A method for quantitative evaluation and visualization of trilineage differentiation potential of mesenchymal stem cells induced by biomaterials

Abstract

The application relates to the technical field of single-cell transcriptomics, biomaterials and bioinformatics, and particularly relates to a method for quantitatively evaluating and visualizing the three-lineage differentiation potential of mesenchymal stem cells induced by biomaterials, which comprises the following steps: obtaining single-cell transcriptome data of a target cell sample; calculating a lineage activity score of each target differentiation lineage and an undifferentiated state score of an undifferentiated state of a single cell based on specific gene sets of each target differentiation lineage and the undifferentiated state; calculating a relative lineage bias and a differentiation degree based on the lineage activity score and the undifferentiated state score, determining a differentiation dominant type of the single cell and counting a population proportion; and mapping the relative lineage bias and the differentiation degree to a multi-dimensional space coordinate system to generate a visual model. The application realizes accurate quantification and intuitive evaluation of the heterogeneity of the multi-lineage differentiation proportion at a single-cell resolution.

Description

Technical Field

[0001] This application relates to the interdisciplinary fields of single-cell transcriptomics, biomaterials and bioinformatics, specifically to a method for quantitative evaluation and visualization of the trilineage differentiation potential of mesenchymal stem cells induced by biomaterials. Background Technology

[0002] In the field of osteochondral tissue engineering, mesenchymal stem cells (MSCs) are widely used as seed cells due to their multi-lineage differentiation potential. Biomaterials, acting as cell carriers and microenvironment regulators, directly influence the regeneration and repair outcomes by inducing MSCs to differentiate into osteoblasts, chondrocytes, and adipocytes. The trilineage differentiation of MSCs does not proceed in a single direction; under the same induction conditions, they may differentiate into multiple lineages simultaneously, forming a complex differentiation ratio. Precise control of this differentiation ratio is crucial for successful osteochondral regeneration—the coordinated ratio of osteoblasts and chondrocytes determines the quality of tissue interface integration, while abnormal adipogenic differentiation can lead to adverse outcomes such as fatty infiltration. Therefore, establishing an evaluation method that can simultaneously quantify the trilineage differentiation ratio is of significant value for the functional screening and optimized design of biomaterials.

[0003] Currently, the biological evaluation of medical biomaterials mainly relies on the ISO 10993 series of standards, which focus on biosafety evaluation and lack standardized evaluation methods for functional efficacy. In evaluating stem cell differentiation efficacy, existing in vitro evaluations often use single endpoint indicators, such as alkaline phosphatase activity reflecting osteogenic differentiation, type II collagen expression reflecting chondrogenic differentiation, and Oil Red O staining reflecting adipogenic differentiation. These methods can only independently confirm the existence of a specific differentiation direction and cannot simultaneously analyze the proportional relationship of three differentiation lines within the same cell population. While traditional histochemical staining (such as Alizarin Red S, Alecidine Blue, and Oil Red O) combined with image analysis can provide semi-quantitative evaluation, it suffers from problems such as uneven staining depth, subjective threshold setting, and difficulty in multi-parameter correlation analysis. Although real-time quantitative PCR can detect marker gene expression, there is no direct correlation between gene expression levels and cell differentiation ratios. More importantly, the above methods mostly detect the average value of marker expression at the population level. This averaged data masks the heterogeneity of the cell population and cannot distinguish whether "all cells are differentiated to a moderate degree" or "some cells are fully differentiated while others are undifferentiated". The latter has completely different biological significance for evaluating the function of materials.

[0004] In terms of transcriptomics-based evaluation methods, existing technologies include schemes for predicting the function of biological materials using population-level transcriptomics data. For example, Chinese patent application CN113604544A discloses a method for predicting and evaluating the function of biological materials based on population transcriptomics, which extracts total RNA from cells for sequencing and uses a machine learning model to calculate the confidence level of samples for different cell types. However, this method relies on the average signal at the population level, averaging the expression data of heterogeneous cell populations, and requires multiple independent induction experiments. It cannot reflect the competitive relationship of differentiation among the three lineages in the same microenvironment, and its output is a discrete four-level confidence score, which is essentially a simple classification and difficult to characterize the continuity and proportional pattern of differentiation. Chinese patent application CN116884484A uses single-cell sequencing technology to evaluate the biological function of mesenchymal stem cells, but its output differentiation potential is a comprehensive score based on non-lineage-specific genes, lacking lineage specificity and unable to analyze the direction of cell differentiation into a specific lineage or the competitive relationship between multiple lineages. Chinese patent application CN120290682A relies on traditional experimental methods such as flow cytometry and histochemical staining for multi-dimensional weighted scoring. Its three-lineage differentiation detection is only a qualitative judgment, with low throughput and strong subjectivity. Different laboratories use different induction protocols, detection time points and evaluation standards, resulting in a lack of standardized quantitative system for evaluation results, making it difficult to make horizontal comparisons.

[0005] In summary, existing technologies are all unable to simultaneously and quantitatively analyze the proportional patterns and continuous changes of multiple target differentiation lineages at single-cell resolution, and there is also a lack of quantitative methods to integrate the three-lineage differentiation data into standardized comparable indices, which restricts the accurate evaluation and optimized design of the induction efficacy of biomaterials. Summary of the Invention

[0006] To address the problem that existing technologies can mask cellular heterogeneity and thus distort the evaluation of multi-lineage differentiation ratios, this application proposes a quantitative evaluation and visualization method for the trilineage differentiation potential of mesenchymal stem cells induced by biomaterials. This method enables precise quantification of multi-lineage differentiation ratios and intuitive evaluation of heterogeneity at single-cell resolution.

[0007] To achieve the above objectives, this application adopts the following technical solution: In a first aspect, this application provides a method for quantitative evaluation and visualization of the trilineage differentiation potential of mesenchymal stem cells induced by biomaterials, comprising: acquiring single-cell transcriptome data of target cell samples; calculating the lineage activity score of a single cell in each target differentiation lineage and the undifferentiated state score in the undifferentiated state based on the specific gene sets of each target differentiation lineage and undifferentiated state; calculating the relative lineage bias and differentiation degree based on the lineage activity score and undifferentiated state score, determining the dominant differentiation type of a single cell and statistically analyzing the population proportion of each dominant differentiation type in the target cell sample; and mapping the relative lineage bias and differentiation degree to a multidimensional spatial coordinate system to generate a visualization model for displaying the lineage competition pattern and differentiation continuum.

[0008] The above scheme overcomes the masking of cell heterogeneity by population averaging by independently calculating the activity scores of each lineage and undifferentiated state at the single-cell level, and combining relative bias and differentiation degree to determine the dominant type and visualize the mapping. It achieves synchronous, continuous and quantitative analysis of the differentiation ratio of multiple lineages.

[0009] As one implementation method, based on the specific gene sets of each target differentiation lineage and undifferentiated state, the lineage activity score of a single cell in each target differentiation lineage and the undifferentiated state score in the undifferentiated state are calculated, including: using a gene set enrichment analysis algorithm, based on the gene expression profile of a single cell, calculating the enrichment degree of each specific gene set; and based on the enrichment degree, obtaining the lineage activity score of a single cell in each target differentiation lineage and the undifferentiated state score in the undifferentiated state.

[0010] This implementation method transforms high-dimensional gene expression data into a lineage activity measure with clear biological significance through a gene set enrichment analysis algorithm, avoiding the one-sidedness of evaluation by a single marker and improving the accuracy and stability of differentiation status identification.

[0011] As one implementation method, the target differentiation lineage includes osteoblastic lineage, chondrogenic lineage, and adipogenic lineage, and the undifferentiated state is the mesenchymal stem cell state; the specific gene set is obtained by means of: extracting specific gene sets corresponding to osteoblastic lineage, chondrogenic lineage, adipogenic lineage, and mesenchymal stem cell state from the population transcriptome database through differential expression analysis.

[0012] This implementation method specifically limits the scope of protection to the trilineage differentiation scenario of mesenchymal stem cells, and ensures the lineage specificity of the gene set through differential expression analysis, providing a reliable basis for subsequent accurate scoring.

[0013] As one implementation method, relative lineage bias and degree of differentiation are calculated based on lineage activity score and undifferentiated state score, including: determining relative lineage bias based on the relative proportion of the lineage activity score of each target differentiated lineage in the total lineage activity, where the total lineage activity is the sum of the lineage activity scores of each target differentiated lineage; and determining the degree of differentiation deviation of a single cell from the undifferentiated state based on the relationship between the undifferentiated state score and the total lineage activity, and using the degree of differentiation deviation as the degree of differentiation.

[0014] This implementation eliminates the influence of absolute expression level differences through normalization, making the data between different samples comparable. At the same time, it defines the degree of differentiation as the degree of deviation from the undifferentiated state, which is consistent with the biological nature of stem cell differentiation.

[0015] As one implementation method, the formula for calculating relative spectral bias is: in, , and These represent the relative lineage bias of cells compared to the osteogenic, chondrogenic, and adipogenic lineages, respectively. , and These represent the lineage activity fractions of the osteogenic, chondrogenic, and adipogenic lineages, respectively; the formula for calculating the degree of differentiation is: in, Indicates the degree of differentiation. Indicates the undifferentiated state score; determines the differentiation dominance type of a single cell, including: when , and The maximum value in is Furthermore, when a single cell is not located in the central region of the multidimensional spatial coordinate system, it is determined to be of the osteogenic lineage-dominant type; when the maximum value is Furthermore, when a single cell is not located in the central region, it is determined to be of the chondrogenic lineage-dominant type; when the maximum value is Furthermore, when a single cell is not located in the central region, it is determined to be an adipogenic lineage-dominant type; when the following conditions are met... , and When the condition is met, it is determined to be a non-dominant type, where 0.1 is the preset center threshold.

[0016] This implementation provides a standardized multi-parameter scoring system and clear rules for determining the dominant lineage type, transforming the evaluation of differentiation status from subjective qualitative to objective quantitative. Furthermore, the setting of a preset center threshold effectively distinguishes between truly unbiased cells and weakly biased cells, avoiding classification ambiguity.

[0017] As one implementation method, the multidimensional spatial coordinate system is a ternary graph coordinate system; the relative spectral bias and differentiation degree are mapped to the multidimensional spatial coordinate system to generate a visualization model for displaying the spectral competition pattern and differentiation continuum, including: mapping the relative spectral bias of the three target differentiation spectral systems to the plane coordinates of the three vertices of the ternary graph coordinate system respectively; mapping the differentiation degree to the vertical depth axis of the ternary graph coordinate system to construct a three-dimensional ternary visualization model.

[0018] This implementation utilizes the natural suitability of ternary graphs for expressing the competitive relationships among the three components, combined with the vertical dimension to characterize the differentiation process, to achieve simultaneous visualization of lineage bias and differentiation degree, and to intuitively present the distribution pattern of cells in the multi-directional differentiation space.

[0019] As one implementation method, generating a visualization model for displaying the pedigree competition pattern and differentiation continuum also includes: generating a pedigree composition statistical ring chart based on the population proportion of each dominant differentiation type in the target cell sample; generating a differentiation score violin chart based on the pedigree activity score and differentiation degree of each target differentiation pedigree in the target cell sample; and combining the three-dimensional ternary visualization model, the pedigree composition statistical ring chart, and the differentiation score violin chart to generate a visualization evaluation report.

[0020] This implementation method, through a combination of multi-dimensional charts, preserves the details of heterogeneity at the single-cell level while providing a statistical overview at the population level, meeting the analytical needs at different levels and enhancing the comprehensiveness and interpretability of the evaluation results.

[0021] In one implementation method, the target cell sample is a mesenchymal stem cell sample cultured with biological materials under conditions free of exogenous inducing factors; this method is used to evaluate the induction efficacy of biological materials on the multi-lineage differentiation potential of mesenchymal stem cell samples. Those skilled in the art will understand that the biological materials to be evaluated include, but are not limited to, different types of materials, or the same material modified with different composition ratios, different microstructures, or different surface treatments, or even any exogenous additive factors that will be used to induce mesenchymal stem cell differentiation in the future.

[0022] This implementation method is applied to the evaluation of the intrinsic induction efficacy of biomaterials, eliminating the interference of exogenous inducing factors, so that the evaluation results truly reflect the ability of the material microenvironment to regulate cell fate, and is suitable for material screening and optimization.

[0023] Secondly, this application also provides a device for quantitative evaluation and visualization of the trilineage differentiation potential of biomaterial-induced mesenchymal stem cells, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the method for quantitative evaluation and visualization of the trilineage differentiation potential of biomaterial-induced mesenchymal stem cells as described above.

[0024] The aforementioned device automates the execution of the aforementioned methods through a hardware carrier, making it easy to integrate into biomaterials R&D platforms or stem cell quality control systems, thereby improving evaluation efficiency and standardization.

[0025] Thirdly, this application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method for quantitative evaluation and visualization of the trilineage differentiation potential of biomaterial-induced mesenchymal stem cells as described above.

[0026] The beneficial effects of the technical solution in this application are reflected in: This application overcomes the masking effect of population averaging on cellular heterogeneity by independently calculating the activity scores of each target differentiation lineage and undifferentiated state at the single-cell level, and combining normalized relative bias and differentiation deviation to determine the dominant type. This avoids the risk of false positives or false negatives due to signal dilution, significantly improving the accuracy and resolution of differentiation ratio evaluation. Furthermore, by mapping relative lineage bias to a ternary graph plane coordinate and differentiation degree to a vertical depth axis, a three-dimensional ternary visualization model is constructed, simultaneously presenting the multi-lineage competition pattern and differentiation continuum. This allows for an intuitive and quantitative expression of complex single-cell differentiation states, solving the technical challenge of existing technologies that cannot simultaneously consider lineage bias and differentiation process in a single view. Simultaneously, based on a standardized scoring system and clear lineage dominance type determination rules, a cross-sample and cross-laboratory comparable evaluation benchmark is established, providing a reliable tool for high-throughput screening and precise optimization of biomaterial induction efficacy, especially suitable for the objective evaluation of material function under conditions without exogenous inducing factors.

[0027] In summary, this application develops an evaluation method that "transforms from population to single cell, from discrete classification to continuous proportion, from single indicator to multidimensional index, and from invisible to visualized," overcoming the fundamental defect of "population averaging masking heterogeneity" and achieving quantitative analysis of trilineage differentiation ratios at single-cell resolution. This method accurately assesses cell differentiation tendency at single-cell resolution, improving the accuracy and stability of differentiation potential evaluation. This application can be used for the evaluation and comparative analysis of trilineage differentiation potential in mesenchymal stem cells from different sources, biomaterial-induced cells, and tissue engineering-related samples, providing a quantitative evaluation tool for biomaterial screening, tissue regeneration research, and stem cell quality control. Furthermore, this method is highly versatile, widely applicable, and scalable, providing technical support for cell function evaluation and intelligent prediction in the field of regenerative medicine. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the existing technology for evaluating the differentiation of biomaterial-induced stem cells based on population-level biomarker detection.

[0029] Figure 2 This is a comparison diagram of the differentiation evaluation mode between the MeD-P method in the prior art and the embodiment of this application.

[0030] Figure 3 This is a method and system architecture diagram for quantitative evaluation and visualization of the trilineage differentiation potential of biomaterial-induced mesenchymal stem cells according to embodiments of this application.

[0031] Figure 4This application presents a three-dimensional ternary diagram of the three-lineage differentiation potential of a multi-dataset trilineage and a statistical chart of the proportion of lineages. A. A three-dimensional ternary plot integrated from four public datasets; B. Cell lineage percentages from the four public datasets; C. A three-dimensional ternary plot and violin plot of cell differentiation scores from the GSE113253 dataset, colored by sample; D. A three-dimensional ternary plot from the GSE113253 dataset, colored by lineage; E. Cell lineage percentages from three samples in the GSE113253 dataset; F. A three-dimensional ternary plot and violin plot of cell differentiation scores from the PRJNA748552 dataset, colored by sample; G. A three-dimensional ternary plot from the PRJNA748552 dataset, colored by lineage; H. Cell lineage percentages from three samples in the PRJNA748552 dataset; I. A three-dimensional ternary plot and violin plot of cell differentiation scores from the GSE143753 dataset, colored by sample; J. A three-dimensional ternary plot from the GSE143753 dataset, colored by lineage; K. Cell lineage percentages from three samples in the GSE143753 dataset; L. M. Three-dimensional ternary plot of GSE160625 dataset colored by sample and violin plot of cell differentiation score; N. Three-dimensional ternary plot of GSE160625 dataset colored by lineage; N. Cell lineage percentage of three samples in GSE160625 dataset.

[0032] Figure 5 This is a graph showing the quantitative evaluation results of the three-lineage differentiation potential of mesenchymal stem cells in an embodiment of this application. A. Three-dimensional ternary graph of the Blank group lineage dominance and its projection at different angles; B. Proportion of each lineage dominance in the Blank group; C. Three-dimensional ternary graph of the Os_D7 group lineage dominance and its projection at different angles; D. Proportion of each lineage dominance in the Os_D7 group; E. Comparison of Blank and Os_D7 lineage dominance; F. Comparison of osteogenic activity and differentiation degree of MSCs in the Blank and Os_D7 groups.

[0033] Figure 6 This is a BTO / P(VDF-TrFE) membrane characterization and stem cell differentiation evaluation diagram according to an embodiment of this application. A. Three-dimensional ternary diagram of the UP_D7 lineage-dominated group and its projection at different angles; B. Proportion of each lineage dominated by the UP_D7 lineage; C. Three-dimensional ternary diagram of the P_D7 lineage-dominated group and its projection at different angles; D. Proportion of each lineage dominated by the P_D7 lineage; E. Difference in lineage proportion between UP_D7 and P_D7; F. Comparison of osteogenic activity and differentiation degree of MSCs in the UP_D7 and P_D7 groups. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0035] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0036] Terminology Explanation scGSEA (Single-cell gene set enrichment analysis): It is a computational method for analyzing single-cell RNA sequencing (scRNA-seq) data. Its core objective is to assess the activity level of a specific gene set (such as a biological pathway or function) at the resolution of a single cell.

[0037] Mesenchymal stem cells (MSCs) are a type of pluripotent adult stem cells found in various human tissues. Their most crucial characteristic is their ability to both self-renew (continuously replicate themselves) and differentiate into various mesodermal-derived cell types. Under specific induction conditions, they can differentiate into osteoblasts (bone repair), chondrocytes (joint repair), adipocytes (energy storage), and tendons, muscles, etc.

[0038] Example 1 like Figure 3 As shown, this embodiment provides a method for quantitative evaluation and visualization of the trilineage differentiation potential of mesenchymal stem cells induced by biological materials. This method constructs a processing framework from single-cell data to multidimensional visualization evaluation. It should be understood that although this embodiment is described in a specific process order, in practical applications, the execution order of some steps can be adjusted according to computing resources or data processing needs, as long as the logical dependencies are maintained. The method mainly includes the following steps: Step S100: Obtain single-cell transcriptome data of the target cell sample.

[0039] Single-cell transcriptome data are digital matrices obtained through high-throughput sequencing technology, reflecting gene expression levels within each individual cell. This step introduces a single-cell resolution data source, enabling subsequent analysis to capture subpopulation differences and asynchronous differentiation within the cell population.

[0040] Considering that traditional evaluation methods based on the average expression level of the entire cell population are insufficient to distinguish the differentiation states of different cell subpopulations, this application uses a single cell as the smallest evaluation unit. For each cell, the activity of three lineages—osteogenic, chondrogenic, and adipogenic—is calculated separately. This enables quantitative analysis of cellular heterogeneity within the sample and further statistically analyzes the proportion of dominant cells in each lineage, more accurately reflecting the overall differentiation characteristics of the biomaterial after induction. The high-throughput sequencing platform is not particularly limited, as long as it can acquire single-cell transcriptome sequencing data.

[0041] Step S200: Based on the specific gene sets of each target differentiation lineage and undifferentiated state, calculate the lineage activity score of a single cell in each target differentiation lineage and the undifferentiated state score in the undifferentiated state.

[0042] A specific gene set is a pre-constructed set of molecular tags that can represent a specific cell fate or state. In this embodiment, not only is the lineage activity score, representing the direction of differentiation, calculated, but an undifferentiated state score, representing the initial state, is also specifically introduced. The lineage activity score characterizes the magnitude of the driving force for cell development towards a specific terminal direction, while the undifferentiated state score characterizes the degree to which the cell retains stem or progenitor cell characteristics. By calculating these two types of scores independently, misjudgments that may occur when determining differentiation state based on only a single dimension are avoided.

[0043] Step S300: Based on the lineage activity score and the undifferentiated state score, calculate the relative lineage bias and differentiation degree, determine the dominant differentiation type of a single cell, and count the population proportion of each dominant differentiation type in the target cell sample.

[0044] Relative lineage bias is a proportional parameter obtained by normalizing the activity scores of multiple lineages, used to reflect the directional selection of cells in the multi-lineage differentiation space. Differentiation degree is a scalar indicator derived by comprehensively considering lineage activity and undifferentiated state, used to quantify the overall process of cells leaving the primitive state and moving towards the terminal mature state. These two parameters make "where to differentiate" (direction) and "to what extent differentiation has occurred" (process) two orthogonal evaluation dimensions. Based on this, continuous values ​​are mapped to discrete dominant differentiation types according to preset rules, and the proportion of each type in the entire sample is further statistically analyzed.

[0045] Step S400: Map the relative phylogenetic bias and degree of differentiation to a multi-dimensional coordinate system to generate a visualization model for displaying the phylogenetic competition pattern and the differentiation continuum.

[0046] In this embodiment, the coordinate system is configured to simultaneously encode directional and process information. Planar coordinates represent the competitive balance between different lineages, while the vertical or depth axis represents the evolution of differentiation. Through this mapping, researchers can observe the distribution characteristics of cell populations migrating from the central undifferentiated region to the peripheral differentiated region, as well as the competitive dynamics between different lineages, in a single view.

[0047] Specifically, this embodiment provides a method for quantitatively analyzing and visually evaluating the trilineal differentiation potential of mesenchymal stem cells induced by biomaterials. The method first extracts characteristic gene sets for osteogenic, chondrogenic, and adipogenic differentiation based on a population transcriptome database of mesenchymal stem cell trilineal differentiation. Then, it constructs a single-cell-level trilineal differentiation reference atlas using single-cell transcriptome sequencing data and scores individual cells based on the aforementioned characteristic gene sets to quantitatively analyze their lineage differentiation tendency. Subsequently, it statistically analyzes the distribution characteristics of trilineal differentiation scores for all cell populations in each sample, generating a visual report on the trilineal differentiation potential evaluation of the samples. In application, only single-cell transcriptome sequencing data of mesenchymal stem cells cultured with biomaterials are required to achieve a systematic evaluation of the trilineal differentiation performance of stem cells induced by biomaterials. The core technical route is as follows: Figure 3 As shown.

[0048] Example 2 Based on the general evaluation framework provided in Example 1, this example further specifies the calculation mechanism for lineage activity score and undifferentiated state score to be applicable to the evaluation scenario of trilineage differentiation potential of mesenchymal stem cells. As one implementation method, a gene set enrichment analysis algorithm is used to calculate the enrichment degree of each specific gene set based on the gene expression profile of a single cell; based on the enrichment degree, the lineage activity score of a single cell in each target differentiation lineage and the undifferentiated state score in the undifferentiated state are obtained.

[0049] This embodiment uses single-cell gene set enrichment analysis (scGSEA) as the specific implementation of the aforementioned gene set enrichment analysis algorithm. scGSEA is specifically optimized for the sparsity and high noise characteristics of single-cell transcriptome data, enabling it to accurately capture the activation state of specific biological pathways within a single cell without relying on external reference samples. During execution, the system takes the standardized gene expression matrix of each cell as input and performs enrichment scoring on four predefined specific gene sets. This independent scoring strategy allows a cell to simultaneously possess high levels of osteogenic activity and a certain degree of chondrogenic activity, or to maintain a high undifferentiated state score while initiating a slight lineage bias, truly reflecting the biological essence of multi-lineage gene co-expression in the early stages of stem cell differentiation. Through this step, each cell is assigned four independent numerical labels: osteogenic lineage activity score (Os), chondrogenic lineage activity score (Ch), adipogenic lineage activity score (Ad), and mesenchymal stem cell state score (MSC), providing the original data foundation for subsequent quantification of relative bias and differentiation degree.

[0050] Furthermore, to ensure that the above scoring has a clear biological orientation, this embodiment specifically defines the target differentiation lineage and undifferentiated state. As one implementation method, the target differentiation lineage includes osteogenic, chondrogenic, and adipogenic lineages, and the undifferentiated state is the mesenchymal stem cell state; the specific gene set is obtained by: extracting specific gene sets corresponding to the osteogenic, chondrogenic, adipogenic, and mesenchymal stem cell states from a population transcriptome database through differential expression analysis.

[0051] In the specific implementation plan, four public population transcriptome datasets—GSE113253, PRJNA748552, GSE143753, and GSE160625—covering different tissue origins and induction conditions, were retrieved and integrated. For each dataset, cell identity annotation and quality control were first performed. Subsequently, differential expression analysis algorithms were used to screen for marker genes that were highly expressed in specific lineages but lowly expressed in other lineages and undifferentiated MSCs. Specific gene sets corresponding to osteogenic, chondrogenic, adipogenic, and mesenchymal stem cell states were constructed respectively.

[0052] Specifically, four independent datasets (GSE113253, PRJNA748552, GSE143753, and GSE160625) were obtained by searching public databases. After cell identity annotation, cell populations associated with osteogenic, chondrogenic, and adipogenic differentiation were extracted, and the scGSEA algorithm was used to score the gene characteristics of the three differentiation lines. Based on the distribution characteristics of the three differentiation lines scores of the cell populations, the cells were divided into four lineage types: osteogenic-dominant, chondrogenic-dominant, adipogenic-dominant, and no dominant differentiation direction. The proportion of each lineage type in each sample was then statistically analyzed. The results are as follows: Figure 4 As shown, the aforementioned public dataset covers the cellular composition of trilineage differentiation and can serve as a reference atlas of stem cell trilineage differentiation. Figure 4 As shown in Figure A, the three-dimensional ternary plot after integrating the four public datasets demonstrates the distribution characteristics of cells in the trilineal differentiation space of each dataset. Figure 4 The stacked bar chart shown in Part B illustrates the cell lineage distribution across the four public datasets. Three-dimensional ternary plots and violin plots of cell differentiation scores for each dataset are also included. Figure 4 (as shown in C, F, I, L), and a three-dimensional ternary diagram colored by spectral system (e.g.) Figure 4 (as shown in D, G, J, M) and the proportion of cell lineages in each sample (e.g.) Figure 4 (As shown in E, H, K, N) further demonstrates the distribution characteristics of trilineage differentiation scores and the consistency of lineage composition of different cell subpopulations in each dataset.

[0053] Example 3 Building upon the lineage activity scores and undifferentiated state scores obtained in Example 2, this example further specifies the calculation mechanism for relative lineage bias and differentiation degree, as well as the rules for determining the dominant differentiation type. As one implementation method, the relative lineage bias and differentiation degree are calculated based on the lineage activity scores and undifferentiated state scores, including: determining the relative lineage bias based on the relative proportion of the lineage activity score of each target differentiated lineage in the total lineage activity, where the total lineage activity is the sum of the lineage activity scores of each target differentiated lineage; and determining the degree of differentiation deviation of a single cell from the undifferentiated state based on the relationship between the undifferentiated state score and the total lineage activity, using the degree of differentiation deviation as the differentiation degree.

[0054] By dividing the activity score of each lineage by the sum of the activity scores of the three lineages, this embodiment transforms absolute expression levels into relative proportions, ensuring that the position of each cell in the trilineal differentiation space is determined solely by its intrinsic fate selection. Simultaneously, the calculation of differentiation degree incorporates the undifferentiated state score as a reference benchmark. By calculating the proportion of the undifferentiated state score in total transcriptomic activity (i.e., the sum of trilineal and MSC activities) and taking the complement, the overall process of cells leaving the primitive stem state and moving towards terminal differentiation is quantified.

[0055] To achieve a standardized and quantitative evaluation of differentiation status, this embodiment provides specific calculation formulas and judgment rules. As one implementation method, the calculation formula for relative lineage bias is: in, , and These represent the relative lineage bias of cells compared to the osteogenic, chondrogenic, and adipogenic lineages, respectively. , and These represent the lineage activity fractions of the osteogenic, chondrogenic, and adipogenic lineages, respectively; the formula for calculating the degree of differentiation is: in, Indicates the degree of differentiation. This represents the score of the undifferentiated state.

[0056] In the above formula system, , and The sum of the three parameters is always 1, intuitively reflecting the resource allocation pattern of cells in the three lineage differentiation directions. The Diff Score theoretically ranges from 0 to 1; a higher value indicates that the cell is further away from the MSC initiation state and has a higher degree of differentiation maturity, while a lower Diff Score indicates that the cell still maintains a high degree of stemness. The above parameters together constitute a standardized multi-parameter scoring system that can simultaneously evaluate lineage activity, lineage bias, and differentiation degree.

[0057] Based on the aforementioned quantitative indicators, this embodiment further defines four discrete differentiation dominance types to facilitate statistical description of population data. Determining the differentiation dominance type of a single cell includes: when... , and The maximum value in is Furthermore, when a single cell is not located in the central region of the multidimensional spatial coordinate system, it is determined to be of the osteogenic lineage-dominant type; when the maximum value is Furthermore, when a single cell is not located in the central region, it is determined to be of the chondrogenic lineage-dominant type; when the maximum value is Furthermore, when a single cell is not located in the central region, it is determined to be an adipogenic lineage-dominant type; when the following conditions are met... , and When the condition is met, it is determined to be a non-dominant type, where 0.1 is the preset center threshold.

[0058] This judgment logic addresses biologically ambiguous states by introducing the concept of a "central region." In this embodiment, a central threshold of 0.1 is set, and a lineage is only considered to have a clear dominant direction when its relative bias is significantly higher than that of the other two lineages; if the relative biases of the three lineages are close to each other (the difference is less than 0.1), they are judged to be non-dominant.

[0059] Specifically, osteogenic induction was performed on self-cultured hBMSCs (Os_D7 group), with the uninduced group (Blank group) serving as a control. Both groups underwent scRNA-seq and the data were input into the platform of this invention to generate a trilineage differentiation potential evaluation report. The results are as follows: Figure 5 As shown: Blank group 3D ternary plot ( Figure 5 A) showed that cells were mainly distributed in the central region, with a slight shift towards Ch; lineage dominant composition analysis ( Figure 5 Further analysis (B) showed that the cells were predominantly Non-dominant (60.9%) and Ch-dominant (38.0%), with almost no Os-dominant cells observed. This was combined with a comparison of the Blank and Os_D7 lineage dominance ( Figure 5 As can be seen from E), the initial state of MSCs mainly exhibits no clear lineage dominance, but there is a certain degree of pre-activation in the cartilage direction. Compared with the Blank group, the Os_D7 group showed a significant overall shift towards the Os direction (E). Figure 5 The proportion of Os-dominant (C) increased significantly to 36.9%, Ch-dominant accounted for 33.5%, and Non-dominant decreased significantly to 29.6%. Figure 5 The D), suggesting that osteogenic induction causes MSCs to shift from a state without a clear lineage dominance towards osteogenic direction, thus forming an osteogenic lineage dominance state. Furthermore, a comparison of bone activity and differentiation degree between Blank and Os_D7 components ( Figure 5 The F-squared results showed that the Os_D7 group had significantly higher Os scores (median increased from 1.3e-03 to 3.6e-03), Os_n (increased from 0.33 to 0.39), and Diff scores (increased from 0.67 to 0.69) than the Blank group. Each figure marked the difference between the two groups with a significance sign, indicating that osteogenic induction not only enhanced osteogenic-related activities but also promoted the overall differentiation degree and the formation of osteogenic lineage bias. In summary, MSCs are not in a completely directionless initial state but exhibit a certain degree of lineage pre-activation; osteogenic induction can further drive MSCs from a state without a clear lineage dominance towards osteogenic orientation, ultimately forming a clearly defined osteogenic lineage dominance state.

[0060] Example 4 Building upon the quantitative calculation of relative phylogenetic bias and differentiation degree completed in Example 3, this example further specifies the data visualization method. As one implementation, the multidimensional spatial coordinate system is a ternary graph coordinate system. The relative phylogenetic bias and differentiation degree are mapped to the multidimensional spatial coordinate system to generate a visualization model for displaying the phylogenetic competition pattern and differentiation continuum. This includes: mapping the relative phylogenetic bias of the three target differentiation phylogenetic systems to the plane coordinates of the three vertices of the ternary graph coordinate system; and mapping the differentiation degree to the vertical depth axis of the ternary graph coordinate system to construct a three-dimensional ternary visualization model.

[0061] Since the sum of the three parameters Os_n, Ch_n, and Ad_n is always 1, the ternary graph naturally fits this three-component competitive system. Its three vertices represent the dominant osteogenic, chondrogenic, or adipogenic states, respectively, while the center point represents the equilibrium of the three lineages. Mapping the DiffScore to the vertical depth axis (Z-axis), the Z-axis height characterizes the process of cells leaving the undifferentiated state. Through this dual information encoding of "planar coordinates = lineage competition" and "vertical axis = differentiation process," those skilled in the art can simultaneously observe the fate selection direction and maturity evolution trajectory of the cell population in a single view. Specifically, Os_n, Ch_n, and Ad_n are respectively used as ternary parameters... Figure 3 The coordinates corresponding to each vertex are used to display the relative positions of different cells in the three-lineage differentiation space through a ternary graph. At the same time, the degree of cell differentiation is mapped to the vertical direction using the Diff Score as the height of the three-dimensional coordinate axis, thereby constructing a three-dimensional ternary visualization model to realize the synchronous display of lineage bias and degree of differentiation, which can more intuitively reflect the overall differentiation pattern induced by different biological materials.

[0062] Furthermore, to compensate for the shortcomings of a single 3D view in displaying statistical overviews and detailed distributions, this embodiment also provides a mechanism for generating a combined evaluation report. As one implementation method, generating a visualization model for displaying the pedigree competition pattern and differentiation continuum further includes: generating a pedigree composition statistical ring chart based on the population proportions of each dominant differentiation type in the target cell sample; generating a differentiation score violin chart based on the pedigree activity score and differentiation degree of each target differentiation pedigree in the target cell sample; and combining the 3D ternary visualization model, the pedigree composition statistical ring chart, and the differentiation score violin chart to generate a visualization evaluation report.

[0063] The three-dimensional ternary plot preserves the individual information of each cell, the lineage composition statistical ring chart visually presents the precise proportion of the four dominant types from a macro-statistical perspective, and the differentiation score violin plot can finely display the activity scores of each lineage and the distribution characteristics of the Diff Score in different cell subpopulations. Figure 3 The system output module shown and Figures 4 to 6Based on the actual evaluation results, these three types of charts were integrated into the same visual evaluation report, forming a multi-dimensional evidence chain from "individual status positioning" to "group structure overview" and then to "molecular characteristic distribution".

[0064] In one implementation method, the target cell sample is a mesenchymal stem cell sample cultured with biological materials under conditions without exogenous inducing factors; this method is used to evaluate the induction efficacy of biological materials on the multi-lineage differentiation potential of mesenchymal stem cell samples.

[0065] In this application scenario, "no exogenous inducing factor conditions" refers to the cell culture system without the addition of any growth factors, hormones, or small molecule compounds that promote specific lineage differentiation. Cell fate determination is driven solely by the microenvironmental signals provided by the physical topology, charge distribution, or chemical components of the biomaterial surface. Specifically, a BTO / P (VDF-TrFE) nanocomposite membrane was prepared using a casting method, in which BTO NPs were uniformly distributed in the basement membrane and polarized (P), resulting in a uniform potential distribution on the composite membrane surface. hBMSCs were seeded onto unpolarized (UP) and polarized (P) composite membranes, respectively. After 7 days of culture, cells were collected for scRNA-seq, and the results were input into the platform of this invention to generate a three-lineage differentiation visualization evaluation report. The UP_D7 group's three-dimensional ternary plot ( Figure 6 A) showed that cells were mainly distributed in the central region, with only a slight shift towards the Os direction; lineage dominant composition analysis ( Figure 6 B) indicates that the UP group was dominated by Non-dominant (55.7%) and Ch-dominant (41.2%), with Os-dominant accounting for only 2.4%, suggesting that most cells have not yet formed a clear osteogenic lineage. In contrast, the three-dimensional ternary plot of the P_D7 group ( Figure 6 The C) showed that the cells as a whole were closer to the Os apex than the UP_D7 group, and the osteogenic direction shift was more obvious; the dominant lineage composition ( Figure 6 The results (D) show that the P group is still dominated by non-dominant (53.5%) and Ch-dominant (44.0%), while the proportion of Os-dominant is similar to that of the UP group. A comparison of the dominance of the UP_D7 and P_D7 lineages ( Figure 6 The E-value further indicates that the overall composition of the two lineage-dominant groups is similar, and it is difficult to distinguish the induction differences between the two materials by simply relying on the lineage dominance ratio. Therefore, the osteogenic activity and differentiation degree of the two groups of cells were further compared. Figure 6 The results showed that the Os score of the P group was significantly higher than that of the UP group (the mean increased from 8.8e-04 to 1.3e-03, as shown in the F). Figure 6 As shown in FI), the Diff score increased from 0.67 to 0.71 (as shown in FI). Figure 6As shown in F-III), P treatment significantly enhanced osteogenic activity and overall differentiation of cells. Although the change in Os_n was small (mean values ​​for both groups were 0.30, as shown in F-III), it indicates that P treatment significantly enhanced osteogenic activity and overall differentiation. Figure 6 As shown in F-II), but combined with the ternary graph distribution ( Figure 6 As shown in C), cells in group P were closer to the Os apex, exhibiting a stronger osteogenic shift. In summary, both groups UP and P promoted lineage shift of MSCs towards osteogenic direction, but group P showed higher osteogenic-related activity and overall differentiation degree, suggesting that it has a stronger osteogenic induction potential.

[0066] Example 5 like Figure 3 As shown, this embodiment provides a device for quantitative evaluation and visualization of the trilineage differentiation potential of mesenchymal stem cells induced by biomaterials. This device is the hardware carrier of the aforementioned method for quantitative evaluation and visualization of the trilineage differentiation potential of mesenchymal stem cells induced by biomaterials, aiming to transform the abstract data processing flow into a concrete physical product. As one implementation, the device includes a memory and a processor.

[0067] Specifically, the memory is used to store computer programs. Here, a computer program refers to a set of instruction code that implements the steps described in any of the above method embodiments (Embodiments 1 to 4). In practical deployment, the memory may include volatile memory, such as random access memory (RAM), for temporarily loading programs and intermediate computational data (such as single-cell expression matrices, lineage activity score matrices, etc.) during device operation; it may also include non-volatile memory, such as read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), or flash memory, for persistently storing the operating system, application modules, and a pre-built database of specific gene sets. It should be understood that the specific selection and capacity configuration of the memory depend on the scale and computational complexity of the single-cell data to be processed, as long as it can meet the functional requirements of program storage and data reading and writing.

[0068] When the processor is used to execute a computer program, it implements a method for quantitative evaluation and visualization of the trilineage differentiation potential of biomaterial-induced mesenchymal stem cells as described in any of the above method embodiments.

[0069] Furthermore, this embodiment also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method for quantitative evaluation and visualization of the trilineage differentiation potential of biomaterial-induced mesenchymal stem cells as described in any of the above method embodiments.

[0070] Specifically, a computer-readable storage medium refers to any tangible medium capable of containing, storing, transmitting, or disseminating a program for use by an instruction execution system. Its specific forms include, but are not limited to, disks (such as floppy disks and hard disks), optical disks (such as CD-ROMs and DVDs), and semiconductor memories (such as USB flash drives, SD cards, and solid-state drives). Unlike transient signals, the storage medium emphasized in this application possesses non-transient physical properties and can preserve program code for a long period. The significance of this implementation lies in covering the independent distribution channels of software products: even if a complete hardware device containing a processor is not sold, distributing a medium storing the method of this application only in the form of a software installation package, firmware update package, or cloud image still falls within the scope of protection. When a user loads the program in this medium onto any compatible computing device and runs it, the device then possesses the ability to execute the quantitative evaluation and visualization method for the induction of trilineage differentiation potential of mesenchymal stem cells using biological materials, as described in this application, thereby achieving standardized and automated evaluation of the induction efficacy of biological materials. It should be understood that, with the development of technology, any new storage technologies that emerge in the future (such as DNA storage, holographic storage, etc.) that can achieve the same program storage and retrieval functions should be regarded as equivalent alternatives to the computer-readable storage medium of this application.

[0071] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for quantitative evaluation and visualization of the trilineage differentiation potential of mesenchymal stem cells induced by biomaterials, characterized in that, The method includes: Obtain single-cell transcriptome data from the target cell sample; Based on the specific gene sets of each target differentiation lineage and undifferentiated state, the lineage activity score of a single cell in each target differentiation lineage and the undifferentiated state score in the undifferentiated state are calculated. Based on the lineage activity score and the undifferentiated state score, the relative lineage bias and differentiation degree are calculated to determine the dominant differentiation type of the individual cell and to count the population proportion of each dominant differentiation type in the target cell sample. The relative phylogenetic bias and the degree of differentiation are mapped to a multidimensional spatial coordinate system to generate a visualization model for displaying the phylogenetic competition pattern and the differentiation continuum.

2. The method according to claim 1, characterized in that, The calculation of the lineage activity score of a single cell in each target differentiation lineage and the undifferentiated state score in each undifferentiated state, based on the specific gene sets of each target differentiation lineage and undifferentiated state, includes: Using a gene set enrichment analysis algorithm, the enrichment degree of each specific gene set is calculated based on the gene expression profile of the individual cell; Based on the enrichment level, the lineage activity score of the individual cell in each of the target differentiation lineages and the undifferentiated state score in the undifferentiated state are obtained.

3. The method according to claim 2, characterized in that, The target differentiation lineage includes osteogenic lineage, chondrogenic lineage and adipogenic lineage, and the undifferentiated state is mesenchymal stem cell state; The method for obtaining the specific gene set includes: extracting the specific gene set corresponding to the osteogenic lineage, the chondrogenic lineage, the adipogenic lineage, and the mesenchymal stem cell state from the population transcriptome database through differential expression analysis.

4. The method according to claim 3, characterized in that, The calculation of relative lineage bias and differentiation degree based on the lineage activity score and the undifferentiated state score includes: The relative pedigree bias is determined based on the relative proportion of the pedigree activity score of each of the target differentiation lineages in the total pedigree activity, and the total pedigree activity is the sum of the pedigree activity scores of each of the target differentiation lineages. Based on the relationship between the undifferentiated state score and the total lineage activity, the degree of differentiation deviation of the individual cell from the undifferentiated state is determined, and the degree of differentiation deviation is taken as the degree of differentiation.

5. The method according to claim 4, characterized in that, The formula for calculating the relative spectral bias is: in, , and The relative lineage biases of the osteogenic, chondrogenic, and adipogenic lineages are respectively represented. , and These represent the lineage activity fractions of the osteogenic lineage, the chondrogenic lineage, and the adipogenic lineage, respectively. The formula for calculating the degree of differentiation is: in, Indicates the degree of differentiation. This represents the score of the undifferentiated state; The determination of the dominant differentiation type of the single cell includes: when , and The maximum value in is Furthermore, if the single cell is not located in the central region of the multidimensional spatial coordinate system, it is determined to be of the osteogenic lineage-dominant type; When the maximum value is Furthermore, when the single cell is not located in the central region, it is determined to be of the chondrogenic lineage-dominant type; when the maximum value is Furthermore, if the single cell is not located in the central region, it is determined to be of the adipogenic lineage-dominant type; when the following conditions are met... , and When the condition is met, it is determined to be a non-dominant type, where 0.1 is the preset center threshold.

6. The method according to claim 1, characterized in that, The multidimensional spatial coordinate system is a ternary graph coordinate system; The step of mapping the relative phylogenetic bias and the degree of differentiation to a multi-dimensional spatial coordinate system to generate a visualization model for displaying the phylogenetic competition pattern and the differentiation continuum includes: The relative spectral biases of the three target differentiation spectral systems are respectively mapped to the plane coordinates of the three vertices of the ternary graph coordinate system; The degree of differentiation is mapped to the vertical depth axis of the ternary graph coordinate system to construct a three-dimensional ternary visualization model.

7. The method according to claim 6, characterized in that, The generation of the visualization model for displaying the phylogenetic competition pattern and the differentiation continuum also includes: Based on the population proportion of each of the dominant differentiation types in the target cell sample, a phylogenetic composition statistical ring diagram is generated; Based on the lineage activity score and the degree of differentiation of each target differentiation lineage in the target cell sample, a differentiation score violin plot is generated; The three-dimensional ternary visualization model, the phylogenetic composition statistical ring chart, and the differentiation score violin chart are combined and output to generate a visualization evaluation report.

8. The method according to claim 1, characterized in that, The biomaterials include different types of materials, or the same material after different composition ratios, different microstructures, or different surface modification treatments; The method is used to evaluate the induction efficacy of the biomaterial on the multi-directional differentiation potential of the mesenchymal stem cell sample.

9. A device for quantitative evaluation and visualization of the trilineage differentiation potential of mesenchymal stem cells induced by biomaterials, characterized in that, The device includes: Memory, used to store computer programs; A processor, configured to execute the computer program to implement a method for quantitative evaluation and visualization of the trilineage differentiation potential of biomaterials induced by biological materials as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements a method for quantitative evaluation and visualization of the trilineage differentiation potential of mesenchymal stem cells induced by biological materials as described in any one of claims 1 to 8.

Images