Single-cell multi-omics triple prediction method and system based on graph causal learning

Abstract

The invention relates to the technical field of bioinformatics and artificial intelligence, and provides a single-cell multi-omics triple prediction method and system based on graph causal learning, and the method comprises the steps: obtaining multi-omics original data, carrying out the preprocessing, constructing a heterogeneous graph, and generating a single-cell multi-omics triple node according to the heterogeneous graph; performing multi-layer propagation and feature aggregation on the triple nodes through a graph neural network, and performing splicing to obtain a splicing vector; obtaining overall semantic features through an overall semantic module; obtaining directivity characterization through a directivity causal module, and fusing the overall semantic features and the directivity characterization to obtain a triple prediction vector; and inputting the triple prediction vector into a classifier to obtain a classification probability vector. According to the method, the ternary structures of the chromatin open regions of various cell genes are accurately predicted, multi-omics information can be fused, various triple modes can be distinguished, causal reasoning and generalization capabilities are achieved, and the biological interpretation and generalization capabilities of the model are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of bioinformatics and artificial intelligence, and particularly relates to a single-cell multi-omics triple prediction method and system based on graph causal learning. BACKGROUND

[0002] Single-cell transcriptome sequencing can analyze gene expression information at the cell level, and single-cell chromatin accessibility sequencing can depict the transcriptional regulation potential by identifying open chromatin regions. However, existing researches mostly analyze two types of omics data separately, or only perform modality alignment at a low-dimensional representation level, lacking systematic modeling of the triple relationship among cells, genes and chromatin open regions. For example, single-cell transcriptome sequencing can reveal the transcriptional state of cells, but there is less research on chromatin, an upstream regulation field. The openness of chromatin is a prerequisite for transcription factors to bind and activate gene promoters, and therefore, the study of chromatin open regions can provide a potential regulatory mechanism level and reveal the upstream regulation mode of gene expression. Joint analysis of chromatin and transcriptome in the same cell can help understand the transcriptional regulation of cells, especially in the fields of tumors, cell development and cell atlas, and can provide a new perspective for in-depth understanding of disease mechanisms and development of diagnosis and treatment strategies. However, single-cell sequencing data has characteristics such as large scale, high dimension, high sparsity and high heterogeneity, and how to effectively fuse expression data and regulation data at the single-cell level and develop new algorithms to effectively process and utilize these data has become an important research direction in bioinformatics today.

[0003] In single-cell chromatin accessibility sequencing analysis, the regions of open chromatin form peaks on the genome, which usually correspond to the promoters of active genes and transcription factor binding sites, providing information about gene expression and regulation in cells. Traditional regulation relationship construction mostly relies on omics data of whole cells, and cannot capture the cell type or state specificity of gene regulation networks. By exploring the triple relationship at the single-cell level: cell-gene-peak, the active regulation paths in different cells can be more accurately depicted. However, most existing graph learning methods in triple mode classification prediction only perform binary classification on a single triple structure (such as a triangular closed loop), and cannot distinguish multiple biological patterns; there are too many high-frequency single-sided structures, and rare but important regulation patterns are easily ignored. Most methods rely on statistical co-occurrence relationships, which are easily affected by confounding factors and lack explicit modeling of causal relationships in the structure. SUMMARY

[0004] The present application aims to at least solve one of the technical problems existing in the related art. To this end, the present application provides a single-cell multi-omics triple prediction method and system based on graph causal learning, which realizes accurate prediction of various cell-gene-chromatin open region triple structures by constructing a heterogeneous graph, fusing overall semantic features and directional causal features, combining a double contrast learning mechanism, and introducing a causal counterfactual enhancement strategy. It can fuse multi-omics information, distinguish various triple patterns, and has causal reasoning and generalization ability, improving the biological interpretability and generalization ability of the model.

[0005] The present application provides a single-cell multi-omics triple prediction method based on graph causal learning, comprising: S1: Obtain multi-omics raw data containing single-cell transcriptome sequencing and single-cell chromatin accessibility sequencing and perform preprocessing to obtain preprocessed cell nodes, peak nodes and gene nodes; S2: Construct a heterogeneous graph according to the preprocessed cell nodes, peak nodes and gene nodes, and generate single-cell multi-omics triple nodes according to the heterogeneous graph; S3: Perform multi-layer propagation and feature aggregation on the single-cell multi-omics triple nodes on the heterogeneous graph through a graph neural network to obtain node embedding features, and splice the node embedding features to obtain a spliced vector; S4: Construct a causal prediction model and train the causal prediction model through double contrast learning and causal counterfactuals, wherein the causal prediction model comprises an overall semantic module and a directional causal module; S5: Fuse the global semantic features of the spliced vector through the overall semantic module to obtain overall semantic features, extract directional features of the single-cell multi-omics triple through the directional causal module to obtain directional representations, and fuse the overall semantic features and the directional representations to obtain a triple prediction vector; S6: Input the triple prediction vector into a classifier to obtain a single-cell multi-omics triple classification probability vector.

[0006] Further, the preprocessing comprises: Peak screening is performed on the single-cell chromatin accessibility sequencing, and feature dimensionality reduction is performed on the retained peak features to obtain a peak feature node set; Gene screening is performed on the single-cell transcriptome sequencing, and peaks related to the genes are screened to obtain a gene node set.

[0007] Further, the step S2 comprises: S21: According to the preprocessed cell nodes, peak nodes and gene nodes, an edge relationship matrix is constructed according to the biological regulation relationship; S22: adopt the embedding learning method based on PyTorch-BigGraph to learn the vectorized representation of the cell nodes, the peak nodes and the gene nodes, and obtain node embeddings; S23: construct a heterogeneous graph according to the node embeddings and the edge relation matrix; S24: construct seven types of three-node structures according to the heterogeneous graph.

[0008] Further, the seven types of three-node structures include a triangle structure, a double-edge structure centered on a cell, a double-edge structure centered on a peak, a double-edge structure centered on a gene, a single-edge structure of a gene and a cell, a single-edge structure of a gene and a peak, and a single-edge structure of a cell and a peak.

[0009] Further, the S3 step includes: S31: extract the initial features of the cell, the initial features of the peak and the initial features of the gene of the triple-node; S32: propagate the initial features of the cell through a multi-layer graph convolutional neural network to obtain cell node embedding features, propagate the initial features of the peak through a multi-layer graph convolutional neural network to obtain peak node embedding features, and propagate the initial features of the gene through a multi-layer graph convolutional neural network to obtain gene node embedding features; S33: vector splicing the cell node embedding features, the peak node embedding features and the gene node embedding features to obtain a splicing vector.

[0010] Further, the double contrast learning includes a node consistency constraint and a structure mode level semantic aggregation. The node consistency constraint includes: S411: randomly mask part of the edges of the heterogeneous graph to obtain a perturbed graph; S412: extract the node embeddings of the heterogeneous graph through a graph convolutional neural network with shared parameters to obtain heterogeneous graph node embeddings, and extract the node embeddings of the perturbed graph through a graph convolutional neural network with shared parameters to obtain perturbed graph node embeddings; S413: calculate the cosine similarity of the heterogeneous graph node embeddings and the perturbed graph node embeddings to obtain a node consistency contrast loss; The structure mode level semantic aggregation includes: S421: calculate the mean of the node embeddings of the triple samples of the heterogeneous graph to obtain a heterogeneous graph prototype vector, and calculate the mean of the node embeddings of the triple samples of the perturbed graph to obtain a perturbed graph prototype vector; S422: calculate a prototype consistency loss according to the heterogeneous graph prototype vector and the perturbed graph prototype vector; S423: obtain a multi-class probability distribution by calculating the similarity of the triple and the heterogeneous graph prototype vector and the similarity of the triple and the perturbed graph prototype vector; S424: calculate cross-entropy loss according to the multi-class probability distribution.

[0011] Further, the causal counterfactual includes: S431: replace the target node of part of the edges of the edge set to generate a counterfactual triple; S432: measure the embedding distribution difference between the counterfactual triple and the triple by KL divergence to generate a counterfactual loss.

[0012] Further, the causal prediction model is jointly optimized by the node consistency comparison loss, the prototype consistency loss, the cross-entropy loss and the counterfactual loss.

[0013] Further, obtaining the directionality representation includes: S521: construct a cell peak direction difference vector, a peak gene direction difference vector and a gene cell direction difference vector; S522: obtain the directionality representation by nonlinear mapping and splicing the cell peak direction difference vector, the peak gene direction difference vector and the gene cell direction difference vector.

[0014] The application also provides a single cell multi-omics triple prediction system based on graph causal learning, which is used to execute the single cell multi-omics triple prediction method based on graph causal learning. The acquisition module acquires multi-omics raw data containing single cell transcriptome sequencing and single cell chromatin accessibility sequencing and performs preprocessing to obtain preprocessed cell nodes, peak nodes and gene nodes; The node generation module constructs a heterogeneous graph according to the preprocessed cell nodes, peak nodes and gene nodes, and generates single cell multi-omics triple nodes according to the heterogeneous graph; The feature splicing module obtains node embedding features by multi-layer propagation and feature aggregation of the single cell multi-omics triple nodes on the heterogeneous graph through the graph neural network, splices the node embedding features to obtain a spliced vector; The model construction module constructs a causal prediction model and trains the causal prediction model through double contrast learning and causal counterfactual, and the causal prediction model includes an overall semantic module and a directional causal module; The prediction module fuses the global semantic features of the spliced vector through the overall semantic module to obtain overall semantic features, extracts the directional features of the single cell multi-omics triple through the directional causal module to obtain a directional representation, and fuses the overall semantic features and the directional representation to obtain a triple prediction vector; The classification module inputs the triple prediction vector into a classifier to obtain a single cell multi-omics triple classification probability vector.

[0015] The one or more technical solutions in the embodiments of the present application have at least one of the following technical effects: The present application realizes accurate prediction of various cell-gene-chromatin open region ternary structures by constructing a heterogeneous graph, fusing global semantics and directional causal features, combining a double contrast learning mechanism, and introducing a causal counterfactual enhancement strategy, can distinguish seven cell-peak-gene ternary structure modes at the same time, realize more fine biological structure prediction, causal reasoning and counterfactual enhancement, significantly alleviate the prediction deviation caused by class imbalance and structural redundancy, improve the biological interpretability and generalization ability of the model, and the training embedding result can be used for cell regulation mechanism analysis and disease research.

[0016] Additional aspects and advantages of the application will be set forth in part in the description which follows, and in part will become apparent to those skilled in the art upon examination of the following description and drawings or can be learned by practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0018] Figure 1 is a flowchart of a single-cell multi-omics ternary group prediction method based on graph causal learning provided by the present application.

[0019] Figure 2 is a technical roadmap of a single-cell multi-omics ternary group prediction model based on graph causal learning provided by the present application.

[0020] Figure 3 is a t-distributed random neighborhood embedding visualization diagram of the ternary relationship embedding of seven categories in the embodiment of the present application.

[0021] Figure 4 is a visualization diagram of the cell embedding obtained by model training using dimension reduction technology in the embodiment of the present application.

[0022] Figure 5 is a visualization diagram of the joint representation of cell embedding and gene embedding using dimension reduction technology in the embodiment of the present application.

[0023] Figure 6 is a visualization diagram of the embedding data of cells, genes and chromatin accessibility regions mapped to a unified low-dimensional space by UMAP in the embodiment of the present application.

[0024] Figure 7It is a structure schematic diagram of a single cell multi-omics triple prediction system based on graph causal learning provided by the application.

[0025] Reference signs: 101, acquisition module; 102, node generation module; 103, feature splicing module; 104, model construction module; 105, prediction module; 106, classification module. DETAILED DESCRIPTION

[0026] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be clearly and completely described below. Obviously, the described embodiments are some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application. The following embodiments are used to illustrate the present application, but cannot be used to limit the scope of the present application.

[0027] In the description of the embodiments of the present application, it should be noted that the terms "first", "second", "third" are only used for description purposes, and cannot be understood as indicating or implying relative importance. In the description of the present application, the description of the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In the present application, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, those skilled in the art can combine and combine the different embodiments or examples described in the present application and the features of the different embodiments or examples without contradiction.

[0028] The following will be described in conjunction with Figures 1 to 7 A single cell multi-omics triple prediction method and system based on graph causal learning of the present application is described.

[0029] As Figure 1 shown, a single cell multi-omics triple prediction method based on graph causal learning, comprising: S1: Obtain multi-omics raw data containing single cell transcriptome sequencing (scRNA-seq) and single cell chromatin accessibility sequencing (scATAC-seq) and pre-process to obtain pre-processed cell nodes, peak nodes and gene nodes; Perform peak screening on single cell chromatin accessibility sequencing and feature dimension reduction on the retained peak features to obtain a peak feature node set.

[0030] The frequency of each peak in all cells was counted, and low-frequency peaks that appeared in fewer than 3 cells were filtered out to remove low-quality signals. Principal component analysis (PCA) is used to reduce the dimensionality of the retained peak features, thereby reducing the feature dimension and lowering the complexity of the model input while preserving as much information as possible.

[0031] Specifically, the chromatin accessibility data matrix of single cells is centered, then the covariance matrix is ​​calculated, and the covariance matrix is ​​decomposed into eigenvalues. The eigenvectors corresponding to the first few eigenvalues ​​are selected as principal components.

[0032] Gene screening was performed on single-cell transcriptome sequencing, and peaks associated with genes were identified to obtain a set of hypervariable gene nodes.

[0033] By using DeepMAPS (a deep learning-based multi-omics analysis platform for single-cell data), spatial relationships, chromatin accessibility, and gene expression characteristics are integrated to identify and quantify the association between peaks and genes, and only peaks that are significantly associated with genes are retained.

[0034] Specifically, two key matrices are given: the gene expression matrix and the gene expression matrix. and chromatin accessibility matrix ; in, For cells genes Gene expression, open chromatin regions In cells Medium chromatin accessibility For the number of genes, Cell count, This represents the number of open regions of chromatin.

[0035] For calculating chromatin open regions On genes Regulation potential weight The distance between the two in the genome plays a decisive role. If Center and gene Distance of transcription start site More than 150kb (kilobase, kilobase pairs), or Located within other nearby genes, the potential regulatory weight is defined as: Long-distance interactions are usually negligible, and peaks located near other genes are more likely to regulate neighboring genes.

[0036] when When located in the exon region of a gene: in, is the exon length.

[0037] Based on observations in batch chromatin accessibility sequencing data, highly expressed genes have open chromatin regions in their exons. Normalization by exon length helps to account for the probability of background readouts.

[0038] In other cases: in, The half-life is the distance value, reflecting the principle that the regulatory effect weakens exponentially with increasing distance.

[0039] In obtaining Afterwards, through Computational cells middle On genes Regulation potential score Finally, the regulatory potential scores of peaks regulating the same gene are summed, and the chromatin accessibility matrix is ​​used to... The results are converted into a potential gene regulatory matrix, yielding a certain number of representative peaks.

[0040] A screening strategy based on highly variable genes (HVGs) was employed to identify gene-related peaks, resulting in a set of highly variable gene nodes. This strategy is based on the core assumption that genes with higher expression variability are more discriminative in distinguishing different cell types or states. The variability score was calculated by statistically analyzing the relationship between the mean expression of each gene across all cells and its variability. Based on this, all genes were sorted according to their variability scores, and the genes with the highest variability were selected as the target for analysis.

[0041] S2: Construct a heterogeneous graph based on the preprocessed cell nodes, peak nodes, and gene nodes, and generate single-cell multi-omics triplet nodes based on the heterogeneous graph; S21: Based on the preprocessed cell nodes, peak nodes, and gene nodes, construct an edge relationship matrix according to biological regulatory relationships; including three types of relationships: cell–peak, cell–gene, and peak–gene.

[0042] S22: An embedding learning method based on PyTorch-BigGraph is used to learn the vectorized representation of cell nodes, peak nodes, and gene nodes to obtain node embeddings; The constructed graph data is converted into a triplet text format conforming to PBG. Before training begins, the system automatically analyzes the basic statistical information of the graph structure and empirically estimates the weight decay parameters based on this information to avoid embedding overfitting. Specifically, a relatively large weight decay factor is used for graphs with fewer edges, while a smaller regularization coefficient is used for graphs with dense edges to retain more structural details. Subsequently, the model iteratively optimizes the embedding representation of each type of node using triplet information and a predefined multi-class cross-entropy loss. In each training step, PBG adjusts the embedding of the corresponding node according to the triplet relationships, so that the edge connections are approximately preserved in the vector space. After training, vector representations of three types of nodes—cell, peak, and gene—are obtained, namely, cell node embeddings, peak node embeddings, and gene node embeddings.

[0043] S23: Construct a heterogeneous graph based on node embeddings and edge relationship matrices; Define heterogeneous graph , , Let be the set of edges. For a set of nodes, It contains three types of nodes: cell, peak, and gene; It contains three types of edges: cell–peak, cell–gene, and peak–gene; S24: Construct seven types of three-node structures based on heterogeneous graphs; like Figure 2 As shown in Figure (a), the seven types of three-node structures include the Triangle structure, the Cell-Center structure (a two-sided structure centered on the cell), the Peak-Center structure (a two-sided structure centered on the peak), the Gene-Center structure (a two-sided structure centered on the gene), the Gene-Cell structure (a one-sided structure centered on the gene and cell), the Gene-Peak structure (a one-sided structure centered on the gene and peak), and the Cell-Peak structure (a one-sided structure centered on the cell and peak).

[0044] Triangle Structure: Triples that simultaneously meet the following conditions are selected: the peak is associated with a gene; the peak is connected to a cell; and the cell and gene have a regulatory relationship. Cell-gene-peak triplets that meet these conditions are retained as triangle structure samples, representing the fully connected biological mechanisms between the three types of nodes.

[0045] Cell-Center: For each cell node, its known associated peak nodes and gene nodes are extracted; and all peak-gene pairs are examined. If no known association exists between peak-gene pairs, the cell-gene-peak triplet is retained. The Cell-Center reflects situations where cells may indirectly influence certain genes through mechanisms that have not yet been identified.

[0046] Peak-Center: Obtain the cell and gene sets associated with each peak. For any cell-gene pair, if there is no direct connection, a ternary structure is formed with the peak as the center. The Peak-Center reflects a possible missing link in the peak regulatory pathway, which helps guide the model to identify unknown cell-gene regulatory pathways.

[0047] Gene-Center: Identify the set of peaks and cell nodes associated with each gene. For each peak-cell pair, if there is no known association between them, construct a ternary relationship centered on the gene.

[0048] The remaining edges in the heterogeneous graph are single-sided structures, and there are three types: gene-cell, gene-peak, and cell-peak.

[0049] S3: Multi-layer propagation and feature aggregation are performed on single-cell multi-omics triple nodes on heterogeneous graphs through graph neural networks to obtain node embedding features, and the node embedding features are concatenated to obtain concatenated vectors. S31: Extract the initial cellular features, initial peak features, and initial gene features of the triplet node; S32: Cell node embedding features are obtained by multi-layer propagation of initial cell features through graph convolutional neural networks; peak node embedding features are obtained by multi-layer propagation of initial peak features through graph convolutional neural networks; gene node embedding features are obtained by multi-layer propagation of initial gene features through graph convolutional neural networks. The calculation expression is: in, For the first Layer node features It is a non-linear activation function. To normalize Laplace, For the first Layer node features As weight.

[0050] Cell node features are encoded using a multi-layer graph convolutional neural network to obtain cell node embedding features. Peak node embedding features are obtained by encoding peak node features through a multi-layer graph convolutional neural network. Gene node embedding features are obtained by encoding gene node features through a multi-layer graph convolutional neural network. .

[0051] Each node's embedding features contain contextual information about itself and its multi-level neighbors.

[0052] S33: Concatenate the cell node embedding features, peak node embedding features, and gene node embedding features into vectors to obtain the concatenated vector. .

[0053] S4: Construct a causal prediction model and train it using dual contrastive learning and causal counterfactual methods. The causal prediction model includes a holistic semantic module and a directional causal module. The causal prediction model is as follows: Figure 2 As shown in Figure (b), Figure 2 In the text, CAT stands for splicing operation.

[0054] like Figure 2 As shown in Figure (c), dual contrastive learning includes node consistency constraints and structural pattern-level semantic aggregation; Node consistency constraints include: S411: Randomly mask some edges of a heterogeneous graph to obtain a perturbation graph; S412: Extract node embeddings of the heterogeneous graph using a graph convolutional neural network with shared parameters to obtain the heterogeneous graph node embeddings. Extract node embeddings of the perturbed graph using a graph convolutional neural network with shared parameters to obtain the perturbed graph node embeddings. The calculation expression is as follows: in, Embedding matrix for heterogeneous graph nodes For graph convolutional neural networks, It is a heterogeneous graph. The node feature matrix, Embedding matrix for nodes in perturbation graph This is a perturbation diagram; Heterogeneous graphs and perturbation graphs share GCN parameters, and node-level consistency constraints are applied to the node characteristics of heterogeneous graphs and perturbation graphs. S413: Calculate the cosine similarity between the node embeddings of the heterogeneous graph and the node embeddings of the perturbed graph to obtain the node consistency contrast loss; The cosine similarity between heterogeneous graph node embeddings and perturbed graph node embeddings is expressed as follows: in, For cosine similarity, For heterogeneous graph nodes Embedded vector, For nodes in the perturbation graph Embedded vector, Let L2 norm be the vector. It is the vector inner product.

[0055] The expression for calculating the node-level contrastive loss is: in, For nodes Node-level contrast loss, For temperature coefficient, For nodes The set of negative sample pairs, This refers to all nodes that have been traversed.

[0056] The consistency loss of each node in the triplet is summarized under both views. The expression for calculating the node consistency comparison loss is as follows: in, The loss is due to node consistency comparison. It is a set of nodes.

[0057] The structural pattern-level semantic aggregation includes: S421: Calculate the mean node embedding of triple samples in the heterogeneous graph to obtain the prototype vector of the heterogeneous graph; calculate the mean node embedding of triple samples in the perturbation graph to obtain the prototype vector of the perturbation graph. S422: Calculate the prototype consistency loss based on the heterogeneous graph prototype vector and the perturbation graph prototype vector. The calculation expression is as follows: in, For prototype consistency loss, For the heterogeneous graph One prototype vector, For the perturbation diagram One prototype vector, The number of prototypes; S423: By calculating the similarity between triples and the prototype vectors of heterogeneous graphs, as well as the similarity between triples and the prototype vectors of perturbation graphs, the multi-class probability distribution is obtained. The calculation expression is as follows: in, for Time of the first The probability of a prototype, For the first One prototype vector, For the first One prototype vector, for The triple vector at time t; S424: Calculate the cross-entropy loss based on the multi-class probability distribution. The calculation expression is as follows: in, For cross-entropy loss, This is an indicator function for the actual label.

[0058] The classification accuracy is optimized by using cross-entropy loss, achieving intra-class aggregation and inter-class separation.

[0059] like Figure 2 As shown in diagram (d), causal counterfactuals include: S431: Replace the target nodes of some edges in the edge set to generate counterfactual triples; Given a set of edges , , For the first edge, For the second edge, For the first A subset of edges is randomly selected for intervention. By modifying the target nodes of these edges, different causal relationships are simulated. The resulting set of edges after intervention is as follows: in, For the set of edges after intervention, The first edge after intervention, The second edge after intervention. For the first day after intervention Edge; exist and The counterfactual node representation and counterfactual triple representation after intervention are obtained through a graph convolutional neural network with shared parameters.

[0060] S432: Use KL divergence to measure the difference in embedding distributions between counterfactual triples and generate counterfactual loss.

[0061] Assumption It is a node The representation after intervention. It is a node The original representation, defined by KL divergence, amplifies the sensitivity to the causal effects of key edges, resulting in the following counterfactual loss: in, For counterfactual losses, Let KL divergence function be used. For the softmax function, the softmax function is a function that maps a real number vector to a probability distribution; Counterfactual loss is obtained by causally comparing the original features and counterfactual features.

[0062] The triplet samples are proportionally divided into training, validation, and test sets; triplet samples and their negative sample sets are batch-sampled to construct two-view and counterfactual graph batches. A learning rate is used. End-to-end training was performed with batch size of 500 and hidden dimension of 128. The causal prediction model was jointly optimized using node consistency contrast loss, prototype consistency loss, cross-entropy loss, and counterfactual loss. The calculation expression is as follows: in, For the total loss, As the first hyperparameter, This is the second hyperparameter. This is the counterfactual loss hyperparameter.

[0063] Node-level consistency loss suppresses structural redundancy and overfitting; structural pattern-level semantic aggregation alleviates class imbalance and maintains consistency between the two views; counterfactual enhancement stabilizes training and improves generalization and interpretability, ultimately resulting in robust triplet pattern recognition performance.

[0064] S5: The global semantic features of the spliced ​​vector are fused through the overall semantic module to obtain the overall semantic features; the directional features of the single-cell multi-omics triples are extracted through the directional causal module to obtain the directional representation; the overall semantic features and the directional representation are fused to obtain the triple prediction vector. S51: By fusing the global semantic features of the concatenated vectors through the overall semantic module, the overall semantic features are obtained, and the calculation expression is as follows: in, This is a vector concatenation operation. As for the overall semantic features, This is the first multilayer perceptron.

[0065] The overall semantic module is a non-linear mapping network; by fusing the full-graph contextual semantics between nodes through the overall semantic module, it emphasizes the joint expression between cells, peaks and genes.

[0066] S52: Extract directional features of single-cell multi-omics triples through the directional causal module to obtain directional characterization; S521: Construct the cell peak direction difference vector, peak gene direction difference vector, and cell gene direction difference vector. The calculation expression is as follows: in, This is the difference vector of the cell peak direction. The peak gene direction difference vector, This is the directional difference vector of the cell genes.

[0067] S522: Directional characterization is obtained by nonlinearly mapping and concatenating the cell peak direction difference vector, peak gene direction difference vector, and gene-cell direction difference vector. The calculation expression is as follows: in, As a directional representation, For the second multilayer perceptron, This is the feature mapping function.

[0068] S53: By fusing overall semantic features and directional representations, a triplet prediction vector is obtained, calculated as follows: in, This is the prediction vector for the triplet.

[0069] S6: Input the triplet prediction vector into the classifier to obtain the single-cell multi-omics triplet classification probability vector.

[0070] The triplet prediction vector is fed into a softmax classifier, which outputs a probability vector that belongs to one of the seven types of three-node structures. Multi-task classification and single-cell multimodal analysis are then performed to obtain the prediction results of the seven types of three-node structures.

[0071] like Figure 3 As shown, the t-distributed random neighborhood embedding (t-SNE) of the seven-class ternary relation embeddings obtained from training is visualized. The seven ternary patterns are labeled with different colors to show the distribution of data points with different prediction labels in two-dimensional space. Figure 3 Data points of different triplet patterns exhibit intra-class clustering and inter-class separation in low-dimensional space, demonstrating that this invention can overcome the limitations of traditional methods and achieve accurate differentiation of seven biological triplet patterns.

[0072] like Figure 4 As shown, dimensionality reduction techniques are used to visualize the cell embeddings obtained from model training and to analyze the distribution patterns of different cell types. Each cell type is labeled with a different color, and different cell types exhibit different clustering patterns in the low-dimensional space, reflecting their similarities and differences. Figure 4The five cell types are: Hair Shaft-cuticle.cortex (hair shaft-keratin / cortical cells), IRS (inner root sheath cells), Medula (hair medulla cells), TAC-1 (Transit-Amplifying Cells, transitional proliferating cells 1), and TAC-2 (transital proliferating cells 2). Regarding the distribution of cell types, TAC-1 and TAC-2 cell types are more dispersed, reflecting the high heterogeneity of these cells in the data space and their complex or diverse biological characteristics.

[0073] like Figure 5 As shown, a joint representation of cell and gene embeddings is visualized using dimensionality reduction techniques. Figure 5 The distribution of genes overlaps with that of cell types, suggesting a potential intrinsic link between genes and cell types. To analyze the correlation between specific genes and cell types, relevant marker genes are labeled in the visualization, revealing the expression patterns of key genes in different cell types. Figure 5 The distribution of Igf1 (Insulin-like growth factor 1), Tsacc (TSSK6-activating co-chaperone), Tmpo (Thymopoietin / LAP2), and Syt7 (Synaptotagmin-7) are shown in the figure.

[0074] like Figure 6 As shown, by mapping the embedding data of cells, genes, and chromatin accessibility regions to a unified low-dimensional space and using UMAP (Uniform Manifold Approximation and Projection) for visualization, the two types of omics data, "single-cell transcriptome (genes)" and "single-cell chromatin accessibility (peaks)," are unified with the features of the "cell" node into the same embedding space, achieving cross-modal fusion of the three types of nodes; the spatial distribution reveals the distribution and potential relationships of cell types, genes, and peaks in the low-dimensional space.

[0075] like Figure 7 As shown, a single-cell multi-omics triplet prediction system based on graph causal learning is used to execute a single-cell multi-omics triplet prediction method based on graph causal learning, including: The acquisition module 101 acquires multi-omics raw data including single-cell transcriptome sequencing and single-cell chromatin accessibility sequencing and preprocesses it to obtain preprocessed cell nodes, peak nodes and gene nodes. The node generation module 102 constructs a heterogeneous graph based on the preprocessed cell nodes, peak nodes, and gene nodes, and generates single-cell multi-omics triplet nodes based on the heterogeneous graph. The feature splicing module 103 performs multi-layer propagation and feature aggregation on the single-cell multi-omics triplet nodes on the heterogeneous graph through graph neural network to obtain node embedded features, and splices the node embedded features to obtain splicing vector. The model building module 104 constructs a causal prediction model and trains the causal prediction model through dual contrastive learning and causal counterfactual methods. The causal prediction model includes an overall semantic module and a directional causal module. The prediction module 105 obtains the overall semantic features by fusing the global semantic features of the spliced ​​vector through the overall semantic module; it obtains the directional features of the single-cell multi-omics triplet by extracting the directional features of the single-cell multi-omics triplet through the directional causal module; and it obtains the directional representation by fusing the overall semantic features and the directional representation to obtain the triplet prediction vector. The classification module 106 inputs the triplet prediction vector into the classifier to obtain the single-cell multi-omics triplet classification probability vector.

[0076] Through the collaborative work of the above modules, this invention constructs heterogeneous graphs, integrates global semantics and directional causal features, combines a dual contrastive learning mechanism, and introduces a causal counterfactual enhancement strategy to achieve accurate prediction of multiple cell-gene-chromatin open region ternary structures. It can simultaneously distinguish seven cell-peak-gene ternary structure patterns, achieving more refined biological structure prediction. Causal reasoning and counterfactual enhancement significantly alleviate prediction bias caused by class imbalance and structural redundancy, improve the model's biological interpretability and generalization ability, and the training embedding results can be used for cell regulatory mechanism analysis and disease research. It can be applied to single-cell multi-omics analysis scenarios such as cell type identification, disease mechanism research, and drug target screening.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting single-cell multi-omics triplets based on graph causal learning, characterized in that, The method comprises the following steps: S1: obtaining multi-omics raw data containing single-cell transcriptome sequencing and single-cell chromatin accessibility sequencing and preprocessing, obtaining preprocessed cell nodes, peak nodes and gene nodes; S2: constructing a heterogeneous graph according to the preprocessed cell nodes, peak nodes and gene nodes, and generating single-cell multi-omics triple nodes according to the heterogeneous graph; S3: performing multi-layer propagation and feature aggregation on the single-cell multi-omics triple nodes on the heterogeneous graph through a graph neural network to obtain node embedding features, and splicing the node embedding features to obtain a spliced vector; S4: constructing a causal prediction model and training the causal prediction model through double contrast learning and causal counterfactual, the causal prediction model comprising an overall semantic module and a directional causal module; S5: fusing the global semantic features of the spliced vector through the overall semantic module to obtain overall semantic features; extracting directional features of the single-cell multi-omics triplets through the directional causal module to obtain directional representations, and fusing the overall semantic features and the directional representations to obtain a triple prediction vector; S6: inputting the triple prediction vector into a classifier to obtain a single-cell multi-omics triple classification probability vector.

2. The single-cell multi-omics triplets prediction method based on graph causal learning according to claim 1, characterized in that, The preprocessing comprises: performing peak screening on single-cell chromatin accessibility sequencing and performing feature dimension reduction on the retained peak features to obtain a peak feature node set; performing gene screening on single-cell transcriptome sequencing and screening peaks related to the genes to obtain a gene node set.

3. The method of claim 1, wherein, The step S2 comprises: S21: constructing an edge relationship matrix according to the preprocessed cell nodes, peak nodes and gene nodes according to biological regulation relationships; S22: using a PyTorch-BigGraph-based embedding learning method to vectorize and represent the cell nodes, peak nodes and gene nodes to obtain node embeddings; S23: constructing a heterogeneous graph according to the node embeddings and the edge relationship matrix; S24: constructing seven types of triple node structures according to the heterogeneous graph.

4. The single-cell multi-omics triplets prediction method based on graph causal learning according to claim 3, characterized in that, The seven types of triple node structures comprise a triangular structure, a double-edge structure centered on a cell, a double-edge structure centered on a peak, a double-edge structure centered on a gene, a single-edge structure of a gene and a cell, a single-edge structure of a gene and a peak, and a single-edge structure of a cell and a peak.

5. The single-cell multi-omics triplets prediction method based on graph causal learning according to claim 1, characterized in that, The step S3 comprises: S31: extracting cell initial features, peak initial features and gene initial features of the triple nodes; S32: performing multi-layer propagation on the cell initial features through a graph convolutional neural network to obtain cell node embedding features, performing multi-layer propagation on the peak initial features through a graph convolutional neural network to obtain peak node embedding features, and performing multi-layer propagation on the gene initial features through a graph convolutional neural network to obtain gene node embedding features; S33: vector splicing the cell node embedding features, peak node embedding features and gene node embedding features to obtain a spliced vector.

6. The single-cell multi-omics triplets prediction method based on graph causal learning according to claim 1, characterized in that, The double contrast learning comprises node consistency constraints and structure mode level semantic aggregation; The node consistency constraints comprise: S411: randomly masking part of the edges of the heterogeneous graph to obtain a perturbed graph; S412: extract node embeddings of the heterogeneous graph and the perturbed graph by the graph convolutional neural network with shared parameters, obtain the heterogeneous graph node embeddings and the perturbed graph node embeddings; S413: calculate the cosine similarity of the heterogeneous graph node embeddings and the perturbed graph node embeddings, obtain the node consistency contrastive loss; The structure mode level semantic aggregation comprises: S421: calculate the node embedding mean of the triple sample of the heterogeneous graph, obtain the heterogeneous graph prototype vector, calculate the node embedding mean of the triple sample of the perturbed graph, obtain the perturbed graph prototype vector; S422: calculate the prototype consistency loss according to the heterogeneous graph prototype vector and the perturbed graph prototype vector; S423: obtain the multi-class probability distribution by calculating the similarity of the triple and the heterogeneous graph prototype vector and the similarity of the triple and the perturbed graph prototype vector; S424: calculate the cross-entropy loss according to the multi-class probability distribution.

7. The single-cell multi-omics triplets prediction method based on graph causal learning according to claim 6, characterized in that, The causal counterfactual comprises: S431: replace the target node of part of the edges in the edge set to generate a counterfactual triple; S432: measure the embedding distribution difference between the counterfactual triple and the triple by KL divergence to generate a counterfactual loss.

8. The single-cell multi-omics triplets prediction method based on graph causal learning according to claim 7, characterized in that, The causal prediction model is jointly optimized by the node consistency contrastive loss, the prototype consistency loss, the cross-entropy loss and the counterfactual loss.

9. The single-cell multi-omics triplets prediction method based on graph causal learning of claim 1, wherein, The directionality representation comprises: S521: construct a cell peak direction difference vector, a peak gene direction difference vector and a gene cell direction difference vector; S522: perform nonlinear mapping and splicing on the cell peak direction difference vector, the peak gene direction difference vector and the gene cell direction difference vector to obtain the directionality representation. 10.A system for single-cell multi-omics triplets prediction based on graph causal learning, comprising: A single-cell multi-omics triple prediction method based on graph causal learning is executed, comprising: An acquisition module acquires multi-omics raw data containing single-cell transcriptome sequencing and single-cell chromatin accessibility sequencing and performs preprocessing to obtain preprocessed cell nodes, peak nodes and gene nodes; A node generation module constructs a heterogeneous graph according to the preprocessed cell nodes, peak nodes and gene nodes, and generates single-cell multi-omics triple nodes according to the heterogeneous graph; A feature splicing module obtains node embedding features by performing multi-layer propagation and feature aggregation on the single-cell multi-omics triple nodes on the heterogeneous graph by a graph neural network, splices the node embedding features to obtain a spliced vector; A model construction module constructs a causal prediction model and trains the causal prediction model by double contrastive learning and causal counterfactuals, wherein the causal prediction model comprises an overall semantic module and a directional causal module; A prediction module fuses global semantic features of the spliced vector by the overall semantic module to obtain overall semantic features, extracts directional features of the single-cell multi-omics triple by the directional causal module to obtain a directionality representation, and fuses the overall semantic features and the directionality representation to obtain a triple prediction vector; A classification module inputs the triple prediction vector into a classifier to obtain a single-cell multi-omics triple classification probability vector.

Images