A method and system for visualizing phenotypic data of potato germplasm resources

Abstract

The application discloses a kind of phenotypic data visualization analysis method and system of potato germplasm resources, it is related to agricultural information technology and data visualization technical field.The multidimensional dataset containing variety identification, multi-time point phenotypic trait value, environmental factor parameter and genetic distance is obtained;Variety is mapped to the first dimension axis of three-dimensional space based on genetic distance;Multi-phenotypic trait is weighted and projected to the second dimension axis;Time and environmental factor are complexly weighted and mapped to the third dimension axis;The control point of each variety in three-dimensional space is generated, and the phenotypic development trajectory curve is generated using spline interpolation;Three-dimensional space and trajectory curve are rendered to display device and provide interactive interface.Can simultaneously integrate four-dimensional information of genetic background, multi-phenotypic trait performance, time dynamics and environmental response, realize the dynamic visualization of phenotypic development process and the automatic identification of abnormal variety, significantly improve the analysis efficiency and cognitive intuitiveness of potato germplasm resources.

Description

Technical Field

[0001] This invention relates to the field of agricultural information technology and data visualization technology, and in particular to a method and system for visual analysis of phenotypic data of potato germplasm resources. Background Technology

[0002] Potatoes are the world's third largest food crop, and the collection, preservation, and evaluation of their germplasm resources are of great significance for ensuring food security and variety improvement. With the development of high-throughput phenotyping technology, researchers can obtain large-scale, multi-time-point, multi-trait, and multi-environmental potato phenotypic data. How to quickly identify superior germplasm, understand the relationships between traits, and analyze the interaction effects between genotype and environment from this high-dimensional, complex, and dynamic data has become a major challenge facing current potato breeding research.

[0003] Existing methods for visualizing potato phenotypic data mainly include: (1) Statistical chart-based methods: such as scatter plots, bar charts, box plots, etc., used to display the distribution characteristics of a single or a few traits. These methods are difficult to display the complex relationships between multiple traits at the same time, especially in terms of time dynamics and environmental response information. (2) Dimensionality reduction-based methods: such as principal component analysis (PCA), t-distributed random neighborhood embedding (t-SNE), etc., which project high-dimensional phenotypic data into two-dimensional or three-dimensional space. Although these methods can display the overall similarity between varieties, they lose the time dimension information, and the interpretability of the dimensionality reduction results is poor. (3) Genomics-based visualization methods: such as Manhattan plots used to display the results of genome-wide association analysis, which can locate genetic loci associated with phenotypes, but cannot intuitively display the comprehensive performance of varieties in multiple traits and their dynamic changes over time. (4) Biplot-based methods: such as genotype-genotype-environment interaction (GGE) biplots, used to analyze the adaptability and stability of varieties, but mainly for single traits such as final yield, and difficult to handle complex phenotypic data with multiple traits and multiple time points. (5) Network-based methods: such as chord diagrams, co-expression network diagrams, etc., used to show the relationship between traits or genes, but lack spatial location metaphors, making it difficult to intuitively understand the genetic background and environmental response patterns of varieties.

[0004] In summary, existing methods cannot simultaneously display information on four dimensions of a variety: genetic background, multiple phenotypic expression, temporal dynamics, and environmental response. Furthermore, they lack the ability to dynamically present the phenotypic development process, making it difficult to intuitively observe the phenotypic changes of a variety from emergence to harvest. Therefore, developing a visualization analysis method for potato germplasm resource phenotypic data that can integrate genetic distance, multi-time-point phenotypic traits, and environmental factors, and support dynamic interaction and anomaly detection, has significant practical implications and application value. To this end, a visualization analysis method and system for potato germplasm resource phenotypic data is proposed. Summary of the Invention

[0005] The main objective of this invention is to provide a method and system for visual analysis of phenotypic data of potato germplasm resources, aiming to solve at least one problem existing in the prior art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a method for visual analysis of phenotypic data of potato germplasm resources, comprising the following steps: S1: Obtain a multidimensional phenotypic dataset of potato germplasm resources, the dataset including variety identifiers, phenotypic trait values ​​at multiple time points, multiple environmental factor parameters, and genetic distance information between varieties; S2: Based on the genetic distance information, a multidimensional scaling algorithm is used to map each variety to the first dimension axis of three-dimensional space, so that the spatial distance between varieties is positively correlated with the genetic distance. S3: Normalize the phenotypic trait values ​​and map multiple phenotypic traits to the second dimension axis of three-dimensional space through weighted projection to form the trait coordinates of each variety at each time point; S4: Combine the time variable with at least one environmental factor parameter and weight them together, then map them to the third dimension axis of the three-dimensional space to form the environmental-time composite coordinates of each variety at each time point. S5: Using the first, second, and third dimensional axis coordinates as three-dimensional coordinate components, generate three-dimensional spatial control points for each variety at each time point. Then, use a spline interpolation algorithm to connect the control points of the same variety at different time points to generate the phenotypic development trajectory curves for each variety. ; S6: A dynamic environmental field is superimposed in the three-dimensional space. The environmental field is represented by radial basis functions and is used to show the spatial and temporal distribution changes of environmental factors. S7: Renders the generated 3D space and phenotypic development trajectory curves to the display device and provides a user interaction interface.

[0007] Preferred options also include: S8: Calculate the geometric characteristics of the phenotypic development trajectory curves of each variety, including trajectory length, average curvature and average torsion, and construct a multivariate feature vector based on the geometric characteristics; S9: The deviation of each variety's feature vector from the overall mean is calculated using Mahalanobis distance, and visual emphasis labels are generated for varieties whose deviation exceeds the adaptive threshold.

[0008] Preferably, it also includes a Pareto front extraction step, specifically: A multi-objective optimization problem is constructed based on at least two target traits, and the set of non-dominant varieties is selected as the Pareto front. The Pareto front is projected onto the plane formed by the second and third dimensional axes, and the front curve is fitted using spline interpolation, highlighting the front curve in three-dimensional space.

[0009] Preferred options also include: Three-dimensional reconstructed images of representative varieties at different growth stages in the field are collected in advance. The three-dimensional reconstructed images are used as textures and mapped to the three-dimensional spatial control point positions of the corresponding varieties at the corresponding time points. When the user selects a specific variety and a specific time point through the interactive interface, the three-dimensional model of the field growth status of the variety at the corresponding time point is displayed in a first-person immersive perspective.

[0010] Preferably, the rendering in step S9 employs a layer-of-details technique, specifically: Calculate the projection error ε of each variety in the screen space. i ; According to the projection error ε i The rendering precision is dynamically switched based on the comparison result with a preset threshold, where: When the projection error ε i When the value is less than the first threshold, it is rendered as a point cloud. When the projection error ε i When the value is between the first and second thresholds, it is rendered as a coarse trajectory. When the projection error ε i When the value is greater than or equal to the second threshold, it is rendered in the form of a fine trajectory.

[0011] Preferably, in step S3, the weighted projection further includes: Quantile normalization was used to convert the phenotypic values ​​into a standard normal distribution. The normalized trait values ​​are nonlinearly transformed using one or more combinations of linear mapping, Sigmoid mapping, or Gaussian kernel mapping. The weighting coefficients of the weighted projection satisfy the normalization constraint.

[0012] Preferably, in step S4, the formula for the composite weighting is: ,in For the normalized time variable, It is a weighted projection value of at least one environmental factor parameter. This is a user-adjustable balance coefficient, with a value range of [0,1].

[0013] Preferably, in step S5, the spline interpolation algorithm uses Catmull-Rom splines for the four adjacent control points of variety i. , , , The interpolation formula for the parameter s∈[0,1] within the segment is: .

[0014] Preferably, the dynamic environment field is represented as a weighted sum of multiple radial basis functions: ,in In three-dimensional space coordinates, Let l be the location of the center of influence of the l-th environmental factor at time t. These are the weighting coefficients. For bandwidth parameters; The environmental field is superimposed in three-dimensional space in the form of a semi-transparent thermal cloud map, and changes dynamically in sync with the user dragging the timeline.

[0015] Preferably, in step S8, the trajectory length of the phenotypic development trajectory curve of variety i Mean curvature and mean torsion They are defined as follows: ; ; ;in , , These represent the phenotypic development trajectory curves. The first, second, and third derivatives with respect to the parameter s; The multivariate feature vector also includes the final performance value of each variety in the target trait and its time variance.

[0016] Preferably, the visual emphasis marker includes a pulsating halo surrounding the aberrant variety, the intensity of which... and pulse frequency They are determined by the following formulas respectively: ; ; in Let be the Mahalanobis distance for the i-th variety. For adaptive threshold, , These are the preset minimum and maximum pulse frequencies.

[0017] Secondly, the present invention provides a phenotypic data visualization and analysis system for potato germplasm resources, used to implement the above method, characterized in that it includes: The data acquisition module is used to acquire multidimensional phenotypic datasets of potato germplasm resources; The spatial mapping module is used to map genetic distance, phenotypic trait values, time, and environmental factors to the three-dimensional axes of three-dimensional space, respectively. The trajectory generation module is used to generate phenotypic development trajectory curves for each variety. The environmental field overlay module is used to overlay dynamic environmental fields in three-dimensional space. Anomaly detection module, used to calculate Mahalanobis distance and generate visual emphasis labels; The rendering and interaction module is used to render 3D space and trajectory curves to the display device and provide a user interaction interface.

[0018] Compared with the prior art, the present invention has the following beneficial effects: By mapping the genetic background (first dimension), multiple trait expression (second dimension), and time process and environmental response (third dimension) of a variety to a unified three-dimensional space, users can perceive information from all four dimensions simultaneously in a single view, overcoming the information fragmentation defects of existing methods.

[0019] By generating phenotypic development trajectory curves, static variety evaluation is transformed into a dynamic display of the developmental process. Users can intuitively observe the trajectory of phenotypic changes from emergence to harvest, and quickly identify the stability and key turning points of phenotypic development.

[0020] A dynamic environmental field is constructed using radial basis functions and superimposed on a three-dimensional space as a semi-transparent thermal cloud map. Users can observe the dynamic changes of the environmental field by dragging the timeline, and simultaneously observe the interaction between the variety trajectory and the environmental field, intuitively understanding the genotype-environment interaction effect.

[0021] Multivariate feature vectors are constructed based on trajectory geometric features, anomaly detection is performed using Mahalanobis distance, and visual emphasis mechanisms such as pulsating halo are used to actively attract users' attention to abnormal varieties, changing the inefficient mode of traditional methods that require manual inspection one by one. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating a method for visualizing and analyzing phenotypic data of potato germplasm resources according to the present invention. Figure 2 This is a schematic diagram of the structure of a potato germplasm resource phenotypic data visualization and analysis system according to the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0024] Example 1: like Figure 1 As shown in the figure, this embodiment provides a method for visualizing and analyzing phenotypic data of potato germplasm resources, specifically including the following steps: Step 1, Data Acquisition and Preprocessing First, a multidimensional phenotypic dataset of potato germplasm resources was obtained. In this embodiment, the dataset contained N=150 potato varieties (lines), and M=12 phenotypic traits (including yield per plant, starch content, reducing sugar content, dry matter content, protein content, vitamin C content, anthocyanin content, late blight resistance score, viral disease resistance score, drought resistance score, single tuber weight, and marketable tuber rate) were observed at T=6 time points (emergence, tuber formation, tuber enlargement, starch accumulation, maturity, and harvest). Simultaneously, L=3 environmental factors (accumulated temperature, precipitation, and sunshine duration) were recorded. Furthermore, a genetic distance matrix between varieties was calculated based on molecular marker data.

[0025] Step 2, 3D spatial coordinate mapping 2.1) Determination of the X-axis (genetic dimension axis) Based on genetic distance matrix We employ Classical Multidimensional Scaling (MDS) to map each variety to the first dimension axis of a three-dimensional space. Specifically, this is achieved by solving the following optimization problem: Take the first dimension coordinate of the optimization result as the X-axis coordinate of each variety. i This mapping places varieties with similar genetic distances close together on the X-axis, and varieties with greater genetic distances far apart on the X-axis.

[0026] 2.2) Determination of the Y-axis (trait dimension axis) For each phenotypic trait value at each time point t Perform quantile normalization: in For normalized, This is the quantile function of the standard normal distribution. After normalization, the trait values ​​follow a standard normal distribution.

[0027] The Y-axis coordinates of each variety at each time point were calculated using weighted linear projection. , specifically: Among them, weight The weighting is set according to the user's analytical objectives. For example, when the user focuses on processing quality, the weighting of starch content and dry matter content can be increased, while the weighting of anthocyanin content can be decreased. The weighting must meet certain criteria. = 1, ≥ 0.

[0028] 2.3) Determination of the Z-axis (environment-time composite axis) First, the time variable is normalized, specifically: Projecting environmental factors, specifically: in Let l be the normalized value of the l-th environmental factor. These are the weighting coefficients. This refers to the number of environmental factors.

[0029] The composite Z-axis coordinate is: .

[0030] In this embodiment, α is set to 0.6, meaning that time accounts for 60% of the weight on the Z-axis, and the environment accounts for 40%. Users can adjust the α value in real time through the interactive interface according to their analysis needs.

[0031] Step 3, Generation of phenotypic development trajectory curves For variety i, its three-dimensional spatial control points at each time point are: .

[0032] The Catmull-Rom spline algorithm is used to interpolate six control points to generate a continuous and smooth trajectory curve. For segments between adjacent control points, the interpolation formula is: .

[0033] The generated trajectory curves follow the following visual coding rules: the curve color represents the variety type (e.g., red represents processing type, green represents fresh type, and purple represents colored type), the curve thickness represents the importance of the variety or data quality, and the curve transparency represents the completeness of the data.

[0034] Step 4, Dynamic environmental field superposition In three-dimensional space Ω = [0,1] 3 A dynamic environmental field is constructed. Taking the temperature factor as an example, the environmental field is expressed as: in The number of spatial sampling points. The position of the l-th sampling point at time t (obtained by interpolation based on actual temperature observation data). σ = 0.15 is the weighting coefficient and the bandwidth parameter.

[0035] The environmental field is overlaid in three-dimensional space as a semi-transparent red thermal cloud map. As the user drags the time axis slider, the environmental field updates in real time according to the t-value, displaying the dynamic changes in temperature distribution over time. By observing the relative position of the variety trajectory and the environmental field, users can intuitively determine the variety's response pattern to temperature—variety whose trajectory deviates towards the high-temperature zone may have heat-resistant characteristics, while variety whose trajectory remains within the suitable temperature zone may be temperature-insensitive.

[0036] Step 5, Anomaly Detection and Visual Emphasis First, calculate the geometric characteristics of the trajectory curve for each variety. For the trajectory curve P(s), s∈[0,1]: trajectory length : ; Mean curvature : ; Mean torsion : ; in , , These represent the phenotypic development trajectory curves. The first, second, and third derivatives with respect to the parameter s.

[0037] Simultaneously extract the final value of the target trait (such as yield) for each variety. and time variance (Reflecting trait stability). The above features are combined into a 5-dimensional feature vector. .

[0038] Calculate the mean μ and covariance matrix Σ of the eigenvectors for all varieties, and then calculate the Mahalanobis distance for each variety. : .

[0039] Set adaptive threshold ,in and represents the mean and standard deviation of the Mahalanobis distance.

[0040] for > Varieties deemed abnormal are identified, and a pulsating halo is generated for visual emphasis. The halo intensity is: ; ; That is, normal varieties have no halo ( The intensity of the halo and the pulse frequency of the anomalous species increase with the degree of deviation, with a maximum pulse frequency of 2 Hz. This dynamic visual emphasis mechanism effectively attracts user attention while encoding information about the degree of anomalousness.

[0041] Step 6, Pareto front extraction When users need to select varieties that excel in both yield and early maturity, the system performs Pareto front extraction.

[0042] Define the objective functions: f1 = yield (maximize), f2 = - number of days to harvest (minimize, i.e., early maturity). Variety i dominates variety j if and only if f1 i ≥ f1 j And f2 i ≥ f2 j And at least one strict inequality holds.

[0043] The set of non-dominant varieties constitutes the Pareto front (P). Projecting the frontier varieties onto the YZ plane (Y-axis representing yield-weighted scores, Z-axis representing harvest time), cubic spline interpolation is used to fit the frontier curve. In three-dimensional space, the trajectory curves of the Pareto frontier varieties are highlighted in gold, and the frontier curve itself is presented as a white, semi-transparent surface. Users can interactively click on points on the frontier curve to view detailed information about the corresponding varieties.

[0044] Step 7, Rendering and Interaction This embodiment uses a WebGL-based 3D rendering engine (Three.js) for visualization rendering. To ensure smooth rendering of 150 trajectory curves (target frame rate ≥ 30fps), Level of Detail (LOD) technology is employed. Calculate the projection error of each variety in the screen space: ,in The spatial distribution scale of the trajectory curve is set to 0.05 in this embodiment. The distance from the variety to the virtual camera, The screen width is 1920 pixels, and the field of view is θ=60°.

[0045] LOD switching logic: when When the value is less than 0.01, it is rendered as a point cloud (only control points are displayed). When 0.01 ≤ When <0.05, render in coarse trajectory form (sampled once every 5 interpolation points); when When the value is ≥ 0.05, it is rendered in the form of a fine trajectory (complete interpolation curve).

[0046] User interaction interfaces include: Mouse drag: Rotate the 3D scene; Mouse wheel: zoom in and out of the view; Timeline slider: When dragged, all tracks move synchronously to the corresponding time point, and the environment field is updated synchronously; Click on the variety trajectory: The right panel displays detailed information about the variety (including original phenotypic data radar chart, field photos, genetic background information, etc.).

[0047] Step 8, Virtual Field Research Mode In this embodiment, 3D reconstructed field images of 15 representative varieties at 6 time points were pre-collected (obtained via UAV oblique photography). The 3D reconstructed models were then used as textures and mapped to the spatial locations of the corresponding varieties at the corresponding time points.

[0048] When a user selects a specific time point on the timeline and clicks on a particular variety, the system provides a "Enter Field" button. Clicking this button switches to a first-person immersive view, allowing the user to "walk" through a virtual field and observe the variety's growth status (plant height, leaf color, tuber enlargement, etc.) from any angle at that time point. This mode supports VR glasses, providing a more immersive experience.

[0049] Example 2: This embodiment is basically the same as Embodiment 1, except that the weighted projection in step S3 uses a nonlinear mapping. For some phenotypic traits, there is a nonlinear relationship between their breeding value and the measured value. For example, the effect of reducing sugar content on the quality of fried food: the difference is not significant when the reducing sugar content is below 0.2%, but the color deteriorates sharply after exceeding 0.3%. Therefore, a Sigmoid mapping is used: , where z is the normalized reducing sugar content. This mapping significantly amplifies small changes in reducing sugar content around 0.25%, which is more consistent with breeding practices.

[0050] Example 3: This embodiment is basically the same as Embodiment 1, except that the α parameter in step S4 adopts an adaptive strategy. For early-maturing breeding objectives, the time dimension is more important, so α is set to 0.8; for adaptive breeding objectives, the environmental dimension is more important, so α is set to 0.3. The system automatically sets the α value according to the analysis template selected by the user, or it can be manually adjusted by the user.

[0051] Example 4: This embodiment is basically the same as embodiment 1, except that the genetic distance mapping uses the t-SNE algorithm instead of MDS.

[0052] t-SNE performs better in handling nonlinear structures and can better preserve the clustering structure of varieties. Specifically, the following optimization objective is adopted: ,in Let be the conditional probability of variety j in the neighborhood of variety i in a high-dimensional space. This represents the conditional probability in the low-dimensional space. After optimization, the first-dimensional coordinate is taken as X. i .

[0053] Example 5: like Figure 2 As shown in the figure, this embodiment provides a visualization and analysis system for phenotypic data of potato germplasm resources, including: Data acquisition module: Used to acquire multidimensional phenotypic datasets of potato germplasm resources, supporting import from various data sources such as Excel, CSV, and databases, and providing data cleaning and missing value imputation functions.

[0054] Spatial mapping module: used to map genetic distance, phenotypic trait values, time and environmental factors to the X, Y and Z axes of three-dimensional space, respectively, and supports a variety of mapping algorithms (MDS, t-SNE, UMAP, linear / nonlinear projection) and weight configuration.

[0055] Trajectory generation module: Used to generate phenotypic development trajectory curves for various varieties, supporting multiple interpolation algorithms such as Catmull-Rom splines, B-splines, and Bézier curves.

[0056] Environmental field overlay module: used to overlay dynamic environmental fields in three-dimensional space, supporting independent or combined display of multiple environmental factors such as temperature, precipitation, and light.

[0057] Anomaly detection module: used to calculate Mahalanobis distance and generate visual emphasis labels, supporting custom feature vector composition and threshold parameters.

[0058] Rendering and Interaction Module: Used to render 3D space and trajectory curves to display devices and provide user interaction interfaces, supporting deployment on web and desktop.

[0059] The system is deployed as software on computer devices, and users access and use it through a browser or client application.

[0060] Industrial applicability The method and system for visualizing and analyzing phenotypic data of potato germplasm resources provided by this invention can be applied to the following scenarios: 1. Digital management of germplasm resource banks: Visualize and integrate the phenotypic data of thousands of potato germplasm accessions stored in the national germplasm resource bank to help resource managers quickly understand the overall structure, diversity and specificity of the resources.

[0061] 2. Breeders' variety selection: Breeders can visually compare the phenotypic performance and developmental dynamics of different varieties in three-dimensional space, and quickly select varieties with excellent comprehensive traits or specific superior traits as parent materials.

[0062] 3. Variety Regional Adaptability Evaluation: By overlaying environmental fields from different ecological zones, the performance differences of varieties in different regions are analyzed, providing decision support for variety layout and promotion.

[0063] 4. Phenotype-genotype association analysis: Linking phenotypic trajectory data with genotype data helps to locate key genes that control the expression of traits at specific developmental stages.

[0064] 5. Teaching and Popular Science Demonstration: Showcase the diversity of potato germplasm resources to students and the public in an intuitive and interactive way to enhance the effectiveness of agricultural science education.

[0065] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for visualizing and analyzing phenotypic data of potato germplasm resources, characterized in that, Includes the following steps: S1: Obtain a multidimensional phenotypic dataset of potato germplasm resources, the dataset including variety identifiers, phenotypic trait values ​​at multiple time points, multiple environmental factor parameters, and genetic distance information between varieties; S2: Based on the genetic distance information, a multidimensional scaling algorithm is used to map each variety to the first dimension axis of three-dimensional space, so that the spatial distance between varieties is positively correlated with the genetic distance. S3: Normalize the phenotypic trait values ​​and map multiple phenotypic traits to the second dimension axis of three-dimensional space through weighted projection to form the trait coordinates of each variety at each time point; S4: Combine the time variable with at least one environmental factor parameter and weight them together, then map them to the third dimension axis of the three-dimensional space to form the environmental-time composite coordinates of each variety at each time point. S5: Using the first, second, and third dimensional axis coordinates as three-dimensional coordinate components, generate three-dimensional spatial control points for each variety at each time point, and use spline interpolation algorithm to connect the control points of the same variety at different time points to generate phenotypic development trajectory curves for each variety. ; S6: A dynamic environmental field is superimposed in the three-dimensional space. The environmental field is represented by radial basis functions and is used to show the spatial and temporal distribution changes of environmental factors. S7: Renders the generated 3D space and phenotypic development trajectory curves to the display device and provides a user interaction interface.

2. The method for visualizing and analyzing phenotypic data of potato germplasm resources according to claim 1, characterized in that, Also includes: S8: Calculate the geometric characteristics of the phenotypic development trajectory curves of each variety, including trajectory length, average curvature and average torsion, and construct a multivariate feature vector based on the geometric characteristics; S9: The deviation of each variety's feature vector from the overall mean is calculated using Mahalanobis distance, and visual emphasis labels are generated for varieties whose deviation exceeds the adaptive threshold.

3. The method for visualizing and analyzing phenotypic data of potato germplasm resources according to claim 1, characterized in that, In step S3, the weighted projection further includes: Quantile normalization was used to convert the phenotypic values ​​into a standard normal distribution. The normalized trait values ​​are nonlinearly transformed using one or more combinations of linear mapping, Sigmoid mapping, or Gaussian kernel mapping. The weighting coefficients of the weighted projection satisfy the normalization constraint.

4. The method for visualizing and analyzing phenotypic data of potato germplasm resources according to claim 1, characterized in that, In step S4, the formula for the composite weighting is: ,in For the normalized time variable, It is a weighted projection value of at least one environmental factor parameter. This is a user-adjustable balance coefficient, with a value range of [0,1].

5. The method for visualizing and analyzing phenotypic data of potato germplasm resources according to claim 1, characterized in that, In step S5, the spline interpolation algorithm uses Catmull-Rom splines for the four adjacent control points of variety i. , , , The interpolation formula for the parameter s∈[0,1] within the segment is: .

6. The method for visualizing and analyzing phenotypic data of potato germplasm resources according to claim 1, characterized in that, The dynamic environment field is represented as a weighted sum of multiple radial basis functions: ,in In three-dimensional space coordinates, Let l be the location of the center of influence of the l-th environmental factor at time t. These are the weighting coefficients. For bandwidth parameters; The environmental field is superimposed in three-dimensional space in the form of a semi-transparent thermal cloud map, and changes dynamically in sync with the user dragging the timeline.

7. The method for visualizing and analyzing phenotypic data of potato germplasm resources according to claim 2, characterized in that, In step S8, the trajectory length of the phenotypic development trajectory curve of variety i Mean curvature and mean torsion They are defined as follows: ; ; ;in , , These represent the phenotypic development trajectory curves, respectively. The first, second, and third derivatives with respect to the parameter s; The multivariate feature vector also includes the final performance value of each variety in the target trait and its time variance.

8. The method for visualizing and analyzing phenotypic data of potato germplasm resources according to claim 1, characterized in that, The method also includes a Pareto front extraction step, specifically: A multi-objective optimization problem is constructed based on at least two target traits, and the set of non-dominant varieties is selected as the Pareto front. The Pareto front is projected onto the plane formed by the second and third dimensional axes, and the front curve is fitted using spline interpolation, highlighting the front curve in three-dimensional space.

9. The method for visualizing and analyzing phenotypic data of potato germplasm resources according to claim 2, characterized in that, The visual emphasis markers include a pulsating halo surrounding the aberrant variety, the intensity of which... and pulse frequency They are determined by the following formulas respectively: ； ； in Let be the Mahalanobis distance for the i-th variety. For adaptive threshold, , These are the preset minimum and maximum pulse frequencies.

10. A system for visualizing and analyzing phenotypic data of potato germplasm resources, used to implement the method described in any one of claims 1-9, characterized in that, include: The data acquisition module is used to acquire multidimensional phenotypic datasets of potato germplasm resources; The spatial mapping module is used to map genetic distance, phenotypic trait values, time, and environmental factors to the three-dimensional axes of three-dimensional space, respectively. The trajectory generation module is used to generate phenotypic development trajectory curves for each variety. The environmental field overlay module is used to overlay dynamic environmental fields in three-dimensional space. Anomaly detection module, used to calculate Mahalanobis distance and generate visual emphasis labels; The rendering and interaction module is used to render 3D space and trajectory curves to the display device and provide a user interaction interface.

Images