Efficient multi-omics element cell inference method based on single cell quantification

Abstract

The invention relates to the technical field of element cell inference, and discloses an efficient multi-omics element cell inference method based on single cell quantization, which comprises the following steps: acquiring and standardizing cell sequencing data to obtain standard cell sequencing data; initializing a quantization codebook, a neural network encoder, a first neural network decoder and a second neural network decoder, inputting standard cell sequencing data for training, and performing feature extraction, cell quantization and quantization reconstruction on the standard cell sequencing data, so as to obtain a quantized cell sequence; obtaining a trained neural network encoder, a first neural network decoder, a second neural network decoder and a quantized codebook; inputting the original single cell sequencing data or the original multi-omics cell sequencing data into a trained neural network encoder for feature extraction, and performing cell quantization based on a trained quantization codebook to obtain an element cell inference result; the method not only realizes efficient inference on a large-scale single cell data set, but also effectively reduces the calculation cost on the premise of ensuring the precision.

Description

Technical Field

[0001] The present invention relates to the technical field of meta-cell inference, and in particular to an efficient multi-omics meta-cell inference method based on single-cell quantification. Background Art

[0002] In recent years, single-cell sequencing technology has made rapid progress, and the number of cells that can be captured has continued to increase. It has shown great advantages in revealing cell heterogeneity and reconstructing cell developmental trajectories. However, with the exponential growth of data size, the analysis of single-cell sequencing data also faces severe computational challenges. For example, a typical single-cell data analysis process (including data integration, clustering, visualization, and differential expression analysis) takes about 16 hours to process 500,000 cells. When the number of cells increases to 600,000, even on a professional computing platform equipped with 512 GB of memory, the program may crash due to insufficient memory.

[0003] In order to cope with the computational pressure brought by large-scale single-cell sequencing data, researchers have proposed a variety of efficient single-cell analysis tools, which are mainly used in tasks such as data interpolation, integration, clustering and cell type annotation. However, these tools are usually designed specifically for specific tasks and are difficult to integrate directly into the existing single-cell data analysis framework. To achieve more general and efficient single-cell data processing, one solution is to compress the original data, thereby reducing data redundancy, so that traditional analysis tools can process large-scale data more efficiently. To achieve single-cell data compression, a representative method is metacell inference, which aggregates several single cells into representative metacells by aggregating biologically similar cell populations, thereby effectively reducing the number of cells to be processed. The metacell inference method has significant advantages in large-scale data processing. First, the data compression brought by metacells reduces the computational overhead and facilitates large-scale data processing. Secondly, by aggregating cells with similar characteristics, metacells alleviate the sparsity of the data, making downstream analysis (such as cell type annotation, developmental trajectory inference, etc.) more efficient and robust. However, although the metacell inference method has shown good results in some application scenarios, it still has shortcomings in terms of accuracy and efficiency on large-scale data sets, namely: the running time required for existing metacell inference methods increases exponentially with the number of cells, limiting its application on larger-scale data sets; at the same time, existing metacell inference algorithms usually only support single-omics data (such as gene sequencing data) and cannot directly process multi-omics data (such as gene + protein sequencing data); and in the field of single-cell data analysis, with the increasing amount of multi-omics data, the applicability of such algorithms is limited. Therefore, how to reduce the huge time overhead of metacell inference methods and realize metacell inference of multi-omics data is an urgent problem to be solved in the current single-cell analysis field. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides an efficient multi-omics meta-cell inference method based on single-cell quantification, which is used to solve the problems of low inference accuracy, low efficiency and high time cost of existing meta-cell inference methods on large-scale data sets.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is: An efficient multi-omics meta-cell inference method based on single-cell quantification, comprising the following steps: S1. Obtain cell sequencing data and standardize it to obtain standard cell sequencing data; Wherein, the cell sequencing data is a number of single-cell sequencing data or a number of multi-omics cell sequencing data; S2, initializing a quantization codebook, a neural network encoder, a first neural network decoder, and a second neural network decoder, and inputting standard cell sequencing data for training, and obtaining a trained neural network encoder, a first neural network decoder, a second neural network decoder, and a quantization codebook by performing feature extraction, cell quantization, and quantization reconstruction on the standard cell sequencing data; Among them, the quantization codebook includes several entries, each entry corresponds to the representation of a meta-cell; S3. Input the original single-cell sequencing data or the original multi-omics cell sequencing data into the trained neural network encoder for feature extraction, and perform cell quantization based on the trained quantization codebook to obtain the meta-cell inference result.

[0006] The present invention has the following beneficial effects: 1. The present invention proposes an efficient multi-omics meta-cell inference method based on single-cell quantification, which can achieve efficient inference on large-scale single-cell data sets. The existing meta-cell inference method has too high computational overhead when processing large-scale data, which increases exponentially with the number of cells, limiting its scalability. The method proposed in the present invention aims to achieve a meta-cell inference with linear time complexity, effectively reducing the computational cost while ensuring accuracy, so that it can support the processing of millions of single-cell data. 2. Support joint analysis of multi-omics data. As existing methods are mostly limited to single-omics data, it is difficult to integrate multi-omics information and fully explore the multi-dimensional characteristics of cells. The method proposed in the present invention can support meta-cell inference of multi-omics single-cell data and improve the ability to understand biological processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 A schematic diagram of a process of an efficient multi-omics meta-cell inference method based on single-cell quantification proposed in the present invention; Figure 2Schematic diagram of the working principles of the neural network encoder, the first neural network decoder, the second neural network decoder and the quantization codebook in the embodiment. DETAILED DESCRIPTION

[0008] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0009] like Figure 1 As shown, an efficient multi-omics meta-cell inference method based on single-cell quantification includes the following steps S1-S3: S1. Obtain cell sequencing data and standardize them to obtain standard cell sequencing data; wherein the cell sequencing data is a number of single-cell sequencing data or a number of multi-omics cell sequencing data.

[0010] In this embodiment, the purpose of standardizing the cell sequencing data is to unify the dimension so that the same set of neural network encoders can be used to extract features from different cells.

[0011] Specifically, step S1 includes S11-S12: S11. Acquire some cell sequencing data and generate a cell sequencing data matrix.

[0012] S12. Scale the total cell counts in each row of the cell sequencing data matrix by S times, take the natural logarithm, and normalize it to obtain standard cell sequencing data.

[0013] In this embodiment, if the cell sequencing data is single-cell sequencing data, that is, the single-cell counting matrix is ​​input ,in is the total number of single cells, Represents the feature dimension, so the standardization of single-cell sequencing data is as follows: scale the count of each cell (i.e., the sum of each row in the single-cell count matrix) by 10,000 times, take the natural logarithm of the scaled single-cell count matrix to smooth out excessive values, and finally standardize it to satisfy 0 mean and 1 variance to obtain standard cell sequencing data; in addition, the standardization process of multi-omics cell sequencing data is the same as that of single-cell sequencing data.

[0014] S2. Initialize a quantization codebook, a neural network encoder, a first neural network decoder and a second neural network decoder, and input standard cell sequencing data for training. By performing feature extraction, cell quantization and quantization reconstruction on the standard cell sequencing data, a trained neural network encoder, a first neural network decoder, a second neural network decoder and a quantization codebook are obtained; wherein the quantization codebook includes a number of entries, each entry corresponding to the representation of a meta-cell.

[0015] In this embodiment, the working principles of the quantization codebook, the neural network encoder, the first neural network decoder and the second neural network decoder are as follows: Figure 2 As shown, the neural network encoder is a fully connected neural network with a dimension of M-512-128-32; the first neural network decoder includes a first fully connected network, a first fully connected network layer, and a second fully connected network layer; the second neural network decoder includes a second fully connected network, a third fully connected network layer, and a fourth fully connected network layer; the dimensions of the first fully connected network and the second fully connected network are 32-128-512, the dimensions of the first fully connected network layer, the second fully connected network layer, the second fully connected network, and the third fully connected network layer are 512-M, and the dimension of the input layer corresponds to the feature dimension M of the input single cell matrix (normalized count matrix), and the quantization codebook has entries, each with a dimension of 32; its working principle is: feature extraction is performed in the neural network encoder, and then the extracted low-dimensional features of the cells are input into the first neural network decoder of the upper layer for reconstruction, and at the same time, the extracted low-dimensional features of the cells are input into the quantization codebook for cell quantization. After obtaining the cell quantization representation, the cell quantization representation is input into the second neural network decoder of the lower layer for quantization reconstruction, thereby helping to optimize the quantization codebook; in addition, the training process of single-cell sequencing data is the same as that of multi-omics cell sequencing data, but because multi-omics cell sequencing data includes several omics, in the training process of multi-omics cell sequencing data, each omics corresponds to a neural network encoder, a first neural network decoder, and a second neural network decoder. The specific training process of single-cell sequencing data or multi-omics cell sequencing data is as follows: Specifically, when the standard cell sequencing data is input for training in step S2, if the standard cell sequencing data is standard single-cell sequencing data, the standard single-cell sequencing data is input into the initialized neural network encoder, the first neural network decoder, the second neural network decoder and the quantization codebook for training to obtain the trained neural network encoder, the first neural network decoder, the second neural network decoder and the quantization codebook. The specific process is as follows: Step 1: Input the standard single-cell sequencing data into the neural network encoder for feature extraction. By mapping the high-dimensional features of the standard single-cell sequencing data into low-dimensional representations, a low-dimensional representation of the single-cell sequencing data is obtained.

[0016] In this embodiment, a fully connected neural network encoder is used to extract a low-dimensional representation of the standard single-cell sequencing data so that the low-dimensional representation can be sent to a neural network decoder for feature reconstruction in a subsequent step.

[0017] Step 2: Input the low-dimensional representation of the single-cell sequencing data into the first neural network decoder for feature reconstruction, and calculate the standard reconstruction loss function; wherein the first neural network decoder includes a first fully connected network, a first fully connected network layer, and a second fully connected network layer.

[0018] In this embodiment, during the cell reconstruction process, negative binomial distribution is used to model the original single-cell sequencing data, and the optimization goal of the neural network encoder and the first neural network decoder is to maximize the following log-likelihood (standard reconstruction loss function) to help extract low-dimensional representations of cells from standard single-cell sequencing data, namely:

[0019] in, represents the standard reconstruction loss function, represents the total number of single cells, represents the logarithmic function, represents the probability density function of the negative binomial distribution, Indicates The original count of single cells, , Respectively represent the first fully connected network layer and the second fully connected network layer, represents the mean of the negative binomial distribution calculated by the first fully connected network layer, represents the variance of the negative binomial distribution calculated by the second fully connected network layer, represents the gamma function, Indicates The scaling factor of the raw count of a single cell in the normalization process, represents a diagonal matrix, represents the exponential function, Indicates The representation of a single cell is calculated by the first fully connected network, Indicates The low-dimensional representation of a single cell extracted by the neural network encoder, represents the first fully connected network, Represents a neural network encoder.

[0020] Step 3: Calculate the cosine similarity between the low-dimensional representation of the single-cell sequencing data and each entry in the quantization codebook and sort them, obtain the entry with the most similar low-dimensional representation of the single-cell sequencing data, assign each single cell in the single-cell sequencing data to the entry that is most similar to it, and calculate the quantization codebook loss function.

[0021] In this embodiment, this step is a cell quantization operation, that is, in the low-dimensional representation space of the neural network, the cosine similarity between the cell representation and each entry in the quantization codebook is calculated, so as to assign each cell to the entry that is most similar to it.

[0022] Specifically, the most similar entry for obtaining the low-dimensional representation of single-cell sequencing data in step 3 is:

[0023] in, represents the quantization codebook, Indicates Quantitative characterization of single cells, represents the cell quantization operation, Represents the quantization codebook No. entries, It means to find the maximum value. represents cosine similarity.

[0024] In this embodiment, since each entry of the quantization codebook corresponds to the representation of a meta-cell, the total number of entries contained therein corresponds to the total number of target meta-cells given by the user.

[0025] Specifically, the quantization codebook loss function in step 3 is:

[0026] in, represents the quantization codebook loss function, Indicates a gradient stop operation.

[0027] In this embodiment, since the quantization codebook is initialized randomly, in order to help each entry in the quantization codebook to be closer to the low-dimensional representation of the cell, a quantization codebook loss function is used to optimize the quantization codebook. , indicating that the loss function is only used to optimize the quantization codebook without affecting the low-dimensional representation of the cell. This design can avoid the relatively random quantization codebook interfering with the low-dimensional representation of the cell in the early stage of training.

[0028] Step 4: Count the number of times each entry is used to assign a single cell, adjust entries with excessive or insufficient usage, update the quantization codebook, and input each entry of the updated quantization codebook into the second neural network decoder to reconstruct the low-dimensional representation of all single-cell sequencing data assigned to the entry, and calculate the quantization reconstruction loss function; wherein the second neural network decoder includes a second fully connected network, a third fully connected network layer, and a fourth fully connected network layer.

[0029] In this embodiment, since the above loss function only optimizes the entries assigned to cells, these optimized entries will be closer and closer to the low-dimensional representation of cells, and thus more and more used. However, those entries that are less used always correspond to fewer cells due to insufficient optimization, resulting in large differences in the number of original cells corresponding to different meta-cells. However, when a meta-cell corresponds to too many original cells, it may cover different cell types and states at the same time, causing some characteristics of the original cells to be diluted in the meta-cell; and when a meta-cell corresponds to too few original cells, it is difficult to effectively alleviate the sparsity of the data and play a role in data compression; therefore, this step counts the number of times each entry is assigned to a single cell, adjusts the entries with too much or too little usage, and updates the quantization codebook to ensure that the original cells are relatively evenly quantized to each entry.

[0030] Specifically, in step 4, the number of times each entry is used to allocate a single cell is counted, and entries with excessive or insufficient usage are adjusted. The specific process of updating the quantization codebook is as follows: First, count the number of times each entry allocates a single cell, that is:

[0031] in, Indicates The quantization codebook in the historical iteration process is counted at the iteration No. Items The number of times it has been used, Indicates The quantization codebook in the historical iteration process is counted at the iteration No. Items The number of times it has been used, Indicates The quantization codebook in the historical iteration process is counted at the iteration No. Items The number of times it has been used, Indicates the smoothing momentum of the control history and current usage times. Indicates Quantization codebook for iteration No. Items The number of times it has been used.

[0032] In this embodiment, The value is 0.9.

[0033] Secondly, adjust the entries with too little usage, and push the entries that are too little used to the cell representation farthest from the quantization codebook, that is, convert the farthest distance between the cell representation and the quantization codebook into the probability that the low-dimensional representation of the single-cell sequencing data is selected as the update target, that is:

[0034] in, Indicates that it is used to adjust the entries that are too little used. The probability that the low-dimensional representation of a single cell is selected as the update target, Indicates taking the maximum value, Represents the quantization codebook entries, Represents the total number of entries in the quantization codebook, that is, the number of target metacells.

[0035] by The probability of Select from low-dimensional representations of single-cell sequencing data Characterization of cells , and update the entries of the quantization codebook to obtain the updated quantization codebook, that is:

[0036] in, It represents the probability that the low-dimensional representation of the first single cell is selected as the update target when adjusting the underused items. Indicates that it is used to adjust the entries that are too little used. The probability that the low-dimensional representation of a single cell is selected as the update target, Indicates the cell representation selected to adjust the first entry which is underused. Indicates the selected one used to adjust the underused Cellular representation of items, Indicates Update the first entry after the less used entry in the iteration entries, Indicates Update the first entry after the less used entry in the iteration entries, represents the weight of controlling the update of entries that are rarely used. Indicates the selected Cellular representation of items, Indicates a minimum value.

[0037] In this embodiment, since the weights corresponding to the less used entries are higher, they will be pushed more to the farthest cell representation of the cluster quantization codebook. Therefore, the parameter is set Controls the weight of updating entries that are rarely used. Used to avoid excessive update weights, the value is 0.001.

[0038] Then, the entries with excessive usage are adjusted to push the overused entries to the cell representation with medium distance from them, that is, the medium distance between the cell representation and the quantization codebook is converted into the probability that the low-dimensional representation of the single-cell sequencing data is selected as the update target, that is:

[0039] in, Indicates that it is used to adjust the entries that are used too much. The low-dimensional features of a single cell are selected as The probability of updating the target of entries, represents the median, Indicates The median distance of the entries to the cell representation, Indicates A low-dimensional representation of a single cell.

[0040] by The probability of Select from low-dimensional representations of single-cell sequencing data Characterization of cells , and update the entries of the quantization codebook to obtain the updated quantization codebook, that is:

[0041] in, When used to adjust overused items, the low-dimensional representation of the first single cell is selected as The probability of updating the target of entries, Indicates that it is used to adjust the entries that are used too much. The low-dimensional representation of a single cell is selected as The probability of updating the target of entries, Indicates the cell representation selected to adjust the overused first entry, Indicates the selected one used to adjust the overused Cellular representation of items, Indicates In the next iteration, update the first entry after the overused entry entries, Indicates the weight of controlling the update of overused entries. Indicates the selected one used to adjust the overused Cellular representation of each item.

[0042] In this embodiment, since the weights corresponding to the entries that are used more frequently are higher, they will be more likely to be pushed to the cell representations in the middle of the distance quantization codebook. Controls the weight of updates for overused entries.

[0043] Specifically, the quantized reconstruction loss function in step 4 is:

[0044] in, represents the quantized reconstruction loss function, Indicates The quantitative representation of a single cell is calculated by the second fully connected network. , They represent the third fully connected network layer and the fourth fully connected network layer respectively. represents the mean of the negative binomial distribution calculated by the third fully connected network layer, represents the variance of the negative binomial distribution calculated by the fourth fully connected network layer, represents the second fully connected network.

[0045] In this embodiment, the meta-cell, as the compression result of the original cell, should fully aggregate the information in the original cell and have the common characteristics of the cell group it represents. In other words, ideally, a meta-cell should be able to effectively reconstruct several original cells it represents. To this end, in the above quantization codebook loss function In addition to the update strategy, each entry of the quantization codebook is also required to reconstruct the original counts of all cells assigned to the entry, that is, the quantization reconstruction loss function is used to optimize the entry of each quantization codebook to reconstruct the original counts of all cells assigned to the entry.

[0046] Step 5: Sum the standard reconstruction loss function, the quantization codebook loss function, and the quantization reconstruction loss function to obtain the total loss function, which is:

[0047] in, represents the total loss function, represents the standard reconstruction loss function, represents the quantization codebook loss function, Represents the quantized reconstruction loss function.

[0048] In this embodiment, the total loss function is used to optimize the neural network encoder, the first neural network decoder, the second neural network decoder and the quantization codebook .

[0049] Step 6: Based on the total loss function, the gradient descent method is used to optimize the neural network encoder, the first neural network decoder, the second neural network decoder and the quantization codebook, and it is determined whether the current number of iterations reaches the maximum number of iterations. If so, the trained neural network encoder, the first neural network decoder, the second neural network decoder and the quantization codebook are obtained. Otherwise, continue to execute steps Step 1-Step 5.

[0050] Specifically, when the standard cell sequencing data is input for training in step S2, if the standard sequencing data is standard multi-omics cell sequencing data, the training process is the same as the training process of standard single-cell sequencing data, and since the standard multi-omics cell sequencing data includes several omics, in the cell quantification process, the low-dimensional representation of each cell is the splicing of the representations of each omics, that is:

[0051] in, Represents the first A low-dimensional representation of a single cell, represents the total number of omics, Represents a splicing operation, Indicates The first low-dimensional representation of single-cell omics, Indicates Low-dimensional representation of single-cell second-omics, Indicates Single cell Low-dimensional representation of omics, Indicates Single cell Low-dimensional representation of omics, Indicates The neural network encoder corresponding to each omics, Indicates Single cell The count matrix after omics standardization.

[0052] In this embodiment, each omics in the standard multi-omics cell sequencing data needs to correspond to a neural network encoder.

[0053] Therefore, the total loss function for standard multi-omics sequencing data training is:

[0054] in, represents the total loss function when training with standard multi-omics sequencing data, Indicates The standard reconstruction loss function for omics, Indicates Quantitative reconstruction loss function for omics.

[0055] S3. Input the original single-cell sequencing data or the original multi-omics cell sequencing data into the trained neural network encoder for feature extraction, and perform cell quantization based on the trained quantization codebook to obtain the meta-cell inference result.

[0056] Specifically, step S3 specifically includes S31-S32: S31. Input the original single-cell sequencing data or the original multi-omics cell sequencing data into the trained neural network encoder for feature extraction, and assign each cell to the entry that is most similar to it based on the similarity between the cell representation and the trained quantization codebook entry.

[0057] S32. For each entry of the trained quantization codebook, the mean of all single-cell sequencing data or original multi-omics sequencing data assigned to the entry is calculated to obtain the meta-cell inference result, that is:

[0058] in, Indicates The inference results of individual cells are represents the first The number of cells per entry, represents the first entries.

[0059] In order to verify the effectiveness of the efficient multi-omics meta-cell inference method based on single-cell quantification proposed in the present invention, a comparative experiment was conducted on a large-scale human embryo single-cell dataset and a multi-omics human bone marrow dataset using the method proposed in the present invention and the most advanced meta-cell inference methods (including SEACell, MetaCell V2 and SuperCell methods) in the world, as follows: The meta-cell-based class-balanced classification accuracy is used to measure the effect of meta-cell inference. After predicting all the original cells corresponding to the same meta-cell to the type with the largest number of cells in the meta-cell, the following class-balanced classification accuracy is calculated:

[0060] in, represents the class-balanced classification accuracy, represents the total number of original single cells, Indicates The weight of a cell is calculated based on its cell type. represents the Dirichlet function, The predicted cell types, Indicates The actual cell type of each cell, Indicates The formula uses a weighting operation to emphasize the importance of rare cell types.

[0061] Experiment 1: A large-scale human embryo single-cell dataset was used, which contains 433,695 cells and 54 cell types. The distribution of the number of cells of each type is shown in Table 1: Table 1 Distribution of cell numbers of various types in large-scale human embryo single-cell dataset

[0062] 500 meta-cells are inferred from the above dataset, and the experimental results are shown in Table 2: Table 2 Comparison results of large-scale human embryo single-cell datasets

[0063] As can be seen from Table 2, the existing SEACell and MetaCell V2 methods will have memory overflow on a 512GB computing server due to huge memory overhead and cannot process the large-scale data set. SuperCell is currently the only method that can process the data set, but its effect is poor. The method proposed in the present invention can not only process millions of single-cell data, but also achieves exponential performance improvement compared to SuperCell, indicating that it can effectively compress large-scale single-cell data.

[0064] Experiment 2: A multi-omics human bone marrow dataset was used, which contained 30,672 cells, covering both gene and protein omics, and 27 cell types. The distribution of the number of cells of each type is shown in Table 3: Table 3 Distribution of cell numbers of various types in the multi-omics human bone marrow dataset

[0065] 613 metacells are inferred from the above dataset (corresponding to 50 times data compression), and the experimental results are shown in Table 4: Table 4 Comparison results of multi-omics human bone marrow datasets

[0066] Among them, since the existing methods do not support the inference of meta-cells from multi-omics data, in order to compare the performance, the data of the two omics need to be spliced ​​at the input level in the experiment and regarded as single-omics data as the input of the existing method. Therefore, it can be seen from Table 4 that the method proposed in the present invention has achieved a higher class-balanced classification accuracy than the existing method, indicating that it can effectively integrate the information of different omics to achieve accurate meta-cell inference. In addition, although the present invention has been verified on single-omics gene single-cell sequencing data and multi-omics gene + protein sequencing data, the proposed efficient multi-omics meta-cell inference method based on single-cell quantification is also applicable to other omics (such as transcriptomics, etc.); in the process of extracting low-dimensional features of cells, the method proposed in the present invention uses negative binomial distribution to model single-cell data, but other distributions (such as Gaussian distribution, Poisson distribution, etc.) can also be used for modeling.

[0067] In summary, the present invention realizes efficient multi-omics meta-cell inference through a single-cell quantization strategy; on the one hand, since both the neural network and the quantization codebook are optimized through batch data, the time overhead of the method proposed in the present invention is linear with the number of input single cells, which significantly reduces the computational overhead compared to the exponential time complexity of the prior art and improves the usability of the meta-cell inference method; on the other hand, the method proposed in the present invention is the first scheme that supports the inference of meta-cells from multi-omics data, and effectively fills the gap in multi-omics meta-cell inference methods at a time when multi-omics sequencing technology is developing rapidly.

[0068] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

[0069] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. An efficient multi-omics meta-cell inference method based on single-cell quantification, characterized in that: The following steps are involved: S1. Obtain cell sequencing data and standardize it to obtain standard cell sequencing data; Wherein, the cell sequencing data is a number of single-cell sequencing data or a number of multi-omics cell sequencing data; S2, initializing a quantization codebook, a neural network encoder, a first neural network decoder, and a second neural network decoder, and inputting standard cell sequencing data for training, and obtaining a trained neural network encoder, a first neural network decoder, a second neural network decoder, and a quantization codebook by performing feature extraction, cell quantization, and quantization reconstruction on the standard cell sequencing data; Among them, the quantization codebook includes several entries, each entry corresponds to the representation of a meta-cell; S3. Input the original single-cell sequencing data or the original multi-omics cell sequencing data into the trained neural network encoder for feature extraction, and perform cell quantization based on the trained quantization codebook to obtain the meta-cell inference result.

2. The efficient multi-omics meta-cell inference method based on single-cell quantification according to claim 1, characterized in that: Step S1 specifically includes: S11, obtaining a number of cell sequencing data and generating a cell sequencing data matrix; S12. Scale the total cell counts in each row of the cell sequencing data matrix by S times, take the natural logarithm, and normalize it to obtain standard cell sequencing data.

3. The efficient multi-omics meta-cell inference method based on single-cell quantification according to claim 2, characterized in that: When the standard cell sequencing data is input for training in step S2, if the standard cell sequencing data is standard single-cell sequencing data, the standard single-cell sequencing data is input into the initialized neural network encoder, the first neural network decoder, the second neural network decoder and the quantization codebook for training to obtain the trained neural network encoder, the first neural network decoder, the second neural network decoder and the quantization codebook. The specific process is as follows: Step 1: Input the standard single-cell sequencing data into the neural network encoder for feature extraction, and obtain the low-dimensional representation of the single-cell sequencing data by mapping the high-dimensional standard single-cell sequencing data features into low-dimensional representations; Step 2: Input the low-dimensional representation of the single-cell sequencing data into the first neural network decoder for feature reconstruction and calculate the standard reconstruction loss function; The first neural network decoder includes a first fully connected network, a first fully connected network layer, and a second fully connected network layer; Step 3: Calculate the cosine similarity between the low-dimensional representation of the single-cell sequencing data and each entry in the quantization codebook and sort them, obtain the entry with the most similar low-dimensional representation of the single-cell sequencing data, assign each single cell in the single-cell sequencing data to the entry that is most similar to it, and calculate the quantization codebook loss function; Step 4: Count the number of times each entry is used to allocate a single cell, adjust the entries with excessive or insufficient usage, update the quantization codebook, and input each entry of the updated quantization codebook into the second neural network decoder to reconstruct the low-dimensional representation of all single-cell sequencing data assigned to the entry, and calculate the quantization reconstruction loss function; The second neural network decoder includes a second fully connected network, a third fully connected network layer, and a fourth fully connected network layer; Step 5: Sum the standard reconstruction loss function, the quantization codebook loss function, and the quantization reconstruction loss function to obtain the total loss function, which is: in, represents the total loss function, represents the standard reconstruction loss function, represents the quantization codebook loss function, represents the quantized reconstruction loss function; Step 6: Based on the total loss function, the gradient descent method is used to optimize the neural network encoder, the first neural network decoder, the second neural network decoder and the quantization codebook, and it is determined whether the current number of iterations reaches the maximum number of iterations. If so, the trained neural network encoder, the first neural network decoder, the second neural network decoder and the quantization codebook are obtained. Otherwise, continue to execute steps Step 1-Step 5.

4. The efficient multi-omics meta-cell inference method based on single-cell quantification according to claim 3, characterized in that: The standard reconstruction loss function in step 2 is: in, represents the standard reconstruction loss function, represents the total number of single cells, represents the logarithmic function, represents the probability density function of the negative binomial distribution, Indicates The original count of single cells, , Respectively represent the first fully connected network layer and the second fully connected network layer, represents the mean of the negative binomial distribution calculated by the first fully connected network layer, represents the variance of the negative binomial distribution calculated by the second fully connected network layer, represents the gamma function, Indicates The scaling factor of the raw count of a single cell in the normalization process, represents a diagonal matrix, represents the exponential function, Indicates The representation of a single cell is calculated by the first fully connected network, Indicates The low-dimensional representation of a single cell extracted by the neural network encoder, represents the first fully connected network, Represents a neural network encoder.

5. The efficient multi-omics meta-cell inference method based on single-cell quantification according to claim 4, characterized in that: The most similar entry for obtaining the low-dimensional representation of single-cell sequencing data in step 3 is: in, represents the quantization codebook, Indicates Quantitative characterization of single cells, represents the cell quantization operation, Represents the quantization codebook No. entries, It means to find the maximum value. represents cosine similarity.

6. The efficient multi-omics meta-cell inference method based on single-cell quantification according to claim 5, characterized in that: The quantization codebook loss function in step 3 is: in, represents the quantization codebook loss function, Indicates a gradient stop operation.

7. The efficient multi-omics meta-cell inference method based on single-cell quantification according to claim 6, characterized in that: In step 4, the specific process of counting the number of times each entry is used to allocate a single cell, adjusting the entries with too much or too little usage, and updating the quantization codebook is as follows: First, count the number of times each entry allocates a single cell, that is: in, Indicates The quantization codebook in the historical iteration process is counted at the iteration No. Items The number of times it has been used, Indicates The quantization codebook in the historical iteration process is counted at the iteration No. Items The number of times it has been used, Indicates The quantization codebook in the historical iteration process is counted at the iteration No. Items The number of times it has been used, Indicates the smoothing momentum of the control history and current usage times. Indicates Quantization codebook for iteration No. Items The number of times it has been used; Secondly, adjust the entries with too little usage, and push the entries that are too little used to the cell representation farthest from the quantization codebook, that is, convert the farthest distance between the cell representation and the quantization codebook into the probability that the low-dimensional representation of the single-cell sequencing data is selected as the update target, that is: in, Indicates that it is used to adjust the entries that are too little used. The probability that the low-dimensional representation of a single cell is selected as the update target, Indicates taking the maximum value, Represents the quantization codebook entries, represents the total number of entries in the quantization codebook, i.e., the number of target metacells; by The probability of Select from low-dimensional representations of single-cell sequencing data Characterization of cells , and update the entries of the quantization codebook to obtain the updated quantization codebook, that is: in, It represents the probability that the low-dimensional representation of the first single cell is selected as the update target when adjusting the underused items. Indicates that it is used to adjust the entries that are too little used. The probability that the low-dimensional representation of a single cell is selected as the update target, Indicates the cell representation selected to adjust the first entry which is underused. Indicates the selected Cellular representation of items, Indicates Update the first entry after the less used entry in the iteration entries, Indicates Update the first entry after the less used entry in the iteration entries, represents the weight of controlling the update of entries that are rarely used. Indicates the selected Cellular representation of items, Indicates a minimum value; Then, the entries with excessive usage are adjusted to push the overused entries to the cell representation with medium distance from them, that is, the medium distance between the cell representation and the quantization codebook is converted into the probability that the low-dimensional representation of the single-cell sequencing data is selected as the update target, that is: in, Indicates that it is used to adjust the entries that are used too much. The low-dimensional features of a single cell are selected as The probability of updating the target of entries, represents the median, Indicates The median distance of the entries to the cell representation, Indicates Low-dimensional representation of single cells; by The probability of Select from low-dimensional representations of single-cell sequencing data Characterization of cells , and update the entries of the quantization codebook to obtain the updated quantization codebook, that is: in, When used to adjust overused items, the low-dimensional representation of the first single cell is selected as The probability of updating the target of entries, Indicates that it is used to adjust the entries that are used too much. The low-dimensional representation of a single cell is selected as The probability of updating the target of entries, Indicates the cell representation selected to adjust the overused first entry, Indicates the selected one used to adjust the overused Cellular representation of items, Indicates In the next iteration, update the first entry after the overused entry entries, Indicates the weight of controlling the update of overused entries. Indicates the selected one used to adjust the overused Cellular representation of each item.

8. The efficient multi-omics meta-cell inference method based on single-cell quantification according to claim 7, characterized in that: The quantized reconstruction loss function in step 4 is: in, represents the quantized reconstruction loss function, Indicates The quantitative representation of a single cell is calculated by the second fully connected network. , They represent the third fully connected network layer and the fourth fully connected network layer respectively. represents the mean of the negative binomial distribution calculated by the third fully connected network layer, represents the variance of the negative binomial distribution calculated by the fourth fully connected network layer, represents the second fully connected network.

9. The efficient multi-omics meta-cell inference method based on single-cell quantification according to claim 8, characterized in that: When inputting standard cell sequencing data for training in step S2, if the standard sequencing data is standard multi-omics cell sequencing data, the training process is the same as the training process of standard single-cell sequencing data. Since the standard multi-omics cell sequencing data includes several omics, in the cell quantification process, the low-dimensional representation of each cell is the splicing of the representations of each omics, that is: in, Represents the first A low-dimensional representation of a single cell, represents the total number of omics, Represents a splicing operation, Indicates The first low-dimensional representation of single-cell omics, Indicates Low-dimensional representation of single-cell second-omics, Indicates Single cell Low-dimensional representation of omics, Indicates Single cell Low-dimensional representation of omics, Indicates The neural network encoder corresponding to each omics, Indicates Single cell The count matrix after omics standardization; Therefore, the total loss function for standard multi-omics sequencing data training is: in, represents the total loss function when training with standard multi-omics sequencing data, Indicates The standard reconstruction loss function for omics, Indicates Quantitative reconstruction loss function for omics.

10. The efficient multi-omics meta-cell inference method based on single-cell quantification according to claim 9, characterized in that: Step S3 specifically includes: S31, inputting the original single-cell sequencing data or the original multi-omics cell sequencing data into the trained neural network encoder for feature extraction, and assigning each cell to the entry that is most similar to it according to the similarity between the cell representation and the trained quantization codebook entry; S32. For each entry of the trained quantization codebook, the mean of all single-cell sequencing data or original multi-omics sequencing data assigned to the entry is calculated to obtain the meta-cell inference result, that is: in, Indicates The inference results of individual cells are represents the first The number of cells per entry, represents the first entries.

Images