041] Example 3, which is different from Example 2, takes a stretchable semiconductor polymer as an example for structure and performance characterization and detection of the conjugated polymer; This method focuses on the correlation between the condensed state structure, molecular chain movement and electrical properties of the conjugated polymer, and uses in-situ variable temperature / stretching grazing incidence X-ray diffraction, atomic force microscopy and other technologies. The process is as follows: 1. Sample preparation stage Polymer film preparation: Dissolve the conjugated polymer in a suitable solvent (such as chloroform, toluene), prepare a uniform film on the substrate by spin coating, control the thickness to be 100-200 nm, and place it in a vacuum drying box (60°C, 2h) to remove residual solvents and avoid film defects.
[0042] Device assembly (for electrical property detection): Take the polymer film as the semiconductor charge transport layer, combine metal nanoelectrodes and elastomer dielectric layers to prepare stretchable organic field effect transistors, and ensure good contact between the electrodes and the film without virtual connection phenomenon.
[0043] 2. Structure characterization and detection stage In-situ GIXD detection: Fix the sample on a variable temperature / stretching sample stage, calibrate the X-ray source (wavelength 1.54 Å), set the detection parameters, collect two-dimensional GIXD images at different temperatures and strains, analyze the peak position changes, fit the curves of the changes of interlayer and π-π stacking distance with temperature and strain, and infer the glass transition temperatures of the main chain (Tg,backbone) and the side chain (Tg,sidechain).
[0044] Atomic force microscope observation: The surface of the polymer film is scanned in tapping mode to observe the crack generation, extension and crystalline fiber bundle distribution of the film under tensile strain, record the crack initiation strain, and analyze the correlation between micro morphology and charge transport path.
[0045] 3. Electrical performance detection stage Tensile state electrical test: Gradient strain is applied to OFETs device, and the change of device on-state current and carrier mobility under different strain is tested, the electrical performance retention rate is recorded, and the influence of strain-induced molecular chain orientation on conductive performance is analyzed.
[0046] Variable temperature performance test: Repeat the electrical test in different temperature ranges to explore the regulation mechanism of molecular chain dynamics on the tensile properties and carrier transport efficiency of the polymer.
[0047] 4. Data integration and analysis stage Combine GIXD, atomic force microscope and electrical test data to build a correlation model of molecular structure-micro morphology-macro performance, and determine the influence of stress concentration area and molecular chain orientation on performance to provide basis for polymer material optimization.
[0048] III. General quality control and precautions Experimental environment control: Avoid strong light interference by using the fluorescence detection of Example 1, and maintain stable room temperature; maintain the vacuum environment of the sample table for GIXD detection to prevent air scattering from affecting the diffraction signal.
[0049] Parallel experiment requirement: Set 3 sets of parallel samples for each batch of samples, and the detection result deviation should be controlled within 5% to ensure the data reliability.
[0050] Reagent and sample storage: The conjugated polymer probe needs to be stored in a sealed refrigerator to avoid oxidation failure; the detection sample needs to be freshly treated to avoid degradation or volatilization of the target pollutant.
[0051] In summary, the sample treatment link of this scheme adopts the standardized impurity removal process of 8000-10000r / min centrifugation for 10 minutes, combined with buffer dilution and constant volume and 3 sets of parallel sample setting, which effectively reduces impurity interference and experimental error; In the data analysis stage, the detection limit is determined by fluorescence intensity normalization and linear fitting to form a whole process standardization system from sample preparation to result determination, which significantly improves the reliability and repeatability of the detection data, and makes up for the defects of the existing technology sample processing not standardized.
[0052] The above working process is described in Figures 1 to 7 .
[0053] It is to be understood that the terminology "including", "comprising", or other derivatives from the term "contain" are inclusive and that, in addition to the stated combinations, other combinations are also contemplated. It is to be further understood that the term "comprising" or "comprises" does not exclude other elements being present in addition to those listed. It is to be further understood that the term "including" or "includes" does not exclude other elements being present in addition to those listed.
[0054] While the embodiments of the application have been shown and described, it is to be understood that the application is not limited to these embodiments. Since modifications, variations, replacements and changes can be made to these embodiments without departing from the principles and spirit of the application, the scope of the application is defined by the appended claims and their equivalents.
Claims
1. A deep learning-based method for detecting bioactivity, characterized in that, Includes the following steps: S1, Preparatory work for testing, including the preparation of relevant reagents and materials and the debugging of testing instruments; S2, collect and process the sample to be tested to obtain a sample solution suitable for testing; S3, which causes the treated sample solution to react with the conjugated polymer-related reagents and collects relevant signals during the reaction process; S4 performs data analysis on the collected signals and determines the detection result based on the analysis results.
2. The method for detecting bioactivity based on deep learning according to claim 1, characterized in that, The detection method is fluorescence sensing, which is used to detect pollutants. The preparation of reagents and materials in S1 includes the synthesis of target conjugated polymer probes, the preparation of intermediates through Suzuki coupling reaction, and the introduction of active units through functionalization modification to inhibit aggregation-induced quenching (ACQ) effect. Prepare a series of standard test solutions, buffer solutions and blank control solutions with different concentration gradients.
3. The method for detecting bioactivity based on deep learning according to claim 2, characterized in that, The active unit includes triphenylamine (TPA) and methylimidazole; the standard detection solution is a PFOA / PFOS standard solution with a concentration range of 0-36 μM.
4. The method for detecting bioactivity based on deep learning according to claim 1, characterized in that, Sample processing in S2 includes removing impurities and precipitates from the sample to be tested by centrifugation at a speed of 8000-10000 r / min for 10 min, and taking the supernatant as the test stock solution. Based on the preliminary experimental results, the test stock solution is diluted with buffer solution to the appropriate concentration range and brought to the standard sample cell mark. Three parallel samples are set up at the same time.
5. The method for detecting bioactivity based on deep learning according to claim 1, characterized in that, In step S3, the reaction and signal acquisition process involves adding a quantitative conjugated polymer probe to the treated sample solution, reacting it in a constant-temperature shaker for 10-15 minutes at a temperature of 25°C and a rotation speed of 150 r / min; placing the reacted sample cell into a fluorescence spectrophotometer, using 370 nm as the excitation wavelength, scanning the fluorescence emission spectrum in the range of 400-700 nm, recording the characteristic peak intensity and peak shape changes, and simultaneously acquiring the sample fluorescence color under a UV lamp using a portable device, converting it into RGB digital information via an app; the same operation is performed on the blank control solution, and the fluorescence signal is acquired as a reference value.
6. The method for detecting bioactivity based on deep learning according to claim 1, characterized in that, Data analysis and result determination in S4 include calculating the normalized value of sample fluorescence intensity, plotting the concentration-fluorescence intensity standard curve, determining the limit of detection (LOD) through linear fitting, with a linear fitting correlation coefficient R² ≥ 0.99; determining the concentration of the target pollutant based on the sample fluorescence signal type and the standard curve, distinguishing pollutants of different chain lengths through linear discriminant analysis (LDA) algorithm, and outputting qualitative and semi-quantitative detection results.
7. A bioactivity detection device based on deep learning, applicable to any one of the bioactivity detection methods based on deep learning described in 1-6, comprising a detection stage (1), wherein a sealing cover (11) is rotatably mounted on one side of the upper surface of the detection stage (1), and a power module (12), an X-ray emission module (13), a heat dissipation module (14), and a control module (15) are respectively installed in the detection stage (1), wherein the power module (12) is used for power supply to the X-ray emission module (13), and the heat dissipation module (14) is used for heat dissipation inside the detection stage (1), characterized in that, It also includes a multi-functional linkage mechanism (2) and a self-adjusting protection mechanism (3); The multi-functional linkage mechanism (2) is set on the upper surface of the detection table (1), and the multi-functional linkage mechanism (2) is used for positioning and protecting the detection object; The self-adjusting protection mechanism (3) is installed in the multi-functional linkage mechanism (2), and the self-adjusting protection mechanism (3) is used for the linkage control of the multi-functional linkage mechanism (2).
8. The bioactivity detection device based on deep learning according to claim 7, characterized in that: The multi-functional linkage mechanism (2) includes a sealing chamber (21), which is installed on the upper surface of the detection platform (1). After the sealing chamber (21) and the sealing cover (11) are closed, a sealed detection chamber can be formed. An annular platform (22) is fixedly installed in the sealing chamber (21). A sliding tooth ring (23) is slidably installed on the inner surface of the annular platform (22). A rotating tooth (24) is circumferentially installed on the side of the sealing chamber (21) near the sliding tooth ring (23). The tooth surface of the rotating tooth (24) meshes with the sliding tooth ring (23). A connecting tooth (25) is fixedly installed in the middle of the upper surface of the rotating tooth (24). A sliding tooth strip (26) meshes with the tooth surface of the connecting tooth (25). The sliding tooth strip (26) is slidably installed in the annular platform (22). The sliding tooth strip (26) is located above the sliding tooth ring (23).
9. A bioactivity detection device based on deep learning according to claim 8, characterized in that: The self-adjusting protective mechanism (3) includes an arc-shaped guard plate (31), a guide rail toothed belt (32) is installed in the middle of the arc-shaped guard plate (31), and the inner surface of the guide rail toothed belt (32) is evenly provided with meshing grooves. A guide slide (34) is symmetrically slidably installed in the sliding rack (26). One end of the guide slide (34) is fixedly installed on the arc-shaped guard plate (31), and an auxiliary spring (37) is sleeved on the outer surface of the other end of the guide slide (34). One end of the auxiliary spring (37) is fixedly installed on the sliding rack (26). The other end of the auxiliary spring (37) is fixedly installed on the guide slide (34). The slide rack (26) has a guide groove (35) in the middle of one end near the arc-shaped guard plate (31). The guide groove (35) is provided with a drive tooth (33). The tooth surface of the drive tooth (33) meshes with the guide rail tooth belt (32). The two ends of the drive tooth (33) are fixedly installed with connecting shafts (36). The outer surfaces of the two ends of the connecting shaft (36) are movably installed in the guide groove (35). The two ends of the connecting shaft (36) are rotatably installed on the guide slide (34).