Organ-like toxicity effect evaluation method, system, equipment and medium
Through the organoid toxicity effect evaluation method, microscopic images and machine learning algorithms combined with molecular descriptors, the problems of low flux and insufficient resolution of compound toxicity evaluation were solved, and efficient and accurate toxicity effect evaluation was achieved.
Patent Information
- Application Number
- CN202510610864.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-13
- Publication Date
- 2025-08-15
AI Technical Summary
The existing compound toxicity evaluation methods have low flux and insufficient resolution, which leads to inaccurate evaluation results. Traditional two-dimensional in vitro cell models are difficult to simulate the three-dimensional complex microenvironment of human organs and cannot fully reflect the toxic effect of compounds on cells.
The organoid toxicity effect evaluation method is used to obtain microscopic images of experimental organoids and control organoids, and the first preset scoring criteria and machine learning algorithms are used to obtain the first toxicity score and the second toxicity score, and a structure-effect relationship is constructed in combination with molecular descriptors to achieve efficient and accurate toxicity effect evaluation.
High-throughput and accurate evaluation of compound toxicity is achieved, and multi-dimensional toxicity parameters can be obtained simultaneously at the single-cell level, improving the accuracy and stability of experimental test results.
Smart Images
Figure CN120485328A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of biomedicine and toxicology, and in particular to a method, system, equipment and medium for evaluating the toxic effects of organoids. Background Art
[0002] In the process of compound toxicity identification and screening, accurate assessment of compound toxicity requires establishing a reliable correlation between human health effects and experimental exposure results, and achieving high-throughput rapid screening. However, current toxicity assessment methods have the following key flaws:
[0003] Traditional two-dimensional in vitro cell models struggle to simulate the complex three-dimensional microenvironment of human organs and tissues, effectively reconstructing intercellular interactions and physiological metabolic functions. Consequently, they fail to fully reflect the toxic effects of compounds on cells. Animal exposure studies are limited by long experimental cycles and low throughput, making them inadequate for large-scale compound screening. Human effect studies, on the other hand, struggle to fully elucidate the specific toxic mechanisms of compound exposure.
[0004] Organoids have significant advantages as biological models for toxic effect evaluation. Organoids are three-dimensional cell aggregates formed in vitro by self-organization of stem cells or primary cells, which can partially reproduce the characteristics of organs in vivo in terms of morphology, structure and function. By inducing the differentiation of embryonic stem cells during the culture process, organoids of different species, including humans, can be obtained to realize exposure experiments on human-related toxic effects. Organoids contain a variety of cells, which can simulate the cell-cell interactions in real organs and tissues. In addition, the organoid culture cycle is 20-30 days, and large-scale culture can be achieved to realize large-scale compound exposure experiments, which is conducive to the realization of high-throughput toxic effect evaluation. However, in the existing organoid toxicity evaluation methods, traditional detection methods (such as fluorescence imaging) are limited by the number of labeling channels and analysis depth, and it is impossible to synchronously obtain multidimensional toxicity parameters at the single-cell level, resulting in poor accuracy and insufficient stability of experimental test results.
[0005] In view of this, there is an urgent need to develop a new method for evaluating the toxic effects of organoids. Summary of the Invention
[0006] The present invention provides a method, system, device and medium for evaluating the toxic effects of organoids, which are used to solve the defects of existing methods for evaluating the toxic effects of compounds, such as low throughput and insufficient resolution, resulting in inaccurate evaluation results.
[0007] The present invention provides a method for evaluating organoid toxic effects, comprising:
[0008] Obtaining microscopic images of the experimental organoid at multiple time points during exposure to the target compound, and obtaining microscopic images of the control organoid at the same multiple time points when the control organoid is not exposed to any compound, wherein the control organoid is the same type and prepared as the experimental organoid;
[0009] Obtaining a first toxicity score of the experimental organoid under the target compound using a first preset scoring standard based on a plurality of microscopic images of the experimental organoid at a plurality of time points and a plurality of microscopic images of the control organoid;
[0010] Based on a plurality of microscopic images of the experimental organoid, a machine learning algorithm is used to obtain the pixel area of the experimental organoid on each microscopic image, and a second preset scoring standard is used to obtain a second toxicity score of the experimental organoid under the target compound;
[0011] Obtain molecular descriptors of experimental organoids;
[0012] Based on the molecular descriptors of the experimental organoids and the first toxicity scores and second toxicity scores of the experimental organoids under the target compound, the toxic effect evaluation of the experimental organoids under the target compound is achieved.
[0013] According to a method for evaluating the toxic effects of organoids provided by the present invention, the experimental organoids can be pluripotent stem cell (PSC)-derived organoids or adult stem cell (ASC)-derived organoids, wherein the pluripotent stem cell-derived organoids include any one of the following or any combination thereof: brain organoids, intestinal organoids, liver organoids, and adult stem cell-derived organoids include any one of the following or any combination thereof: small intestine / colon organoids, gastric organoids, lung organoids, pancreatic organoids, kidney organoids, and breast organoids.
[0014] According to a method for evaluating the toxic effects of organoids provided by the present invention, the target compounds can be conventional pollutants or new pollutants, wherein conventional pollutants include any one of the following or any combination thereof: sulfur dioxide, nitrogen oxides, PM2.5, etc., and new pollutants include any one of the following or any combination thereof: persistent organic pollutants controlled by international conventions, endocrine disruptors, antibiotics, etc., such as perfluorinated and polyfluorinated compounds and their substitutes, endocrine disruptors represented by tetrabromobisphenol A and its derivatives, and other compounds.
[0015] According to a method for evaluating toxic effects of an organoid provided by the present invention, the method comprises obtaining a first toxicity score of the experimental organoid under the condition of a target compound using a first preset scoring standard based on a plurality of microscopic images of the experimental organoid and a plurality of microscopic images of the control organoid, comprising:
[0016] For each microscopic image of the experimental organoid, a corresponding first toxicity score is obtained using a first preset scoring standard. Then, based on the first toxicity scores corresponding to all microscopic images of the experimental organoid, a first toxicity score of the experimental organoid under the target compound is obtained, wherein the first preset scoring standard includes:
[0017] The damage of the experimental organoids was determined based on their microscopic images. When the experimental organoids were determined to have the preset severe damage characteristics, the first toxicity score of the experimental organoids was limited to the range of 60-70 points. When the experimental organoids were determined not to have the preset severe damage characteristics, the number of cell drops of the experimental organoids was evaluated based on the microscopic images of the experimental and control organoids at the same time point, and the first toxicity score of the experimental organoids was limited to the range of 70-75 points, 75-80 points, 80-85 points, 85-90 points or 90-95 points. After the scoring range of the first toxicity score of the experimental organoids was determined, the morphological characteristics of the experimental organoids were evaluated based on the microscopic images of the experimental and control organoids at the same time point, and the first toxicity score of the experimental organoids was adjusted up or down by no more than 3 points.
[0018] According to a method for evaluating toxic effects of an organoid provided by the present invention, the method comprises obtaining the pixel area of the experimental organoid in each microscopic image using a machine learning algorithm based on a plurality of microscopic images of the experimental organoid, and obtaining a second toxicity score of the experimental organoid under the target compound using a second preset scoring standard, including:
[0019] According to each microscopic image of the experimental organoid, the pixel area of the experimental organoid on each microscopic image is obtained by using the watershed algorithm;
[0020] The pixel areas of the experimental organoids in several microscopic images were calculated by multiple ratios to obtain the second toxicity score of the experimental organoids under the target compound.
[0021] According to a method for evaluating the toxic effects of organoids provided by the present invention, when the experimental organoid is a brain organoid and the target compound is a perfluorinated compound, the molecular descriptors of the experimental organoid include any one of the following or any combination thereof: Crippen partition coefficient (CrippenClogP), Labute approximate solvent accessible surface area (LabuteASA), Hall-Kier molecular connectivity index α (HallKierAlpha), topological polar surface area (Topological Polar Surface Area, tpsa), and Bertz Complexity Index (BertzCT).
[0022] According to the present invention, a method for evaluating the toxic effects of an organoid is provided, wherein the method evaluates the toxic effects of the experimental organoid under the target compound based on the molecular descriptor of the experimental organoid and the first toxicity score and the second toxicity score of the experimental organoid under the target compound, comprising:
[0023] Based on the molecular descriptors of the experimental organoids and the first and second toxicity scores of the experimental organoids under the target compounds, a quantitative structure-activity relationship (QSAR) model was constructed to predict potential compounds whose toxic effects met the preset conditions.
[0024] According to the present invention, a method for evaluating organoid toxic effects further comprises:
[0025] Acquire single-cell mass spectrometry imaging data after antibody staining on pre-established antibody channels in experimental organoids;
[0026] Based on the single-cell mass spectrometry imaging data of experimental organoids, deep learning models are combined to perform cell segmentation and cell subtype classification on experimental organoids;
[0027] Based on the cell subtype classification results of the experimental organoids, the toxic effect evaluation of the experimental organoids under the target compound is verified.
[0028] The present invention also provides an organoid toxicity effect evaluation system, comprising:
[0029] a microscopic image acquisition module, configured to acquire microscopic images of the experimental organoid at a plurality of time points during exposure to the target compound, and, when the control organoid is not exposed to any compound, acquire microscopic images of the control organoid at the same plurality of time points, wherein the control organoid and the experimental organoid are of the same type and prepared by the same method;
[0030] A first toxicity scoring module is configured to obtain a first toxicity score of the experimental organoid under the target compound based on a plurality of microscopic images of the experimental organoid at a plurality of time points and a plurality of microscopic images of the control organoid using a first preset scoring standard;
[0031] a second toxicity scoring module, configured to: obtain, based on a plurality of microscopic images of the experimental organoids, a pixel area of each microscopic image of the experimental organoids using a machine learning algorithm, and obtain a second toxicity score of the experimental organoids under the target compound using a second preset scoring standard;
[0032] Structural parameter acquisition module, used to: obtain molecular descriptors of experimental organoids;
[0033] The toxic effect evaluation module is used to: evaluate the toxic effect of the experimental organoid under the target compound based on the molecular descriptor of the experimental organoid and the first toxicity score and the second toxicity score of the experimental organoid under the target compound.
[0034] The present invention also provides an electronic device comprising a processor and a memory storing a computer program, wherein when the processor executes the computer program, any of the above-mentioned organoid toxicity effect evaluation methods is implemented.
[0035] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned methods for evaluating organoid toxic effects.
[0036] The present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute any of the above-mentioned organoid toxicity effect evaluation methods.
[0037] The present invention provides a method, system, device and medium for evaluating the toxic effects of organoids. These methods utilize microscopic images of experimental organoids exposed to a target compound and microscopic images of control organoids not exposed to any compound, and obtain a first toxicity score and a second toxicity score of the experimental organoids under the target compound through a first preset scoring standard, a second preset scoring standard and a machine learning algorithm. The structure-effect relationship of the experimental organoids in toxic effect evaluation is then constructed in combination with molecular descriptors to achieve accurate and efficient toxic effect evaluation. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0039] Figure 1 A schematic flow chart of a method for evaluating organoid toxic effects provided by the present invention.
[0040] Figure 2 The morphological structure of some brain organoids is shown as a reference for the first toxicity scoring of microscopic images of brain organoids.
[0041] Figure 3 Shown is the process of performing the first toxicity scoring on microscopic images of brain organoids.
[0042] Figure 4 Shown is the process of performing a secondary toxicity score on microscopic images of brain organoids.
[0043] Figure 5 The results of structural parameter prediction of the target compound using the Rdkit tool are shown.
[0044] Figure 6 The compounds with obvious toxic effects and compounds with potential research value are shown.
[0045] Figure 7 Results of an antibody validation test are shown.
[0046] Figure 8 Results of a range-expansion experiment are shown.
[0047] Figure 9 Shows the segmentation mask generated locally by Mesmer.
[0048] Figure 10 This is a schematic diagram of the structure of an organoid toxic effect evaluation system provided by the present invention.
[0049] Figure 11 This is a schematic structural diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION
[0050] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the embodiments described are part of the embodiments of the present invention, not all of the embodiments, and they should not be understood as limitations on the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In the description of the present invention, it should be understood that the terms used are only for descriptive purposes and cannot be understood as indicating or implying relative importance.
[0051] Figure 1 This is a flow chart of a method for evaluating organoid toxicity provided by the present invention. The method can be executed by any applicable terminal-side device or network-side device, such as an organoid toxicity evaluation device.
[0052] See also Figure 1 The present invention provides a method for evaluating organoid toxicity effects, which may include:
[0053] S110. Obtain microscopic images of the experimental organoids at several time points during the exposure of the experimental organoids to the target compound, and, when the control organoids are not exposed to any compound, obtain microscopic images of the control organoids at the same several time points, wherein the control organoids and the experimental organoids are of the same type and prepared by the same method.
[0054] In one embodiment, the experimental organoids can be pluripotent stem cell (PSC)-derived organoids or adult stem cell (ASC)-derived organoids, wherein the pluripotent stem cell-derived organoids include any one of the following or any combination thereof: brain organoids, intestinal organoids, liver organoids, and adult stem cell-derived organoids include any one of the following or any combination thereof: small intestine / colon organoids, gastric organoids, lung organoids, pancreatic organoids, kidney organoids, and breast organoids.
[0055] In one embodiment, the target compound can be a conventional pollutant or a new pollutant, wherein conventional pollutants include any one of the following or any combination thereof: sulfur dioxide, nitrogen oxides, PM2.5, etc., and new pollutants include any one of the following or any combination thereof: persistent organic pollutants controlled by international conventions, endocrine disruptors, antibiotics, etc., such as perfluorinated and polyfluorinated compounds and their substitutes, endocrine disruptors represented by tetrabromobisphenol A and its derivatives, and other compounds.
[0056] In this embodiment, human embryonic stem cells were directed to induce differentiation to form brain organoid structures. After 30 days of culture and maturation, experimental organoids and control organoids were obtained. The experimental organoids were then exposed to a perfluorochemical (target compound) at a concentration of 1 μM for seven days, and 4× microscopic observations were performed on the 2nd, 4th, and 7th days, and images were retained. After the 7-day exposure experiment, the samples were immediately fixed with paraformaldehyde and collected. At the same time, the control organoids that were not exposed to any compound were observed and recorded on the 2nd, 4th, and 7th days. Several microscopic images of the experimental organoids and several microscopic images of the control organoids were obtained. When obtaining the microscopic images, the magnification of each picture was consistent, and the ambient light was similar to the position of the organoid. The image format and resolution of all images were unified. Specifically, after a comprehensive comparison of the formats and resolutions of all microscopic images, the minimum resolution (2736*1824) was used as the unified resolution, and all image formats were transposed to the PNG image file format.
[0057] S120. Obtain a first toxicity score of the experimental organoid under the target compound based on a first preset scoring standard according to a plurality of microscopic images of the experimental organoid at a plurality of time points and a plurality of microscopic images of the control organoid.
[0058] In one embodiment, S120 may include: for each microscopic image of the experimental organoid, using a first preset scoring standard to obtain a corresponding first toxicity score, and then summing and averaging the first toxicity scores corresponding to all microscopic images of the experimental organoid to obtain the first toxicity score of the experimental organoid under the target compound.
[0059] In one embodiment, see Figure 2 and Figure 3 , the first preset scoring criteria include:
[0060] The damage of the experimental organoids was determined based on their microscopic images. When the experimental organoids were determined to have preset severe damage features (which may be obvious hollow structures, and the severe damage features can be manually preset), the first toxicity score of the experimental organoids was limited to 60-70 points. When the experimental organoids were determined not to have preset severe damage features, the number of cell drops in the experimental organoids was evaluated based on the microscopic images of the experimental and control organoids at the same time point, and the first toxicity score of the experimental organoids was limited to 70-75 points, 75-80 points, 80-85 points, 85-90 points, or 90-95 points (for example, the background values of the microscopic images of the experimental and control organoids at the same time point were compared. If the difference in the number of shadows from the background value was about 5%, it meant that there was basically no change, and the score range was 90-95 points, which was specifically manifested in a bright culture environment and no small shadows). =The score range is 85-90 points; if the difference between the number of shadows and the background value is 5-15%, the number of dropped cells is small, and the score range is 85-90 points; if the number of shadows exceeds the background value by 15%-30%, the number of dropped cells is small, and the score range is 80-85 points; if it exceeds the background value by 30%-50%, the number of dropped cells is large, and the score range is 75-80 points; if it exceeds the background value by more than 50%, the number of dropped cells is large, and the score range is 70-75 points). After the scoring range of the first toxicity score of the experimental organoid is determined, the morphological characteristics of the experimental organoid are evaluated based on the microscopic images of the experimental organoid and the control organoid at the same time point, and the first toxicity score of the experimental organoid is adjusted up or down by no more than 3 points (for example, the morphological characteristics of the experimental organoid at the same time point are compared with the control organoid, and the score is adjusted according to whether there are edge and internal damage characteristics, or the score is adjusted according to the degree of deviation of the morphological characteristics).
[0061] S130. Based on the plurality of microscopic images of the experimental organoid, a machine learning algorithm is used to obtain the pixel area of the experimental organoid in each microscopic image, and a second preset scoring standard is used to obtain a second toxicity score of the experimental organoid under the target compound.
[0062] In one embodiment, see Figure 4, S130 may include:
[0063] Based on each microscopic image of the experimental organoid, a watershed algorithm is used to obtain the pixel area of the experimental organoid in each microscopic image. Specifically, the original image can be first transposed into a second-order grayscale image, and then a threshold is set to distinguish the organoid from the surrounding environment. After obtaining the organoid pixel distribution, the area of the irregular image is calculated to achieve the area measurement of the organoid. Alternatively, a manually supervised watershed algorithm can be used to delineate the range of the organoid to obtain the pixel area of the organoid.
[0064] A multiple ratio calculation is performed on the pixel areas of the experimental organoids in several microscopic images to obtain a second toxicity score for the experimental organoids exposed to the target compound. Specifically, after obtaining the organoid pixel area, a ratio calculation is performed on the pixel area data of the experimental organoids at different stages of exposure to the same target compound. For example, the pixel area of the microscopic image of the same experimental organoid on day 7 is ratioed to the pixel area of the microscopic image of the same experimental organoid on day 4. The resulting value is recorded as the 7 / 4 ratio, reflecting the growth of the experimental organoid from day 4 to day 7. In this process, multiple ratios are calculated over the entire process to serve as the second toxicity score for the experimental organoids exposed to the target compound. In this example, the 7 / 2 ratio is used as an overall reference to illustrate the degree of perturbation of the experimental organoid growth by the target compound exposure. This value is standardized to 1; a value greater than 1.2 or less than 0.8 indicates that the target compound exposure has affected the growth of the experimental organoid. The process values of the 4 / 2 ratio and the 7 / 4 ratio are used as growth curves to assist in the evaluation of toxic effects.
[0065] S140. Obtain molecular descriptors (also called structural parameters) of experimental organoids.
[0066] In this example, the experimental organoids are brain organoids, and the target compounds are perfluorinated compounds.
[0067] In one embodiment, the properties and structure of the target compound can be quantitatively characterized using a molecular descriptor method. Since the target compound is a perfluorinated compound, except for perfluorinated compounds containing special element atoms, the compounds within a smaller radius centered on the atom are not much different. Therefore, when using molecular descriptors, a system using a single atomic radius as the calculation standard should be avoided.
[0068] In this example, the structural parameters of the compounds to be calculated and predicted are selected based on the environmental behavior of the target compounds and the contact characteristics of organoids with human organs. Considering that the exposure pathways of perfluorinated and polyfluorinated compounds to the human brain in the actual environment must first enter the human body through multiple pathways such as contact and ingestion, and then pass through the blood-brain barrier to enter the human brain, the following structural parameters are selected to calculate and predict the structural parameters of the exposed perfluorinated and polyfluorinated compounds (see Figure 5): Crippen ClogP, used to predict the partition coefficient (logP) of a molecule, that is, the tendency of the molecule to distribute between oil and water; Labute approximate solvent accessible surface area (Labute ASA), used to describe properties such as solubility and biodistribution of a molecule; Hall-Kier molecular connectivity index α (HallKierAlpha), is the calculation result of the Hall-Kier alpha value, which can be used to represent the polarity and charge distribution of a molecule; Topological Polar Surface Area (tpsa), a parameter commonly used in medicinal chemistry, is defined as the total surface area of polar molecules in a compound; Bertz Complexity Index (BertzCT), represents the topological structure of a lattice in space. This molecular descriptor is intended to describe the complexity of a molecule, which to a certain extent can represent the steric hindrance caused by the cyclic or branched structure of the molecule, indicating that it may be associated with the difficulty of the compound binding to proteins and cell permeability during interaction with cells.
[0069] S150. Implement toxic effect evaluation of the experimental organoid under the target compound based on the molecular descriptor of the experimental organoid and the first toxicity score and the second toxicity score of the experimental organoid under the target compound.
[0070] In one embodiment, S150 may include:
[0071] Based on the molecular descriptors of the experimental organoids and the first and second toxicity scores of the experimental organoids under the target compound, a quantitative structure-activity relationship (QSAR) model is constructed to predict potential compounds whose toxic effects meet the preset conditions (for example, the first and second toxicity scores reach a certain preset range, respectively).
[0072] In one embodiment, the molecular descriptors (structural parameters) of the target compound can be associated with the first toxicity score and the second toxicity score of the experimental organoid, so as to find compounds with obvious toxic effects and compounds with potential research value. Figure 6In this embodiment, after sorting the first toxicity score in descending order, it can be found that the perfluorinated compound substitute containing an ether bond has a more obvious toxic tendency. For example, after exposure to sevoflurane, the PAX6 signal is significantly enhanced, indicating that its radial glial cells increase, while the ki67 signal decreases, indicating that the perfluorinated compound inhibits the growth of brain organoids to a certain extent. This can also be confirmed by the pixel ratio in the second toxicity score. After exposure to perfluoro-(2,5,8-trimethyl-3,6,9-trioxadodecanoic acid), its tuj1 and PAX6 signals are significantly enhanced, and ki67 expression is expressed in the boundary part. Combined with the phenomenon of a large number of fallen cells in the early microscopic observation, it is believed that perfluoro-(2,5,8-trimethyl-3,6,9-trioxadodecanoic acid) makes the brain organoid structure loose, but does not have an inhibitory effect on the growth of brain organoids. This can also be confirmed by the pixel ratio in the second toxicity score. Through the above analysis, potential compounds whose toxic effects meet the preset conditions can be obtained, which is helpful for compound research.
[0073] S160. Acquire single-cell mass spectrometry imaging data after antibody staining on pre-constructed antibody channels of experimental organoids.
[0074] In one embodiment, single-cell mass spectrometry imaging data obtained after antibody staining on a pre-constructed antibody channel of an experimental organoid can be obtained by the following process:
[0075] I. Construction of Brain Organoid Antibody Channel
[0076] I-1. Antibody effect test
[0077] In order to avoid antibody waste and achieve efficient use of antibodies, trace antibody effectiveness tests are performed on the antibodies contained in the pre-designed antibody channel.
[0078] For details, see Figure 7 In order to verify the effectiveness of the antibodies in the group, six antibodies that have been verified in mouse brain slices were selected for the first stage of effectiveness and titer range verification to determine the effectiveness of the selected antibodies in human brain organoid immunostaining and to determine the subsequent pre-experimental antibody titer test range to be between 1:50 and 1:100. The six antibodies and their corresponding metal channels are E-cadeherin (158Gd), N-cadeherin (166Er), SOX2 (173Yb), Ki67 (150Nd), DCX (144Nd), and Tuj1 (148Nd).
[0079] I-2. Scope Expansion Experiment
[0080] After completing the antibody effect test, the test scope will be expanded to pre-designed antibodies of the same type, and evaluation criteria for the effectiveness of antibody titers will be established to further verify the antibodies.
[0081] For details, see Figure 8 After determining the effectiveness of the six basic antibodies on brain organoids, a basic cell screening method was established for brain organoids. On this basis, the existing antibodies against nerve cells in the group were expanded for verification. The range expansion experiment will include all existing nerve cell-related antibodies in the group for verification, and the initial experiment of the range expansion experiment will be carried out according to the titer range obtained from the antibody effect test, and the screening and adjustment will be carried out step by step according to the effect. In the process of establishing the evaluation criteria for the effectiveness of antibody titer, the MCDviewer is used for observation, and the ThresholdMax of each antibody is adjusted to about 10 using a single channel to observe the antibody distribution. The background value and the distribution of isotope metal signals with a close mass number are observed, so that the distribution shows a trend of spatial distribution aggregation and there is no channel crosstalk between different metals or the influence can be eliminated due to the difference in spatial distribution. In this way, this antibody can play a role in channel construction and this antibody is included in the antibody channel.
[0082] I-3. Determination of Antibody Titer
[0083] A full-channel scanning method was used to determine the antibody channels and their corresponding titers used in the formal experiment from the expanded range of antibody channels. The overall experimental design was based on the principle of avoiding the waste of individually prepared antibodies and biological samples.
[0084] Specifically, the antibodies included in the expanded scope experiment are E-cadeherin, N-cadeherin, SOX2, Ki67, DCX, Tuj1, PAX6, GFAP, NeuN, Olig2, S100β, Choline, and GAD67. After completing the antibody titer effectiveness evaluation criteria, based on this criterion, the following antibodies were selected for brain organoid single-cell mass spectrometry imaging test analysis to construct expression profiles: E-cadeherin, N-cadeherin, SOX2, Ki67, DCX, Tuj1, PAX6, and NeuN.
[0085] II. Antibody Staining of Paraffin Sections of Brain Organoids for Single-Cell Mass Spectrometry Imaging
[0086] The antibody staining method for paraffin sections of brain organoids for single-cell mass spectrometry imaging is as follows:
[0087] Optical positioning of brain organoid tissue on the slice is performed to determine the position of the organoid tissue on the slice.
[0088] Place the paraffin-embedded sections in a 60°C oven for 2 hours. After 2 hours, remove them and soak them in a well-ventilated xylene vat for 10 minutes. Transfer them to another xylene vat and soak them for another 10 minutes. After rehydration, rehydrate the sections in 100%, 95%, 80%, and 70% anhydrous ethanol solutions for 5 minutes each. Remove them from the 70% ethanol and rinse them in ddH2O for 5 minutes.
[0089] After completing the above steps, incubate the slides in preheated antigen retrieval solution at 96°C for 30 minutes. After incubation, allow to cool naturally at room temperature for 10 minutes to 70°C. After cooling, rinse the slides in ddH2O and then PBS for 10 minutes.
[0090] Wipe the slide dry with lens tissue (being careful not to touch the tissue area). Use a PAP pen to draw a circle around the tissue area and block for 45 minutes at room temperature in DPBS containing 3% BSA. While blocking, prepare the antibody cocktail in 0.5% BSA. After the 45-minute blocking period, aspirate the blocking solution and apply the antibody dropwise to the tissue area. Incubate overnight at 4°C.
[0091] After incubation, remove the sections from the 4°C refrigerator and allow to rest at room temperature for 5 minutes. Rinse the sections with 0.2% Triton-X in DPBS for 8 minutes on a shaker. Transfer the sections to another 0.2% Triton-X solution in DPBS and repeat this process. Rinse with DPBS for 8 minutes twice.
[0092] After rinsing, apply Ir at a 1:1600 ratio to the tissue area at room temperature and place in a humidified chamber for staining for 30 minutes. After staining is complete, rinse in ddH2O for 5 minutes, remove the slide and air dry at room temperature for at least 20 minutes.
[0093] III. Single-cell mass spectrometry imaging, data acquisition, and analysis
[0094] After antibody staining, the sections were placed in a tissue imaging mass spectrometry flow system (Hyperion), and the instrument was scanned and analyzed using CyTOF software to obtain MCD and TXT data. The TXT data were automatically segmented and analyzed using the Mesmer method using a Shell script (e.g. Figure 9 As shown), single-cell mass spectrometry imaging data of experimental organoids were obtained.
[0095] S170. Based on the single-cell mass spectrometry imaging data of experimental organoids, deep learning models are combined to perform cell segmentation and cell subtype classification on experimental organoids.
[0096] In one embodiment, the deep learning model can be pre-trained based on single-cell mass spectrometry imaging data of similar experimental organoids and corresponding cell subtype classification label data.
[0097] S180. Verify the toxic effect evaluation of the experimental organoids under the target compound based on the cell subtype classification results of the experimental organoids.
[0098] After data analysis of organoid single-cell mass spectrometry images, namely morphological analysis, single-cell segmentation and typing, the morphological structure and cell composition types of the organoids were analyzed and evaluated. Specifically, morphological analysis was first performed on the organoids, and the internal structure of the brain organoids presented by the single-cell mass spectrometry images was evaluated. The destruction phenomena of holes and damaged areas inside the organoids and the overall morphological structure and developmental abnormalities were sorted out and evaluated. Combined with molecular targets such as proliferation factors represented by ki67 and cytoskeleton represented by tuj1, a comprehensive toxicity evaluation of the organoids was conducted. Based on the antibody channels constructed in the early stage, the cell types were classified, the various cell types were statistically analyzed, and the independent evaluation based on morphological analysis was verified to complete the overall toxic effect evaluation.
[0099] The present invention provides a method, system, device and medium for evaluating the toxic effects of organoids. These methods utilize microscopic images of experimental organoids exposed to a target compound and microscopic images of control organoids not exposed to any compound, and obtain a first toxicity score and a second toxicity score of the experimental organoids under the target compound through a first preset scoring standard, a second preset scoring standard and a machine learning algorithm. The structure-effect relationship of the experimental organoids in toxic effect evaluation is then constructed in combination with molecular descriptors to achieve accurate and efficient toxic effect evaluation.
[0100] The organoid toxicity effect evaluation system provided by the present invention is described below. The organoid toxicity effect evaluation system described below and the organoid toxicity effect evaluation method described above can be referenced to each other.
[0101] See also Figure 10 The present invention provides an organoid toxicity effect evaluation system, which may include:
[0102] a microscopic image acquisition module, configured to acquire microscopic images of the experimental organoid at a plurality of time points during exposure to the target compound, and, when the control organoid is not exposed to any compound, acquire microscopic images of the control organoid at the same plurality of time points, wherein the control organoid and the experimental organoid are of the same type and prepared by the same method;
[0103] A first toxicity scoring module is configured to obtain a first toxicity score of the experimental organoid under the target compound based on a plurality of microscopic images of the experimental organoid at a plurality of time points and a plurality of microscopic images of the control organoid using a first preset scoring standard;
[0104] a second toxicity scoring module, configured to: obtain, based on a plurality of microscopic images of the experimental organoids, a pixel area of each microscopic image of the experimental organoids using a machine learning algorithm, and obtain a second toxicity score of the experimental organoids under the target compound using a second preset scoring standard;
[0105] Structural parameter acquisition module, used to: obtain molecular descriptors of experimental organoids;
[0106] The toxic effect evaluation module is used to: evaluate the toxic effect of the experimental organoid under the target compound based on the molecular descriptor of the experimental organoid and the first toxicity score and the second toxicity score of the experimental organoid under the target compound.
[0107] In one embodiment, it may further include:
[0108] A single-cell mass spectrometry imaging data acquisition module is used to: acquire single-cell mass spectrometry imaging data obtained after antibody staining on pre-constructed antibody channels of experimental organoids;
[0109] The cell subtype classification module is used to perform cell segmentation and cell subtype classification on experimental organoids based on single-cell mass spectrometry imaging data combined with deep learning models;
[0110] The toxic effect evaluation and verification module is used to: verify the toxic effect evaluation of experimental organoids under target compounds based on the cell subtype classification results of the experimental organoids.
[0111] Figure 11 An example of a physical structure diagram of an electronic device is shown below. Figure 11 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to perform the following steps:
[0112] Obtaining microscopic images of the experimental organoid at multiple time points during exposure to the target compound, and obtaining microscopic images of the control organoid at the same multiple time points when the control organoid is not exposed to any compound, wherein the control organoid is the same type and prepared as the experimental organoid;
[0113] Obtaining a first toxicity score of the experimental organoid under the target compound using a first preset scoring standard based on a plurality of microscopic images of the experimental organoid at a plurality of time points and a plurality of microscopic images of the control organoid;
[0114] Based on a plurality of microscopic images of the experimental organoid, a machine learning algorithm is used to obtain the pixel area of the experimental organoid on each microscopic image, and a second preset scoring standard is used to obtain a second toxicity score of the experimental organoid under the target compound;
[0115] Obtain molecular descriptors of experimental organoids;
[0116] Based on the molecular descriptors of the experimental organoids and the first toxicity scores and second toxicity scores of the experimental organoids under the target compound, the toxic effect evaluation of the experimental organoids under the target compound is achieved.
[0117] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.
[0118] In another aspect, the present invention further provides a computer program product, comprising a computer program, which may be stored on a non-transitory computer-readable storage medium, and wherein when the computer program is executed by a processor, the computer is capable of performing the following steps:
[0119] Obtaining microscopic images of the experimental organoid at multiple time points during exposure to the target compound, and obtaining microscopic images of the control organoid at the same multiple time points when the control organoid is not exposed to any compound, wherein the control organoid is the same type and prepared as the experimental organoid;
[0120] Obtaining a first toxicity score of the experimental organoid under the target compound using a first preset scoring standard based on a plurality of microscopic images of the experimental organoid at a plurality of time points and a plurality of microscopic images of the control organoid;
[0121] Based on a plurality of microscopic images of the experimental organoid, a machine learning algorithm is used to obtain the pixel area of the experimental organoid on each microscopic image, and a second preset scoring standard is used to obtain a second toxicity score of the experimental organoid under the target compound;
[0122] Obtain molecular descriptors of experimental organoids;
[0123] Based on the molecular descriptors of the experimental organoids and the first toxicity scores and second toxicity scores of the experimental organoids under the target compound, the toxic effect evaluation of the experimental organoids under the target compound is achieved.
[0124] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is configured to execute the following steps when executed by a processor:
[0125] Obtaining microscopic images of the experimental organoid at multiple time points during exposure to the target compound, and obtaining microscopic images of the control organoid at the same multiple time points when the control organoid is not exposed to any compound, wherein the control organoid is the same type and prepared as the experimental organoid;
[0126] Obtaining a first toxicity score of the experimental organoid under the target compound using a first preset scoring standard based on a plurality of microscopic images of the experimental organoid at a plurality of time points and a plurality of microscopic images of the control organoid;
[0127] Based on a plurality of microscopic images of the experimental organoid, a machine learning algorithm is used to obtain the pixel area of the experimental organoid on each microscopic image, and a second preset scoring standard is used to obtain a second toxicity score of the experimental organoid under the target compound;
[0128] Obtain molecular descriptors of experimental organoids;
[0129] Based on the molecular descriptors of the experimental organoids and the first toxicity scores and second toxicity scores of the experimental organoids under the target compound, the toxic effect evaluation of the experimental organoids under the target compound is achieved.
[0130] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.
[0131] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.
[0132] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.
Claims
1. A method for evaluating organoid toxic effects, characterized in that: include: Obtaining microscopic images of the experimental organoid at multiple time points during exposure to the target compound, and obtaining microscopic images of the control organoid at the same multiple time points when the control organoid is not exposed to any compound, wherein the control organoid is the same type and prepared as the experimental organoid; Obtaining a first toxicity score of the experimental organoid under the target compound using a first preset scoring standard based on a plurality of microscopic images of the experimental organoid at a plurality of time points and a plurality of microscopic images of the control organoid; Based on a plurality of microscopic images of the experimental organoid, a machine learning algorithm is used to obtain the pixel area of the experimental organoid on each microscopic image, and a second preset scoring standard is used to obtain a second toxicity score of the experimental organoid under the target compound; Obtain molecular descriptors of experimental organoids; Based on the molecular descriptors of the experimental organoids and the first toxicity scores and second toxicity scores of the experimental organoids under the target compound, the toxic effect evaluation of the experimental organoids under the target compound is achieved.
2. The method for evaluating organoid toxic effects according to claim 1, wherein: Experimental organoids are organoids derived from pluripotent stem cells or adult stem cells, among which pluripotent stem cell-derived organoids include any one of the following or any combination thereof: brain organoids, intestinal organoids, and liver organoids; adult stem cell-derived organoids include any one of the following or any combination thereof: small intestine / colon organoids, gastric organoids, lung organoids, pancreatic organoids, kidney organoids, and breast organoids.
3. The method for evaluating organoid toxic effects according to claim 1, wherein: The step of obtaining a first toxicity score of the experimental organoid under the target compound using a first preset scoring standard based on the plurality of microscopic images of the experimental organoid and the plurality of microscopic images of the control organoid includes: For each microscopic image of the experimental organoid, a corresponding first toxicity score is obtained using a first preset scoring standard. Then, based on the first toxicity scores corresponding to all microscopic images of the experimental organoid, a first toxicity score of the experimental organoid under the target compound is obtained, wherein the first preset scoring standard includes: The damage of the experimental organoids was determined based on their microscopic images. When the experimental organoids were determined to have the preset severe damage characteristics, the first toxicity score of the experimental organoids was limited to the range of 60-70 points. When the experimental organoids were determined not to have the preset severe damage characteristics, the number of cell drops of the experimental organoids was evaluated based on the microscopic images of the experimental and control organoids at the same time point, and the first toxicity score of the experimental organoids was limited to the range of 70-75 points, 75-80 points, 80-85 points, 85-90 points or 90-95 points. After the scoring range of the first toxicity score of the experimental organoids was determined, the morphological characteristics of the experimental organoids were evaluated based on the microscopic images of the experimental and control organoids at the same time point, and the first toxicity score of the experimental organoids was adjusted up or down by no more than 3 points.
4. The method for evaluating organoid toxic effects according to claim 3, wherein: The method further comprises obtaining the pixel area of the experimental organoid on each microscopic image by a machine learning algorithm based on the plurality of microscopic images of the experimental organoid, and obtaining a second toxicity score of the experimental organoid under the target compound using a second preset scoring standard, including: According to each microscopic image of the experimental organoid, the pixel area of the experimental organoid on each microscopic image is obtained by using the watershed algorithm; The pixel areas of the experimental organoids in several microscopic images were calculated by multiple ratios to obtain the second toxicity score of the experimental organoids under the target compound.
5. The method for evaluating organoid toxic effects according to any one of claims 1 to 4, characterized in that: When the experimental organoid is a brain organoid and the target compound is a perfluorinated compound, the molecular descriptors of the experimental organoid include any one of the following or any combination thereof: Krabben partition coefficient, Labute approximate solvent accessible surface area, Hall-Kiln molecular connectivity index α, topological polar surface area, and Belz complexity index.
6. The method for evaluating organoid toxic effects according to any one of claims 1 to 4, characterized in that: The method of implementing the toxic effect evaluation of the experimental organoid under the target compound according to the molecular descriptor of the experimental organoid and the first toxicity score and the second toxicity score of the experimental organoid under the target compound includes: Based on the molecular descriptors of the experimental organoids and the first and second toxicity scores of the experimental organoids under the target compounds, a quantitative structure-activity relationship model is constructed, and potential compounds whose toxic effects meet the preset conditions are predicted.
7. The method for evaluating organoid toxic effects according to any one of claims 1 to 4, characterized in that: Also includes: Acquire single-cell mass spectrometry imaging data after antibody staining on pre-established antibody channels in experimental organoids; Based on the single-cell mass spectrometry imaging data of experimental organoids, deep learning models are combined to perform cell segmentation and cell subtype classification on experimental organoids; Based on the cell subtype classification results of the experimental organoids, the toxic effect evaluation of the experimental organoids under the target compound is verified.
8. An organoid toxicity effect evaluation system, characterized in that: include: a microscopic image acquisition module, configured to acquire microscopic images of the experimental organoid at a plurality of time points during exposure to the target compound, and, when the control organoid is not exposed to any compound, acquire microscopic images of the control organoid at the same plurality of time points, wherein the control organoid and the experimental organoid are of the same type and prepared by the same method; A first toxicity scoring module is configured to obtain a first toxicity score of the experimental organoid under the target compound based on a plurality of microscopic images of the experimental organoid at a plurality of time points and a plurality of microscopic images of the control organoid using a first preset scoring standard; a second toxicity scoring module, configured to: obtain, based on a plurality of microscopic images of the experimental organoids, a pixel area of each microscopic image of the experimental organoids using a machine learning algorithm, and obtain a second toxicity score of the experimental organoids under the target compound using a second preset scoring standard; Structural parameter acquisition module, used to: obtain molecular descriptors of experimental organoids; The toxic effect evaluation module is used to: evaluate the toxic effect of the experimental organoid under the target compound based on the molecular descriptor of the experimental organoid and the first toxicity score and the second toxicity score of the experimental organoid under the target compound.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the organoid toxic effect evaluation method according to any one of claims 1 to 7 is implemented.
10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the organoid toxic effect evaluation method according to any one of claims 1 to 7.