A method, device and medium for analyzing single-cell transcriptome data of a PDX model

Abstract

The present invention discloses an analysis method, device and medium for single-cell transcriptome data of a PDX model, and belongs to the field of bioinformatics technology. The analysis method includes: comparing the single-cell transcriptome sequencing data of the PDX model with a mixed genome library to obtain a cell-gene expression spectrum matrix; identifying cells as human cells, mouse cells or double cells according to the ratio of human or mouse genes; extracting cell sequences from the single-cell transcriptome sequencing data based on the cell barcode, and comparing with the reference genome of the source species to obtain the corresponding cell-gene expression spectrum, which can be used for differential gene analysis, functional enrichment analysis, etc., to obtain more data and provide technical support for the clinical treatment of cancer. Further, the cell-integrated gene expression spectrum of multiple species can be obtained based on the integrated gene sequence set, and the interaction mechanism between human and mouse cells can be further obtained through cell clustering analysis.

Description

[0001] Related patents

[0002] This application is a divisional application of the Chinese invention patent application with application number 2023100214777, application date January 7, 2023, and invention name “PDX-based single-cell transcriptome data analysis method, system, device and medium”. Technical Field

[0003] The present invention belongs to the technical field of biological data processing, and in particular, relates to an analysis method, device and medium for single-cell transcriptome data of a PDX model. Background Art

[0004] The PDX (patient-derived tumor xenograft) model is a tumor model constructed by transplanting tumor tissue or primary cells from patients into NSG (immunodeficient) mice. This model is a direct transplantation of tumor tissue into NSG mice without any artificial culture, so most of the characteristics of primary tumors are retained at the histopathological, molecular biological and genetic levels, and are more similar to clinical samples. The PDX model is the tumor model that is closest to clinical samples so far, and this model is of great significance for clinical tumor evaluation, treatment and prognosis.

[0005] Using PDX models for single-cell transcriptome sequencing can deeply analyze the cell types and gene expression characteristics at different stages of the tumor, thereby providing guidance for tumor treatment. At present, although the 10× official analysis software cellranger can perform data analysis on multiple species mixed libraries of PDX models and obtain the corresponding expression spectrum matrix and clustering results. However, due to the large number of homologous genes between human and mouse species, even if cells from humans are analyzed through a mixed genome of humans and mice, some reads can be compared to the mouse genome. The cell clustering results directly based on the expression spectrum and the results obtained by downstream data analysis and mining will be inaccurate. And because there are homologous sequences in the human and mouse genomes, it is not possible to directly use sequences that are compared to the human genome for analysis, which will cause some cells from mice to be mistakenly identified as human cells.

[0006] In addition, in the study of PDX models, in addition to studying the differences in gene expression changes of human-related tumor cells under different treatment regimens to find corresponding drug treatment targets, it is also necessary to study how cells in mice interact with introduced human cells. Therefore, how to separate the cell gene expression profile of a single species from the cell sequencing data obtained from the PDX model and how to obtain the cell gene expression profile of multiple species are crucial. However, there is currently no method for systematically analyzing the single-cell transcriptome based on the PDX model. Summary of the invention

[0007] In order to solve at least one of the above technical problems, the technical solution adopted by the present invention is as follows:

[0008] The first aspect of the present invention provides a single-cell transcriptome data analysis method based on PDX, comprising the following steps:

[0009] S1, comparing the obtained single-cell transcriptome sequencing data of the PDX model with a mixed genome library of humans and mice to obtain a cell-gene expression profile matrix based on the mixed genome library, wherein the mixed genome library is obtained by merging the reference genome files and gene annotation files of humans and mice;

[0010] S2, identifying the cell as a human cell, a mouse cell, or a double cell according to whether the proportion of human genes expressed in the cell or the proportion of mouse genes expressed in the cell is greater than or equal to a first preset threshold value P1;

[0011] S3, based on the cell barcode, extracting sequences identified as human cells and sequences identified as mouse cells from the single-cell transcriptome sequencing data;

[0012] S4, compare the obtained human cell sequence with the human reference genome, and compare the obtained mouse cell sequence with the mouse reference genome to obtain the corresponding cell-gene expression profile,

[0013] Among them, the setting of P1 makes the difference between the double cell rate and half of the multicellular rate not exceed 5%. The double cell rate and multicellular rate are calculated by the following formula:

[0014] Double cell rate = (double cell number / (human cell number + mouse cell number + double cell number)) × 100%;

[0015] Multicellularity = (number of captured cells × 7.589 × 10 -6 +5.272×10 -4 )×100%.

[0016] Since the multicellularity rate includes a mixture of human and human cells, mouse and mouse cells in addition to double cells (mixed human and mouse cells), theoretically, the double cell rate is equal to 1 / 2 of the multicellularity rate. In the present invention, the difference between the double cell rate and half of the multicellularity rate does not exceed 5%, which means: |(double cell rate-multicellularity rate / 2)| / (multicellularity rate / 2)×100%≤5%. If P1 is set too high, too many cells will be judged as double cells, which is inconsistent with the actual situation.

[0017] In some specific embodiments of the present invention, P1 = 70%.

[0018] In some embodiments of the present invention, in step S1, when merging the genome files and gene annotation files of humans and mice, in order to avoid duplication of genes and chromosomes, specific labels are added before the gene ID, gene name, and chromosome to distinguish them. For example, "human" is added before the gene ID, gene name, and chromosome of humans, and "mouse" is added before the gene ID, gene name, and chromosome of mice. Further, a library file that can be used for comparison is generated based on the merged genome file and gene annotation file.

[0019] In some embodiments of the present invention, step S2 specifically comprises:

[0020] S21, count the number of human genes expressed in cells Nh and the number of mouse genes expressed Nm;

[0021] S22, calculating the ratio Ph of human genes expressed in the cell and the ratio Pm of mouse genes expressed, where Ph = Nh / (Nh+Nm), Pm = Nm / (Nh+Nm);

[0022] S23, if Ph is greater than or equal to the first preset threshold value P1, the cell is identified as a human cell, and if Pm is greater than or equal to the first preset threshold value P1, the cell is identified as a mouse cell.

[0023] In some embodiments of the present invention, cells that neither satisfy Ph greater than or equal to the first preset threshold P1 nor Pm greater than or equal to the first preset threshold P1 are determined to be double cells, that is, cells that have both human gene expression and mouse gene expression.

[0024] Barcode is also called index, i.e. barcode or label, which is usually used to distinguish different sources of sequences in sequencing technology. In the present invention, barcode is used to distinguish different cells, i.e. sequencing sequences with the same barcode in the sequencing results mean that they come from the same cell, so different barcodes can represent different cells. In some descriptions of the present invention, barcode and cell have the same meaning.

[0025] In some embodiments of the present invention, in step S3, the extraction of sequences identified as human cells and sequences identified as mouse cells specifically includes:

[0026] S31, identifying the barcode of the sequencing sequence, comparing it with the barcode of the cell to obtain a base matching coefficient Mi, where Mi = Lm / Lb, Lm is the number of bases in the barcode of the sequencing sequence that match the barcode of the cell, and Lb is the number of bases in the barcode of the cell;

[0027] S32, extracting sequences according to Mi and a second preset threshold P2:

[0028] If Mi = 100%, directly extract the corresponding sequence; when P2 ≤ Mi < 100% and the sequencing quality value of the bases corresponding to the sequencing reads that are not fully matched < 10, correct the sequencing reads to the correct bases and then extract the sequence; if Mi < P2, do not extract.

[0029] Among them, P2 ≥ 80%.

[0030] After identifying human cells and mouse cells using step S2, although the corresponding cell-gene expression profiles of human cells or mouse cells can also be obtained according to the comparison results of step S1. However, the comparison in step S1 is based on the mixed genomic library, and affected by homologous genes, the gene expression profile of each cell may not be accurate. Therefore, based on the extraction of the original sequencing data from human and mouse cells in step S3, it can be further compared with the reference genomes of humans and mice respectively to obtain the correct cell-gene expression profile matrix of humans or mice.

[0031] Different sequencing platforms usually have different barcode lengths. For example, for the Illumina 10× single-cell sequencing platform, the barcode length is 16, and for the MobiNova platform, the barcode length is 28. Different P2 values are selected according to different barcode lengths, and the generally selected standard is to accept only 1 to 2 base mismatches. In some embodiments of the present invention, the barcode length is 16, and P2 is set to 90%, allowing only 1 base to be unmatched.

[0032] In some embodiments of the present invention, in step S32, the correction of the sequencing reads to the correct bases means correcting the bases with a sequencing quality value < 10 in the barcode of the sequencing sequence to the bases at the corresponding positions of the barcode of the matching cell. If the sequencing quality is ≥ 10, correction cannot be performed, and this sequencing read does not belong to this cell, and other barcodes should be used for extraction or discarded.

[0033] In some embodiments of the present invention, in step S4, when constructing the cell-gene expression profile matrix, only genome alignment and UMI correction need to be performed, and there is no need to perform a cell identification process.

[0034] In some embodiments of the present invention, after step S2, the following steps are performed:

[0035] S3', comparing the single-cell transcriptome sequencing data with the integrated gene sequence set of human and mouse homologous genes to obtain a comparison result with the integrated genes;

[0036] S4', based on the human cell and mouse cell barcodes identified in step S2, obtain the cell-integrated gene expression profile from the comparison results obtained in step S3',

[0037] Wherein, the integrated gene sequence set of the homologous genes is obtained based on the following steps:

[0038] (1) Splicing the sequences of each homologous gene of humans and mice to obtain the integrated gene sequence of each homologous gene; filling the gap between the gene sequences of humans and mice with 60 to 100 N bases;

[0039] (2) Based on the integrated gene sequence, the corresponding annotation file is constructed: the homologous gene is regarded as an independent chromosome, and the homologous gene sequences from humans and mice are regarded as two transcripts.

[0040] In some embodiments of the present invention, 60 to 100 N bases are set in the integrated gene sequence in order to add species information to the collated annotation file to facilitate subsequent reads filtering. In some specific embodiments of the present invention, 80 N bases are set in the integrated gene sequence. That is, in an integrated gene sequence, 80 N bases are inserted between the homologous gene sequence from humans and the homologous gene sequence from mice. On the other hand, by designing N bases, it is also possible to prevent reads from being aligned across species during subsequent comparisons, that is, part of a read is aligned to a homologous gene in humans, and part of a read is aligned to a homologous gene in mice.

[0041] In some embodiments of the present invention, in steps S3' and S4', a step of filtering the comparison results is further included:

[0042] If the only position of an integrated gene is compared, the corresponding comparison information is retained and one comparison information is extracted; if multiple positions of the same integrated gene are compared, only one comparison information is retained; if different integrated genes are compared, the corresponding comparison information is filtered.

[0043] The second aspect of the present invention provides a single-cell transcriptome data analysis system based on PDX, comprising:

[0044] Data input module, used to obtain single-cell transcriptome sequencing data of PDX models;

[0045] A database storage module, used to store a human reference genome, a mouse reference genome, and a mixed genome library of humans and mice, wherein the mixed genome library is obtained by merging the reference genome files and gene annotation files of humans and mice;

[0046] A first comparison module, connected to the data input module and the database storage module, respectively, for comparing the single-cell transcriptome sequencing data with the mixed genome library;

[0047] A cell identification module, connected to the first comparison module, for identifying cells as human cells, mouse cells or double cells according to whether the proportion of human genes expressed in the cells or the proportion of mouse genes expressed in the cells is greater than or equal to a first preset threshold value P1;

[0048] A sequence acquisition module, connected to the cell identification module and the data input module, respectively, for extracting sequences identified as human cells and sequences identified as mouse cells from the single-cell transcriptome sequencing data based on the cell barcode;

[0049] The cell-gene expression profile construction module is connected to the sequence acquisition module and the database storage module respectively, and is used to compare the obtained human cell sequence with the human reference genome, and compare the obtained mouse cell sequence with the mouse reference genome, to obtain the corresponding cell-gene expression profile.

[0050] Among them, the setting of P1 makes the difference between the double cell rate and half of the multicellular rate not exceed 5%. The double cell rate and multicellular rate are calculated by the following formula:

[0051] Double cell rate = (double cell number / (human cell number + mouse cell number + double cell number)) × 100%;

[0052] Multicellularity = (number of captured cells × 7.589 × 10 -6 +5.272×10 -4 )×100%.

[0053] Since the multicellularity includes a mixture of human and human cells, mouse and mouse cells in addition to double cells (mixed human and mouse cells), theoretically, the double cell rate is equal to 1 / 2 of the multicellularity. In the present invention, the difference between the double cell rate and half of the multicellularity is no more than 5%, which means: |(double cell rate-multicellularity / 2)| / (multicellularity / 2)×100%≤5%.

[0054] In some specific embodiments of the present invention, P1 = 70%.

[0055] In some embodiments of the present invention, in the database storage module, when merging the genome files and gene annotation files of humans and mice, in order to avoid duplication of genes and chromosomes, specific labels are added before the gene ID, gene name, and chromosome for distinction. For example, "human" is added before the gene ID, gene name, and chromosome of humans, and "mouse" is added before the gene ID, gene name, and chromosome of mice. Further, a library file that can be used for comparison is generated based on the merged genome file and gene annotation file.

[0056] In some embodiments of the present invention, the cell identification module performs cell identification based on the following steps:

[0057] Count the number of human genes expressed in the cells, Nh, and the number of mouse genes expressed in the cells, Nm;

[0058] Calculate the ratio of human genes expressed in the cell, Ph, and the ratio of mouse genes expressed, Pm, where Ph = Nh / (Nh+Nm), Pm = Nm / (Nh+Nm);

[0059] If Ph is greater than or equal to the first preset threshold value P1, the cell is identified as a human cell, if Pm is greater than or equal to the first preset threshold value P1, the cell is identified as a mouse cell, and the rest are double cells.

[0060] In some embodiments of the present invention, cells that neither satisfy Ph greater than or equal to the first preset threshold P1 nor Pm greater than or equal to the first preset threshold P1 are determined to be double cells, that is, cells that have both human gene expression and mouse gene expression.

[0061] In some embodiments of the present invention, the sequence acquisition module extracts sequences identified as human cells and sequences identified as mouse cells by the following steps:

[0062] Identify the barcode of the sequencing sequence and compare it with the barcode of the cell to obtain the base matching coefficient Mi, Mi = Lm / Lb, where Lm is the number of bases that match the barcode of the sequencing sequence and the barcode of the cell, and Lb is the number of bases of the barcode of the cell;

[0063] Sequence extraction is performed according to Mi and the second preset threshold P2:

[0064] If Mi = 100%, the corresponding sequence is directly extracted; when P2 ≤ Mi < 100% and the sequencing quality value of the base corresponding to the uncompletely matched sequencing reads < 10, the sequencing reads are corrected to the correct base and then the sequence is extracted; if Mi < P2, no extraction is performed.

[0065] Among them, P2 ≥ 80%.

[0066] In some embodiments of the present invention, the database storage module is further used to store the integrated gene sequence set of human and mouse homologous genes, and the system further includes or does not include the sequence acquisition module and the cell-gene expression profile construction module but includes:

[0067] A second alignment module, which is respectively connected to the data input module and the database storage module, and is used to align the single-cell transcriptome sequencing data with the integrated gene sequence set of human and mouse homologous genes to obtain the alignment result with the integrated gene.

[0068] A cell-integrated gene expression profile construction module, which is respectively connected to the cell recognition module and the second alignment module, and is used to obtain the cell-integrated gene expression profile from the alignment result obtained by the second alignment module based on the human cell and mouse cell barcodes recognized by the cell recognition module.

[0069] Among them, the integrated gene sequence set of the homologous genes is obtained based on the following steps:

[0070] (1) The sequences of each homologous gene of human and mouse are spliced to obtain the integrated gene sequence of each homologous gene; 60-100 N bases are filled between the gene sequences of human and mouse.

[0071] (2) According to the integrated gene sequence, a corresponding annotation file is constructed: the homologous gene is used as an independent chromosome, and the homologous gene sequences from human and mouse are used as 2 transcripts.

[0072] The third aspect of the present invention provides a computer device, including:

[0073] A memory, which is used to store a computer program.

[0074] A processor, which is used to implement the steps of a method for analyzing single-cell transcriptome data based on PDX as described in any one of the first aspects of the present invention when executing the computer program.

[0075] The fourth aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of a method for analyzing single-cell transcriptome data based on PDX as described in any one of the first aspects of the present invention are implemented.

[0076] Beneficial Effects of the Invention

[0077] Compared with the prior art, the present invention has achieved the following beneficial effects:

[0078] The method and system of the present invention can be used to construct the cell-gene expression profile of human and / or mouse samples before and after PDS mouse model treatment (such as drugs or other treatment methods), and further differential gene analysis and functional enrichment analysis can be performed to obtain more data and provide technical support for the clinical treatment of cancer.

[0079] By using the method and system of the present invention to obtain the expression profile of cell-integrated genes, downstream cell clustering and differential gene search analysis can be performed. The clustering results obtained in this way can also cluster human and mouse cells well together, which can be used to understand the interaction mechanism between human and mouse cells. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 A graph showing the number of captured cells versus the multicellularity rate is shown.

[0081] Figure 2 A schematic diagram of the single-cell transcriptome sequencing results of the PDX model in Example 1 of the present invention is shown.

[0082] Figure 3 The clustering results based on mixed library analysis in Example 1 of the present invention are shown.

[0083] Figure 4 The results of cluster analysis based on sequencing data corresponding to extracted human cells in Example 1 of the present invention are shown.

[0084] Figure 5 The number of genes obtained based on mixed library analysis and analysis of sequencing data corresponding to extracted human cells in Example 1 of the present invention is shown.

[0085] Figure 6 The clustering results obtained using the cell-gene expression profiles constructed based on the hybrid database are shown.

[0086] Figure 7 Human and mouse homologous gene information is shown.

[0087] Figure 8 The mouse Cd3d gene sequence is shown.

[0088] Fig. 9 The integrated gene sequence obtained by splicing the human and mouse homologous Cd3d gene sequences is shown.

[0089] Fig.10The annotation file of the integrated gene sequence obtained by splicing the human and mouse homologous Cd3d gene sequences is shown.

[0090] Fig.11 The results of the expression profile matrix cluster analysis based on cell-integrated genes in Example 2 of the present invention are shown.

[0091] Fig.12 A schematic diagram of a single species cell-gene expression profile construction system in Example 3 of the present invention is shown.

[0092] Fig.13 A schematic diagram of a system for constructing cell-gene expression profiles of multiple species in Example 4 of the present invention is shown.

[0093] Fig.14 A schematic diagram of a single-cell transcriptome sequencing data analysis system for the PDX model in Example 5 of the present invention is shown. DETAILED DESCRIPTION

[0094] Unless otherwise indicated, implied from the context, or customary in the prior art, all parts and percentages in this application are based on weight, and the tests and characterization methods used are all synchronized with the filing date of this application. Where applicable, the contents of any patent, patent application or disclosure involved in this application are fully incorporated herein by reference, and their equivalent patent families are also introduced as references, especially the definitions of relevant terms in the art disclosed in these documents. If the definition of a specific term disclosed in the prior art is inconsistent with any definition provided in this application, the definition of the term provided in this application shall prevail.

[0095] In order to make the technical problems, technical solutions and beneficial effects solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments.

[0096] Example

[0097] The following examples are used to demonstrate preferred embodiments of the present invention. It will be appreciated by those skilled in the art that the techniques disclosed in the following examples represent techniques discovered by the inventors that can be used to implement the present invention and therefore can be considered as preferred embodiments of the present invention. However, it will be appreciated by those skilled in the art based on this specification that many modifications may be made to the specific embodiments disclosed herein and still achieve the same or similar results without departing from the spirit or scope of the present invention.

[0098] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, and the disclosure and materials cited therein are hereby incorporated by reference.

[0099] The experimental methods in the following examples are conventional methods unless otherwise specified. The instruments and equipment used in the following examples are conventional laboratory instruments and equipment unless otherwise specified; the experimental materials used in the following examples are purchased from conventional biochemical reagent stores unless otherwise specified.

[0100] Example 1 Construction of cell-gene expression profiles of a single species based on single-cell transcriptome data of PDX models

[0101] This embodiment provides a method for constructing a cell-gene expression profile of a single species based on single-cell transcriptome data of a PDX model, comprising the following steps:

[0102] 1. Obtaining single-cell transcriptome data of PDX models

[0103] The target tissue is dissociated to obtain a single cell suspension, and then the gel beads containing barcode information are combined with a mixture of cells and enzymes, and then wrapped by oil surfactant droplets in a microfluidic system to form GEMs (GelBeads-In-Emulsions). The gel beads in the GEM are dissolved, the cells are lysed to release mRNA, and barcoded cDNA for sequencing is produced by reverse transcription. After the liquid oil layer is destroyed, cDNA amplification, purification and library construction are performed. The completed library is sequenced to obtain single-cell transcriptome sequencing data.

[0104] 2. Obtaining cell-gene expression profiles based on a mixed mouse and human database

[0105] Construction of a mixed genome library of humans and mice: The genome files of humans and mice (GRCh38 and GRCm39) and gene annotation files (ensemble V105 version: Homo_sapiens.GRCh38.105.chr.gtf.gz, Mus_musculus.GRCm39.105.chr.gtf.gz) were merged. To avoid duplication of genes and chromosomes, "human" and "mouse" were added before the gene ID, gene name, and chromosome to distinguish them. Then, based on the merged genome file and gene annotation file, the cellranger software was used to generate a library file that can be used for star alignment.

[0106] For the single-cell transcriptome data of the PDX model obtained in step 1, the following process was performed using the cellranger software:

[0107] (1) Barcode extraction and correction

[0108] The first 16bp sequence of the R1 end of the sequencing was compared with the 10× whitelist. If the sequence came from the whitelist, the corresponding double-end reads were retained. If only one base of the sequenced reads was different from the whitelist, and the quality value of the sequenced bases at different positions was <10, the incorrectly matched bases were modified according to the whitelist and the reads were retained. Otherwise, the reads were not included in the subsequent expression profile matrix calculation.

[0109] (2) Sequence genome alignment

[0110] The sequenced R2 reads were aligned with the reference genome using the star software to obtain alignment information, and reads that only uniquely matched the positive transcript were retained and included in the statistical barcode-gene expression matrix.

[0111] (3) Cell recognition

[0112] According to the barcode-gene expression matrix, all the UMI counts measured for each barcode are counted and sorted from high to low. For the identification of cells with high RNA content, taking the capture of 10,000 cells as an example, if the UMI count of a barcode is greater than 1 / 10 of the UMI count value of the 101st barcode, it is considered a cell. Some cells with low RNA content may not meet the identification conditions just now. Select a barcode with a lower expression as the background, count the gene expression of the barcode, and then compare the gene expression corresponding to the barcode that was not identified as a cell in the first step with the background. If the gene expression is consistent, it is considered to be the background, otherwise the barcode is considered to be a cell.

[0113] Thus, a cell-gene expression profile matrix based on the mixed genomic library was obtained.

[0114] The specific expression profile data format is as follows:

[0115] Gene barcode 1 barcode 2 barcode 3 barcode 4 …… barcode n human_TNFRSF4 0 1 0 0 …… 2 human_CD3E 1 0 0 3 …… 0 …… …… …… …… …… …… …… mouse_Gm28417 0 0 1 0 0

[0116] 3. Identification of human and mouse cells

[0117] The number of human and mouse genes expressed in each cell (barcode) was counted based on the gene name prefix, and recorded as Nh and Nm respectively. The ratio of human and mouse genes in each barcode was calculated, and recorded as Ph and Pm respectively, where Ph = Nh / (Nh+Nm) and Pm = Nm / (Nh+Nm).

[0118] Since the comparison is done using a mixed human and mouse genomic library, and humans and mice have a large number of homologous genes, even cells from humans may express a small amount of mouse genes. If the Ph of a barcode is > 70%, the barcode is from human cells. If the Pm of a barcode is > 70%, the barcode is from mouse cells. The remaining barcodes are considered to be double cells. The "70%" here is the threshold set for identifying barcodes from human or mouse cells.

[0119] In the 10× single-cell sequencing process, the multicellularity rate is related to the number of cells captured, e.g. Figure 1 As shown, the relationship between the multicellularity rate and the number of captured cells is shown in the following formula:

[0120] Multicellularity = (number of captured cells × 7.589 × 10 -6 +5.272×10 -4) ×100%

[0121] However, among the above multi-cells, in addition to human and mouse cells mixed in one oil droplet (double cells), there are also cases where human and human cells, and mouse and mouse cells are mixed together. Therefore, theoretically, the double cell rate should be half of the multi-cell rate, that is, ((the number of captured cells × 7.589 × 10 -6 +5.272×10 -4 ) / 2). The inventors found that when the threshold is 70%, the obtained double cell rate is very close to half of the multi-cell rate.

[0122] The specific statistical table is shown in the following table (display part):

[0123] barcode Nh Nm Ph(%) Pm(%) Cell Type AAACCTGAGCTAACAA 1871 61 96.84 3.16 Human cells AAACCTGAGTTGCAGG 645 42 93.89 6.11 Human cells AAGGCAGAGCACACAG 897 1284 41.13 58.87 Double Cell AAGGCAGGTATCTGCA 102 1053 8.83 91.17 Mouse cells ACGAGCCGTAGCCTAT 1082 192 84.93 15.07 Human cells …… …… …… …… …… …… TATGCCCCAGCATGAG 321 1673 16.10 83.90 Mouse cells TGCACCTTCAATCACG 1295 825 61.08 38.92 Double Cell

[0124] Due to the influence of PDX models, sample sampling sites, sampling size, single cell dissociation, cell capture efficiency, etc., the proportion of human and mouse cells obtained from different samples is not completely consistent. The following table shows the actual number and proportion of human and mouse cells obtained from two different PDX samples:

[0125]

[0126] From the table above, we can see that the number of cells captured by sample 1 (number of human cells + number of mouse cells + number of double cells) is 8565 in total, and the multicellularity rate is calculated to be (8565×7.589×10 -6 +5.272×10 -4)×100%=6.55%. When the threshold is set to 70%, the obtained double cell rate is 3.19%, which is only 2.7% different from half of the multicellular rate (3.28%). The number of cells captured by sample 2 (human cells + mouse cells + double cells) is 8639 in total, and the calculated multicellular rate is (8636×7.589×10 -6 +5.272×10 -4 )×100%=6.61%. When the threshold is set to 70%, the obtained double cell rate is 3.25%, which is only 1.8% different from half of the multi-cell rate (3.31%). It can be seen that the above method is more accurate in identifying cells.

[0127] 4. Sequence Extraction

[0128] The sequence information corresponding to the barcodes from human and mouse cells determined above was extracted from the raw sequencing data. According to the library structure of the single-cell transcriptome, the barcode sequence comes from the first n bases of the R1 end of the double-end sequencing (for example, 10× single-cell platform, n=16, such as Figure 2 As shown in the figure, the cell barcode corresponding to the domestic single-cell platform is longer, for example, the cell barcode length of the MoZhuo platform is 28). The specific sequence extraction method is as follows:

[0129] 1) Calculate the base matching coefficient corresponding to each sequence: the cell barcode length is recorded as Lb, the actual matching base length is recorded as Lm, and the base matching coefficient is Mi, Mi = Lm / Lb.

[0130] 2) Considering that the cell barcode lengths set on different single-cell platforms are different, the base matching coefficient threshold is set to 90%, that is, when the cell barcode is 16 bases, only 1 base is allowed to fail to match. When Mi = 100%, the corresponding reads pair is directly output. When 90% ≤ Mi < 100%, the sequencing quality value of the corresponding base of the sequencing reads that is not fully matched is calculated. If the sequencing quality value of this position is < 10, it is considered that the sequencing error caused the failure to match, and the sequencing reads are corrected to the correct base, that is, the sequencing reads are corrected to the barcode that has been identified as the cell and the corresponding reads pair is output.

[0131] 5. Recover the cell-gene expression profile matrix of a single species

[0132] Although the expression spectrum matrix corresponding to human cells can be directly obtained based on step 3, since the cell-gene expression spectrum is based on a mixed library of humans and mice, the gene expression spectrum of each cell may not be accurate due to the influence of homologous genes.

[0133] Therefore, based on the original data from human and mouse cells obtained in step 4, the inventors re-analyzed the data using the human or mouse reference genome as the library file to obtain the correct cell-gene expression profiles of humans and mice.

[0134] Since all the sequences extracted in step 4 are from actual cells, unlike normal single-cell transcriptome sequencing, which may also contain empty oil-in-water data, when generating the expression matrix, only genome alignment and UMI correction are required, and cell identification is no longer required. If cell identification is performed, it is very likely that the same type of cells will be removed as background barcodes, which will affect subsequent data mining.

[0135] Using the above method, for PDX mouse models, the gene expression spectrum matrix from human or mouse cells in the samples can be obtained before and after treatment, and further data mining such as differential gene analysis and functional enrichment analysis can be performed.

[0136] Figure 3 The clustering results based on mixed library analysis are shown, showing that human and mouse cells are divided into left and right sides. The sequencing data corresponding to human cells were extracted and the results of clustering analysis are shown in Figure 1. Figure 4 shown.

[0137] The following table shows the differentially expressed genes based on mixed library analysis and analysis of raw data extracted from human cells:

[0138]

[0139] From the statistical results, it was found that there were a lot of differentially upregulated genes based on the mixed library. Since there were both human cells and mouse cells in it, using all cells for differentiation might result in too many differentially expressed genes. Therefore, only seven clusters from humans, clusters 3, 4, 5, 8, 9, 10, and 11, were analyzed for differences. The analysis results showed that a total of 5066 differentially expressed genes were found, of which 1209 were from mouse genes. This was obviously due to the mixed library and did not conform to the actual situation.

[0140] For the same barcode (from human cells), using mixed library analysis will result in a higher number of genes due to alignment with the mouse genome, such as Figure 5 As shown, the upper figure shows the mixed library analysis, and the lower figure shows the statistical results of the library analysis using extracted reads.

[0141] Example 2 Construction of multi-species cell-integrated gene expression profiles

[0142] Example 1 only analyzes the cell expression profile of a single species. In actual applications, it is found that some PDX models only introduce primary tumor cells. If you want to know how the mouse immune cells act on tumor cells and how the cells interact with each other, you still need to generate an expression profile matrix based on mixed data. Based on the cell-gene expression profile data of the mouse and human mixed database obtained in step 1 of Example 1, due to the different gene names between different species, when clustering cells, it will lead to a situation where they are clearly divided into two species groups (such as Figure 6 ).

[0143] In this case, the following steps are used to obtain the corrected multi-species cell-gene expression profile matrix:

[0144] (1) Obtain homologous gene information of human and mouse species from the Ensemble database, such as Figure 7 As shown, the first two columns are human gene ID and gene name information, and the last two columns are the corresponding mouse gene ID and gene name information. Humans and mice have a total of 18973 homologous genes (based on human genes, there are 18973 gene IDs that can find homologous genes from mice).

[0145] (2) Based on the gene chromosome location information and gene positive and negative strand information in the genome annotation file, the sequence information corresponding to each gene is extracted from the human and mouse genomes respectively. Taking the Cd3d gene as an example, the extracted sequence information is as follows Figure 8 shown.

[0146] (3) According to the homologous gene information obtained in step (1), the sequences are merged (80 N bases are used to fill the gap between the genes of two species in order to add species information to the sorted annotation file) to obtain the integrated gene. Taking the Cd3d gene as an example, its integrated gene sequence is as follows Fig. 9 shown.

[0147] (4) Based on the integrated gene sequence, construct the corresponding annotation file (homologous genes are treated as an independent chromosome, and homologous genes from two species are treated as two transcripts). Take merge0000001 as an example. Fig.10 shown.

[0148] (5) The reads obtained by sequencing are compared with the integrated gene sequence set as the reference genome to obtain the comparison results with the integrated gene sequence.

[0149] (6) Filtering of alignment results: For alignments on a unique position of an integrated gene (and the aligned sequences are only from the same species, i.e., not spanning N bases), retain the corresponding alignment information and extract a read multiple alignment information; if alignments are on multiple positions of the same integrated gene (and the multiple positions on the alignments are all from the same species, i.e., not spanning N bases), retain one alignment information; if alignments are on different integrated genes, filter out the corresponding alignment information. This step obtains about 40% of the alignment information.

[0150] (7) Based on the barcode information of humans and mice obtained in Example 1 (excluding double cell information), the expression spectrum matrix of the cell-integrated gene is obtained from the filtered comparison information obtained in step (6).

[0151] Based on the expression profile matrix of cell-integrated genes, downstream cell clustering and differential gene search analysis can be performed. In this way, the clustering results obtained can also cluster human and mouse cells well (such as Fig.11 can be used to understand the interaction mechanism between human and mouse cells.

[0152] Example 3 Single species cell-gene expression profile construction system

[0153] This embodiment provides a system for constructing a single-species cell-gene expression profile based on single-cell transcriptome data of a PDX model, such as Fig.12 As shown, including:

[0154] Data input module, used to obtain single-cell transcriptome sequencing data of PDX models;

[0155] A database storage module, used to store a human reference genome, a mouse reference genome, and a mixed genome library of humans and mice, wherein the mixed genome library is obtained by merging the reference genome files and gene annotation files of humans and mice, as described in Example 1;

[0156] A first comparison module, connected to the data input module and the database storage module, respectively, for comparing the single-cell transcriptome sequencing data with the mixed genome library;

[0157] A cell identification module, connected to the first comparison module, for identifying a cell as a human cell or a mouse cell according to whether the proportion of human genes expressed in the cell or the proportion of mouse genes expressed in the cell is greater than or equal to 70%;

[0158] A sequence acquisition module is connected to the cell identification module and the data input module respectively, and is used to extract sequences identified as human cells and sequences identified as mouse cells from the single-cell transcriptome sequencing data based on the cell barcode;

[0159] The cell-gene expression profile construction module is connected to the sequence acquisition module and the database storage module respectively, and is used to compare the obtained human cell sequence with the human reference genome, and to compare the obtained mouse cell sequence with the mouse reference genome, to obtain the corresponding cell-gene expression profile.

[0160] Example 4: Construction of a multi-species cell-integrated gene expression profiling system

[0161] This example provides a multi-species cell-integrated gene expression profile construction system based on Example 3, such as Fig.13 As shown, including:

[0162] Data input module, used to obtain single-cell transcriptome sequencing data of PDX models;

[0163] A database storage module, used to store a human reference genome, a mouse reference genome, and a mixed genome library of humans and mice, wherein the mixed genome library is obtained by merging the reference genome files and gene annotation files of humans and mice, as described in Example 1;

[0164] A first comparison module, connected to the data input module and the database storage module, respectively, for comparing the single-cell transcriptome sequencing data with the mixed genome library;

[0165] A cell identification module, connected to the first comparison module, for identifying a cell as a human cell, a mouse cell, or a double cell according to whether the proportion of human genes expressed in the cell or the proportion of mouse genes expressed in the cell is greater than or equal to 70%;

[0166] The second comparison module is connected to the data input module and the database storage module respectively, and is used to compare the single-cell transcriptome sequencing data with the integrated gene sequence set of human and mouse homologous genes to obtain a comparison result with the integrated genes, wherein the construction method of the integrated gene sequence set refers to Example 2.

[0167] The cell-integrated gene expression profile construction module is connected to the cell recognition module and the second comparison module respectively, and is used to obtain the cell-integrated gene expression profile from the comparison results obtained by the second comparison module based on the human cell and mouse cell barcodes identified by the cell recognition module.

[0168] Example 5 Single-cell transcriptome sequencing data analysis system for PDX models

[0169] This embodiment provides a single cell transcriptome sequencing data analysis system for a PDX model, combining the systems of Embodiments 3 and 4, such as Fig.14 As shown, including:

[0170] Data input module, used to obtain single-cell transcriptome sequencing data of PDX models;

[0171] A database storage module, used to store a human reference genome, a mouse reference genome, and a mixed genome library of humans and mice, wherein the mixed genome library is obtained by merging the reference genome files and gene annotation files of humans and mice, as described in Example 1;

[0172] A first comparison module, connected to the data input module and the database storage module, respectively, for comparing the single-cell transcriptome sequencing data with the mixed genome library;

[0173] A cell identification module, connected to the first comparison module, for identifying a cell as a human cell, a mouse cell, or a double cell according to whether the proportion of human genes expressed in the cell or the proportion of mouse genes expressed in the cell is greater than or equal to 70%;

[0174] A sequence acquisition module is connected to the cell identification module and the data input module respectively, and is used to extract sequences identified as human cells and sequences identified as mouse cells from the single-cell transcriptome sequencing data based on the cell barcode;

[0175] The cell-gene expression profile construction module is connected to the sequence acquisition module and the database storage module respectively, and is used to compare the obtained human cell sequence with the human reference genome, and to compare the obtained mouse cell sequence with the mouse reference genome, to obtain the corresponding cell-gene expression profile.

[0176] Also includes:

[0177] The second comparison module is connected to the data input module and the database storage module respectively, and is used to compare the single-cell transcriptome sequencing data with the integrated gene sequence set of human and mouse homologous genes to obtain a comparison result with the integrated genes, wherein the construction method of the integrated gene sequence set refers to Example 2.

[0178] The cell-integrated gene expression profile construction module is connected to the cell recognition module and the second comparison module respectively, and is used to obtain the cell-integrated gene expression profile from the comparison results obtained by the second comparison module based on the human cell and mouse cell barcodes identified by the cell recognition module.

[0179] All documents mentioned in the present invention are cited as references in this application, just as each document is cited as reference individually. In addition, it should be understood that after reading the above teachings of the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the claims attached to this application.

Claims

1. A method for analyzing single-cell transcriptome data of a PDX model, characterized in that: The following steps are involved: S1, comparing the obtained single-cell transcriptome sequencing data of the PDX model with a mixed genome library of humans and mice to obtain a cell-gene expression profile matrix based on the mixed genome library, wherein the mixed genome library is obtained by merging the reference genome files and gene annotation files of humans and mice; S2, identifying the cell as a human cell, a mouse cell, or a double cell according to whether the proportion of human genes expressed in the cell or the proportion of mouse genes expressed in the cell is greater than or equal to a first preset threshold value P1; S3, based on the cell barcode, extracting sequences identified as human cells and sequences identified as mouse cells from the single-cell transcriptome sequencing data; S4, compare the obtained human cell sequence with the human reference genome, and compare the obtained mouse cell sequence with the mouse reference genome to obtain the corresponding cell-gene expression profile, S5, comparing the single-cell transcriptome sequencing data with the integrated gene sequence set of human and mouse homologous genes to obtain a comparison result with the integrated genes; S6, based on the human cell and mouse cell barcodes identified in step S2, obtaining a cell-integrated gene expression profile from the comparison results obtained in step S5, Wherein, the integrated gene sequence set of the homologous genes is obtained based on the following steps: (1) Splicing the sequences of each homologous gene of humans and mice to obtain the integrated gene sequence of each homologous gene; filling the gap between the gene sequences of humans and mice with 60 to 100 N bases; According to the integrated gene sequence, the corresponding annotation file is constructed: the homologous gene is regarded as an independent chromosome, and the homologous gene sequences from humans and mice are regarded as two transcripts. P1 is set so that the difference between the double cell rate and half of the multicellularity rate does not exceed 5%. The double cell rate and multicellularity rate are calculated by the following formula: Double cell rate = (double cell number / (human cell number + mouse cell number + double cell number)) × 100%; Multicellularity = (number of captured cells × 7.589 × 10 -6 +5.272×10 -4 )×100%.

2. The method for analyzing single-cell transcriptome data of a PDX model according to claim 1, characterized in that: Step S2 specifically includes: S21, count the number of human genes expressed in cells Nh and the number of mouse genes expressed Nm; S22, calculating the ratio Ph of human genes expressed in the cell and the ratio Pm of mouse genes expressed, where Ph = Nh / (Nh+Nm), Pm = Nm / (Nh+Nm); S23, if Ph is greater than or equal to the first preset threshold value P1, the cell is identified as a human cell, if Pm is greater than or equal to the first preset threshold value P1, the cell is identified as a mouse cell, and the rest are double cells.

3. The method for analyzing single-cell transcriptome data of a PDX model according to claim 1, characterized in that: In step S3, extracting sequences identified as human cells and sequences identified as mouse cells specifically includes: S31, identifying the barcode of the sequencing sequence, comparing it with the barcode of the cell, and obtaining the base matching coefficient Mi, Mi = Lm / Lb, where Lm is the number of bases that match the barcode of the sequencing sequence and the barcode of the cell, and Lb is the number of bases of the barcode of the cell; S32, extracting a sequence according to Mi and a second preset threshold value P2: If Mi = 100%, directly extract the corresponding sequence; when P2 ≤ Mi < 100% and the sequencing quality value of the base corresponding to the sequencing reads that are not fully matched < 10, correct the sequencing reads to the correct base and then extract the sequence; if Mi < P2, do not extract. Among them, P2 ≥ 80%.

4. The method for analyzing single-cell transcriptome data of a PDX model according to claim 3, characterized in that: Correcting the sequencing reads to the correct base means correcting the base with a sequencing quality value < 10 in the barcode of the sequencing sequence to the base at the corresponding position of the barcode of the matching cell.

5. The method for analyzing single-cell transcriptome data of a PDX model according to claim 1, characterized in that: In steps S5 and S6, it further includes a step of filtering the alignment results: If it aligns to a unique position of an integrated gene, retain the corresponding alignment information and extract one alignment information; if it aligns to multiple positions of the same integrated gene, only retain one of the alignment information; if it aligns to different integrated genes, filter the corresponding alignment information.

6. A computer device, characterized in that: It includes: A memory for storing computer programs; A processor for implementing the steps of an analysis method for single-cell transcriptome data of a PDX model as described in any one of claims 1 - 5 when executing the computer program.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the steps of an analysis method for single-cell transcriptome data of a PDX model as described in any one of claims 1 - 5.

Images