Single cell barcode identity recognition method based on SNP (Single Nucleotide Polymorphism)

Abstract

The invention relates to the technical field of bioinformatics, and discloses a single cell barcode identity recognition method based on SNP (Single Nucleotide Polymorphism), which comprises the following steps: selecting a plurality of high-variation SNP sites based on population genetics analysis; the selection of SNP sites is optimized by adopting a genome data integration method, so that the uniqueness of barcodes is ensured; noise filtering is carried out on barcodes of low-abundance cells by utilizing a machine learning algorithm, so that the recognition sensitivity is improved; in combination with phenotype information of cells, a multi-dimensional data fusion analysis technology is adopted, so that the diversity and accuracy of barcode are improved; redundant information is added into the barcode sequence, so that the stability of the barcode is enhanced, and the barcode is prevented from being lost. Through multiple SNP site combination and selection of high-variation SNP sites, it is ensured that barcode of each cell has high specificity and stability, meanwhile, the recognition rate of low-abundance cells is increased through a machine learning algorithm, and especially in a complex sample, accurate recognition can be achieved, and false positive and false negative can be reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of bioinformatics, and in particular to a single cell barcode identification method based on SNP polymorphism. Background Art

[0002] The single-cell barcode identification method based on SNP (single nucleotide polymorphism) polymorphism is a technology that uses genetic markers (SNPs) within cells to efficiently identify individual cells. In this method, a unique "barcode" is generated in the genome of each cell through specific SNP sites. These SNP markers are individual-specific and can therefore be used to distinguish different individual cells. Through high-throughput sequencing technology, researchers can obtain sequence information of these SNP sites at the single-cell level and accurately identify the identity of each cell through computational analysis. This method can not only accurately track and label individual cells, but also realize the diversity analysis of cell populations, which is of great significance for single-cell transcriptomics, genomics and other studies.

[0003] Although the single-cell barcode identification method based on SNP polymorphism has high resolution and specificity, it still has some shortcomings in the existing technology: Limitations of SNP site selection: SNP markers rely on specific genetic sites, but in different individuals or different species, the number of SNP sites suitable for use as barcodes is limited. For some populations, the diversity of SNP sites may be insufficient, resulting in reduced identification accuracy or difficulty in distinguishing very similar cells; high cost and high throughput requirements: SNP-based barcode identification methods usually require large-scale genome sequencing, which involves high costs, especially when high-throughput sequencing is performed at the single-cell level. The genomic information of each cell needs to be accurately extracted and analyzed, which places high demands on equipment and technology, resulting in increased overall costs; sample processing and data complexity: single-cell sequencing is usually accompanied by high background noise, especially in low-abundance samples, which may lead to the loss or misidentification of some SNP information. In addition, the identification of single-cell barcodes involves a large amount of data analysis, which requires powerful computing resources and algorithm support. The data processing process is complex and prone to errors. The sensitivity of SNP detection: For some SNP sites, especially in cells with low expression or low gene copy number, the detection sensitivity may be insufficient, resulting in the barcodes of these cells cannot be accurately identified. Different SNP sites may have different mutation frequencies. The mutation frequency of some SNP sites is low and may not be stably detected in large-scale single-cell sequencing. The challenge of intercellular heterogeneity: Genetic heterogeneity at the single-cell level may cause similar SNP mutations to appear in different cells, resulting in cross-recognition or mislabeling between "barcodes". Especially in highly heterogeneous cell populations, the genetic differences between cells are small, which may make it difficult to make accurate distinctions; It is time-consuming: When performing SNP-based identification, a series of experiments and data processing steps are required to extract and analyze the genetic information of each single cell. For large-scale single-cell analysis, the entire process may take a long time, reducing the overall experimental efficiency; dependence on a specific reference genome: This method usually needs to rely on an existing reference genome to select SNP sites. For some species that are not fully annotated or lack a reference genome, the SNP-based barcode identification method may not be effectively applied, limiting its versatility.

[0004] To this end, we propose a single-cell barcode identification method based on SNP polymorphism. Summary of the Invention

[0005] To achieve the above objectives, the present invention provides the following technical solution: a single cell barcode identification method based on SNP polymorphism, comprising the following steps:

[0006] S1: Select multiple highly variable SNP sites based on population genetics analysis;

[0007] S2: Use genomic data integration methods to optimize the selection of SNP sites to ensure the uniqueness of barcodes;

[0008] S3: Use machine learning algorithms to filter noise from barcodes of low-abundance cells to improve recognition sensitivity;

[0009] S4: Combined with cell phenotypic information, multi-dimensional data fusion analysis technology is used to improve the diversity and accuracy of barcodes;

[0010] S5: Add redundant information to the barcode sequence to enhance the stability of the barcode and avoid barcode loss;

[0011] S6: Decode barcode data through deep learning algorithms to ensure recognition accuracy.

[0012] Preferably, the selection of the SNP sites is based on genetic diversity analysis and population variation assessment based on whole genome data, and the selected SNP sites include at least 10 highly variable sites.

[0013] Preferably, the low-abundance cell recognition algorithm adopts a support vector machine (SVM) or convolutional neural network (CNN) model to further improve the recognition sensitivity of low-abundance cells.

[0014] Preferably, the redundant information enhances the stability of the barcode by splicing short sequences or inserting repeated sequences. The design of the redundant information ensures that the barcode is not lost during high-throughput sequencing.

[0015] Preferably, the data fusion analysis uses a statistical learning method, combined with information such as cell morphology, functional characteristics and marker expression, to further enhance the diversity and accuracy of the barcode.

[0016] Preferably, the data decoding is implemented through a deep learning network or other efficient computing models, and the decoding process further optimizes the recognition accuracy of the barcode information.

[0017] Preferably, the SNP site selection basis further includes the spanning range of the gene segment, ensuring that the selected SNP sites can be widely distributed in different gene regions, thereby increasing the diversity of the cell barcode.

[0018] Preferably, during the decoding process of the barcode data, population variation data is combined to optimize the impact of low-variance regions on the barcode information, thereby further improving accuracy.

[0019] Preferably, the SNP site selection strategy further includes analyzing the variability between populations through population genetic diversity assessment, and selecting the SNP site combination with the largest variability.

[0020] Compared with the existing technology, the present invention provides a single cell barcode identification method based on SNP polymorphism, which has the following beneficial effects:

[0021] 1. This single-cell barcode identification method based on SNP polymorphism ensures the high specificity and stability of each cell's barcode through the combination of multiple SNP sites and the selection of highly variable SNP sites. At the same time, it improves the recognition rate of low-abundance cells through machine learning algorithms, especially in complex samples, and can accurately identify and reduce the occurrence of false positives and false negatives.

[0022] 2. This single-cell barcode identification method based on SNP polymorphism adopts genomic data integration and genetic diversity analysis strategies, making it applicable to different species and individual populations. It has strong versatility and can meet different research needs. By introducing redundant information, it avoids the problems of barcode loss or mismatching in traditional methods and improves the stability and reliability of the barcode, especially in high-throughput sequencing processes.

[0023] 3. This single-cell barcode identification method based on SNP polymorphism can more accurately identify and classify cells by integrating cell phenotypic data, further improving the diversity and accuracy of the method. Through optimized SNP site selection and efficient data decoding methods, it reduces the cost of high-throughput sequencing and improves data processing efficiency, making it suitable for large-scale single-cell analysis. DETAILED DESCRIPTION

[0024] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention.

[0025] Example

[0026] An embodiment of a single cell barcode identification method based on SNP polymorphism

[0027] A single cell barcode identification method based on SNP polymorphism includes the following steps:

[0028] S1: Select multiple highly variable SNP sites based on population genetics analysis;

[0029] S2: Use genomic data integration methods to optimize the selection of SNP sites to ensure the uniqueness of barcodes;

[0030] S3: Use machine learning algorithms to filter noise from barcodes of low-abundance cells to improve recognition sensitivity;

[0031] S4: Combined with cell phenotypic information, multi-dimensional data fusion analysis technology is used to improve the diversity and accuracy of barcodes;

[0032] S5: Add redundant information to the barcode sequence to enhance the stability of the barcode and avoid barcode loss;

[0033] S6: Decode barcode data through deep learning algorithms to ensure recognition accuracy.

[0034] Specifically, the selection of SNP sites is based on genetic diversity analysis and population variation assessment based on whole genome data, and the selected SNP sites include at least 10 highly variable sites.

[0035] Specifically, the low-abundance cell recognition algorithm uses a support vector machine (SVM) or convolutional neural network (CNN) model to further improve the recognition sensitivity of low-abundance cells.

[0036] Specifically, the redundant information enhances the stability of the barcode by splicing short sequences or inserting repeated sequences. The design of the redundant information ensures that the barcode is not lost during high-throughput sequencing.

[0037] Specifically, data fusion analysis uses statistical learning methods to combine information such as cell morphology, functional characteristics, and marker expression to further enhance the diversity and accuracy of barcodes.

[0038] Specifically, data decoding is achieved through a deep learning network or other efficient computing models, and the decoding process further optimizes the recognition accuracy of the barcode information.

[0039] Specifically, the selection criteria for SNP sites further include the spanning range of gene segments, ensuring that the selected SNP sites can be widely distributed in different gene regions, thereby increasing the diversity of cell barcodes.

[0040] Specifically, during the decoding process of barcode data, population variation data is combined to optimize the impact of low-variance regions on barcode information and further improve accuracy.

[0041] Specifically, the SNP site selection strategy further includes analyzing the variability between populations through population genetic diversity assessment and selecting the SNP site combination with the greatest variability.

[0042] Through the above technical solutions, in the present invention, through the combination of multiple SNP sites and the selection of highly variable SNP sites, the barcode of each cell is ensured to have high specificity and stability. At the same time, the recognition rate of low-abundance cells is improved through machine learning algorithms. Especially in complex samples, it can accurately identify and reduce the occurrence of false positives and false negatives. By adopting genomic data integration and genetic diversity analysis strategies, the method is applicable to different species and individual populations, has strong versatility, and can meet different research needs. By introducing redundant information, the problem of barcode loss or mismatch in traditional methods is avoided, and the stability and reliability of the barcode are improved. Especially in the high-throughput sequencing process, by integrating the phenotypic data of the cells, more accurate identity identification and classification analysis of the cells can be performed, further improving the diversity and accuracy of the method. Through optimized SNP site selection and efficient data decoding methods, the cost of high-throughput sequencing is reduced, and the data processing efficiency is also improved, making it suitable for large-scale single-cell analysis.

[0043] SNP site selection and optimization

[0044] SNP site selection:

[0045] Obtain a set of highly variable SNPs from a human genome database (e.g., the 1000 Genomes Project). Using genomic variation analysis methods and population genetics analysis, calculate the variation frequency of each site in the target population. Select at least 10 highly variable SNPs with high variation rates across individuals and distributed across multiple genomic regions to ensure the uniqueness and reliability of the cell barcode.

[0046] SNP site optimization and data integration:

[0047] Using genomic data integration methods, data from different individuals or species are integrated and analyzed across genomic regions to optimize SNP selection. This strategy ensures that the selected SNPs are not only effective within a specific species or population but also have the same diversity effects across different backgrounds. This approach improves the adaptability of barcodes across different samples.

[0048] Low-abundance cell identification and data denoising

[0049] Low-abundance cell identification algorithm:

[0050] Deep learning algorithms, such as support vector machines (SVMs) or convolutional neural networks (CNNs), are used to analyze high-throughput sequencing data. By training on sample data from low-abundance cells, the model can identify and remove background noise. This is particularly true when the number of cells is small or in complex samples. Machine learning models can effectively enhance the sensitivity of identifying low-abundance cells.

[0051] Denoising and accuracy improvement:

[0052] Using training samples, the model is able to identify the unique pattern of each cell barcode. This algorithm not only removes background noise from the data but also optimizes the matching of cell barcodes, avoiding misidentification of low-abundance cells. Ultimately, through high-precision recognition, it can accurately restore the genetic signature of each cell.

[0053] Redundant barcode design

[0054] Redundant information added:

[0055] To ensure the stability of the barcode and prevent loss, redundant information is added to each barcode using techniques such as short sequence splicing or repeated sequence insertion. This redundant information ensures that even if some SNP sites cannot be correctly read, the complete barcode information can still be supplemented through other redundant sites.

[0056] Redundancy enhancement effect:

[0057] Through this redundant design, even if data loss or sequencing errors occur in high-throughput sequencing, the stability and accuracy of the barcode information can be guaranteed, thereby significantly reducing the problem of barcode loss caused by sequencing errors.

[0058] Multi-dimensional data fusion analysis

[0059] Fusion of cell phenotypic data:

[0060] Based on SNP data, we further combine cell phenotypic information (such as morphology, functional characteristics, marker expression, etc.) and use multi-dimensional data fusion analysis technology to enhance the diversity of each cell barcode. Through comprehensive analysis of cell functional characteristics and combined with SNP variation information, the accuracy of cell identification is further improved.

[0061] Algorithm optimization for data fusion:

[0062] Statistical learning methods (such as principal component analysis (PCA) and correlation analysis) are used to fuse cell phenotypic data and genomic data, ensuring that during the analysis process, the cell's phenotypic characteristics and genetic information can work together to generate and identify barcodes, thereby improving the reliability of the results.

[0063] Data decoding and identification

[0064] Deep Learning Decoding:

[0065] Single-cell barcode data is decoded using a deep learning network, such as a convolutional neural network (CNN) or a long short-term memory network (LSTM). The barcodes are then pattern-recognized by the neural network model to accurately restore the identity of each cell.

[0066] Population variation analysis combined with:

[0067] During the decoding process, the impact of low-variance regions is optimized and adjusted by combining population variation data. By combining population genetic data, the interference of low-frequency variation between cells can be further eliminated, ensuring the accuracy and stability of barcode decoding.

[0068] Application Verification

[0069] Verification and experiment:

[0070] The selected SNP sites and designed barcode scheme were experimentally validated, and single-cell samples were processed through high-throughput sequencing. The experimental results showed that the single-cell barcodes generated by this method can efficiently and accurately identify each cell, and significantly outperform traditional methods in terms of diversity and sensitivity.

[0071] Sample suitability:

[0072] This method can be applied to a variety of sample types, including humans, animals, plants, and microorganisms. The method can effectively generate highly specific and diverse barcodes in cell populations of different species, demonstrating strong universality.

[0073] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A single cell barcode identification method based on SNP polymorphism, characterized by: The following steps are involved: S1: Select multiple highly variable SNP sites based on population genetics analysis; S2: Use genomic data integration methods to optimize the selection of SNP sites to ensure the uniqueness of barcodes; S3: Use machine learning algorithms to filter noise from barcodes of low-abundance cells to improve recognition sensitivity; S4: Combined with cell phenotypic information, multi-dimensional data fusion analysis technology is used to improve the diversity and accuracy of barcodes; S5: Add redundant information to the barcode sequence to enhance the stability of the barcode and avoid barcode loss; S6: Decode barcode data through deep learning algorithms to ensure recognition accuracy.

2. The single cell barcode identification method based on SNP polymorphism according to claim 1, characterized in that: The selection of the SNP sites is based on genetic diversity analysis and population variation assessment based on whole genome data, and the selected SNP sites include at least 10 highly variable sites.

3. The single cell barcode identification method based on SNP polymorphism according to claim 1, characterized in that: The low-abundance cell recognition algorithm adopts a support vector machine (SVM) or convolutional neural network (CNN) model to further improve the recognition sensitivity of low-abundance cells.

4. The single cell barcode identification method based on SNP polymorphism according to claim 1, characterized in that: The redundant information enhances the stability of the barcode by splicing short sequences or inserting repeated sequences. The design of the redundant information ensures that the barcode is not lost during high-throughput sequencing.

5. The single cell barcode identification method based on SNP polymorphism according to claim 1, characterized in that: The data fusion analysis uses statistical learning methods to combine information such as cell morphology, functional characteristics, and marker expression to further enhance the diversity and accuracy of barcodes.

6. The single cell barcode identification method based on SNP polymorphism according to claim 1, characterized in that: The data decoding is achieved through a deep learning network or other efficient computing models, and the decoding process further optimizes the recognition accuracy of the barcode information.

7. The single cell barcode identification method based on SNP polymorphism according to claim 1, characterized in that: The SNP site selection criteria further include the spanning range of the gene segment, ensuring that the selected SNP sites can be widely distributed in different gene regions, thereby increasing the diversity of cell barcodes.

8. The single cell barcode identification method based on SNP polymorphism according to claim 1, characterized in that: During the decoding process of the barcode data, the population variation data is combined to optimize the impact of low-variance regions on the barcode information, thereby further improving accuracy.

9. The single cell barcode identification method based on SNP polymorphism according to claim 1, characterized in that: The SNP site selection strategy further includes analyzing the variability between populations through population genetic diversity assessment and selecting the SNP site combination with the largest variability.