Predictive markers for novel coronavirus susceptibility and prediction methods and devices

Abstract

The present invention discloses a predictive marker for novel coronavirus susceptibility and a prediction method and device, which are used to solve the problems of unstable biomarker discovery, large risk prediction limitations and low feasibility in existing related methods. The method includes: first, providing a predictive marker for novel coronavirus susceptibility, wherein the predictive marker is a differentially expressed gene set extracted based on a human blood sample; second, providing a corresponding method for predicting novel coronavirus susceptibility, obtaining a blood sample to be tested, and extracting a PBMC sample to be tested from the blood sample to be tested; extracting the gene expression data of the predictive marker in the PBMC sample to be tested from the gene expression spectrum; inputting the gene expression data to be tested into a pre-trained novel coronavirus susceptibility prediction model for susceptibility risk judgment, and obtaining the novel coronavirus susceptibility risk assessment result of the human body corresponding to the blood sample to be tested, thereby simply, quickly and accurately predicting the human body's novel coronavirus susceptibility risk.

Description

Technical Field

[0001] The present invention relates to the technical field of virus infection research, and in particular to a predictive marker for novel coronavirus susceptibility, a method for predicting novel coronavirus susceptibility, a device for predicting novel coronavirus susceptibility, an electronic device and a storage medium. Background Art

[0002] The new coronavirus (hereinafter referred to as COVID-19) is a highly contagious virus.

[0003] Therefore, identifying and paying attention to susceptible populations, studying susceptibility, and increasing non-drug intervention measures in advance will help to scientifically and rationally allocate medical resources, increase treatment effects, and improve treatment rates. Studying susceptibility can also guide vaccination strategies. Vaccines are one of the important means to control the epidemic. Understanding the susceptibility of the population can help determine vaccination priorities and strategies to minimize the risk of infection. It can also help evaluate the effectiveness and protective effects of vaccines, and provide a scientific basis for subsequent vaccine development and improvement.

[0004] The current research on susceptibility to the new coronavirus mainly uses prediction methods based on susceptibility genes, biomarker discovery methods based on metabolomics data, or TCR (T-Cell Receptor) / BCR (B-Cell Receptor) sequencing methods.

[0005] However, although the susceptibility gene prediction method can discover susceptibility genes, it can only reveal some genetic factors and cannot fully cover all factors affecting susceptibility, which lacks comprehensiveness. Moreover, this method is based on static genotype data for prediction, which cannot reflect the dynamic changes of individuals at different time points and environments. It is easily affected by the interaction between genetics and the environment and lacks dynamism.

[0006] When using biomarker discovery methods based on metabolomics data for prediction, it is easily affected by factors such as individual metabolic status and dietary habits, and sample heterogeneity may also lead to unstable biomarker discovery. In addition, although metabolomics data can provide biomarkers, they often lack in-depth explanations of the biological mechanisms behind them, limiting the understanding of susceptibility to COVID-19 infection.

[0007] The TCR / BCR sequencing method is limited to the immune system and cannot fully consider other factors that affect the risk of COVID-19 infection. Moreover, this method often requires complex experimental operations and data processing, which invisibly increases the experimental cost and time cost, limiting its feasibility in large-scale applications. Summary of the invention

[0008] The present invention provides a predictive marker for novel coronavirus susceptibility, a predictive method for novel coronavirus susceptibility, a predictive device for novel coronavirus susceptibility, an electronic device and a storage medium, which are used to solve or partially solve the technical problems existing in the existing novel coronavirus susceptibility risk prediction methods, such as unstable biomarker discovery, large limitations in risk prediction and low feasibility.

[0009] The present invention provides a predictive marker for susceptibility to the new coronavirus, wherein the predictive marker is a differentially expressed gene set extracted from human blood samples, and the predictive marker is used to predict a human's risk of susceptibility to the new coronavirus.

[0010] Optionally, the differentially expressed gene set includes up-regulated genes and down-regulated genes, the up-regulated genes include ZSWIM3, EGR2, EGR3, FFAR2, LIMA1 and APOL1, and the down-regulated genes include TNFRSF21, CDYL, RELL1, USP3, CTSK and FAM156B.

[0011] The present invention also provides a method for predicting susceptibility to the novel coronavirus, wherein the method uses the prediction markers as described above to predict the risk of susceptibility to the novel coronavirus in a human body, and the method comprises:

[0012] Obtaining a blood sample to be tested, and extracting a PBMC sample to be tested from the blood sample to be tested;

[0013] Extracting the gene expression data of the prediction marker in the PBMC sample to be tested from the gene expression profile;

[0014] The gene expression data to be tested is input into a pre-trained COVID-19 susceptibility prediction model to judge the susceptibility risk, and the COVID-19 susceptibility risk assessment result of the human body corresponding to the blood sample to be tested is obtained.

[0015] Optionally, the construction and training process of the COVID-19 susceptibility prediction model includes:

[0016] Obtain a pre-constructed COVID-19 susceptibility risk dataset, and randomly divide the COVID-19 susceptibility risk dataset into a COVID-19 susceptibility training set and a COVID-19 susceptibility test set according to a preset division ratio;

[0017] Build several different types of classification prediction models;

[0018] Based on the COVID-19 susceptible training set, each of the classification prediction models is trained for susceptible risk judgment, and parameters of each of the classification prediction models are adjusted and optimized during the training process;

[0019] The classification prediction model with the highest prediction accuracy obtained after training is used as the COVID-19 susceptibility prediction model;

[0020] The new coronavirus susceptibility test set is used to evaluate the model prediction effect of the new coronavirus susceptibility prediction model, obtain multiple model evaluation indicators, draw the ROC curve corresponding to the new coronavirus susceptibility risk data set, and calculate the feature importance.

[0021] Optionally, the process of constructing the COVID-19 susceptibility risk dataset includes:

[0022] Obtain blood sample data from the trial population and their symptoms during the COVID-19 pandemic;

[0023] Perform multi-omics sequencing analysis based on the blood sample data to obtain antibody level data;

[0024] According to the antibody level data and the symptom conditions, the test population is divided into an uninfected group, an asymptomatic group, and a symptomatic group;

[0025] Extracting a test PBMC sample of each test blood sample in the blood sample data, and performing sequencing processing on each of the test PBMC samples to obtain transcriptome sequencing data;

[0026] Aligning each of the transcriptome sequencing data to a reference genome to obtain gene alignment information, and converting the gene alignment information into an original gene expression matrix;

[0027] Filtering the original gene expression matrix according to a preset filtering condition to obtain a filtered gene expression matrix;

[0028] Screening protein-coding genes from the filtered gene expression matrix, removing duplicate genes, obtaining a target gene expression matrix, and performing standardization on the target gene expression matrix to obtain a TPM value, wherein the TPM value is used to compare gene expression levels between different experimental PBMC samples;

[0029] Performing statistical analysis on the target gene expression matrix, taking the symptomatic group as a first susceptible group, and taking the asymptomatic group and the uninfected group as a second susceptible group;

[0030] Perform gene differential expression analysis on the first susceptible group and the second susceptible group to screen out differentially expressed gene sets that meet the preset significant expression change conditions as predictive markers for human susceptibility to COVID-19;

[0031] The test gene expression data of the predictive marker in each of the test PBMC samples are extracted from the gene expression spectrum, and the test gene expression data of each are integrated into a new coronavirus susceptibility risk data set.

[0032] Optionally, the test population includes multiple test individuals, the symptom condition is asymptomatic or symptomatic, and the test population is divided into an uninfected group, an asymptomatic group, and a symptomatic group according to the antibody level data and the symptom condition, including:

[0033] Test subjects with IgG antibody levels ≤100 (AU / mL) were classified into the uninfected group;

[0034] The test subjects with IgG antibody level >100 (AU / mL) and asymptomatic condition were classified as asymptomatic group;

[0035] The test subjects with IgG antibody levels >100 (AU / mL) and symptomatic symptoms were classified as the symptomatic group.

[0036] Optionally, performing gene differential expression analysis on the first susceptible group and the second susceptible group to screen out differentially expressed gene sets that meet preset significant expression change conditions as predictive markers for human susceptibility to COVID-19 includes:

[0037] The first susceptible group and the second susceptible group were subjected to gene differential expression analysis, and padj≤0.05 and log2FC>0 were used as the significant up-regulation criteria of genes to screen out up-regulated genes, which included ZSWIM3, EGR2, EGR3, FFAR2, LIMA1 and APOL1;

[0038] Using log2FC<0 as the significant down-regulation standard of genes, down-regulated genes were screened out, including TNFRSF21, CDYL, RELL1, USP3, CTSK and FAM156B;

[0039] The differentially expressed gene set consisting of 12 genes, including ZSWIM3, EGR2, EGR3, FFAR2, LIMA1, APOL1, TNFRSF21, CDYL, RELL1, USP3, CTSK and FAM156B, will be used as a predictive marker for human susceptibility to COVID-19.

[0040] The present invention also provides a device for predicting novel coronavirus susceptibility, which is applied to a method for predicting novel coronavirus susceptibility, wherein the method uses the above-mentioned prediction markers to predict the novel coronavirus susceptibility risk of a human body, and the device comprises:

[0041] A PBMC sample extraction module to be tested, used to obtain a blood sample to be tested, and extract a PBMC sample to be tested from the blood sample to be tested;

[0042] A gene expression data extraction module to be tested, used to extract the gene expression data to be tested of the prediction marker in the PBMC sample to be tested from the gene expression spectrum;

[0043] The COVID-19 susceptibility risk judgment module is used to input the gene expression data to be tested into a pre-trained COVID-19 susceptibility prediction model to perform susceptibility risk judgment, and obtain the COVID-19 susceptibility risk assessment result of the human body corresponding to the blood sample to be tested.

[0044] The present invention also provides an electronic device, the device comprising a processor and a memory:

[0045] The memory is used to store program code and transmit the program code to the processor;

[0046] The processor is used to execute the method for predicting susceptibility to the new coronavirus as described in any of the above items according to the instructions in the program code.

[0047] The present invention also provides a computer-readable storage medium, which is used to store program code, and the program code is used to execute the method for predicting susceptibility to the new coronavirus as described in any of the above items.

[0048] It can be seen from the above technical solutions that the present invention has the following advantages:

[0049] In view of the deficiencies in the prior art, a predictive marker for virus susceptibility and a corresponding prediction method are proposed. First, a predictive marker for novel coronavirus susceptibility is provided, and the predictive marker is a differentially expressed gene set extracted based on a human blood sample. Secondly, a corresponding method for predicting novel coronavirus susceptibility is provided: obtaining a blood sample to be tested, and extracting a PBMC sample to be tested from the blood sample to be tested; extracting the gene expression data of the predictive marker in the PBMC sample to be tested from the gene expression spectrum; inputting the gene expression data to be tested into the pre-trained novel coronavirus susceptibility prediction model for susceptibility risk judgment, and obtaining the novel coronavirus susceptibility risk assessment result of the human body corresponding to the blood sample to be tested. Therefore, by using the predictive markers located in the human blood sample, combined with the corresponding novel coronavirus susceptibility prediction method, the human body's novel coronavirus susceptibility risk can be simply, quickly and accurately predicted. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0051] Figure 1 This is a schematic diagram of the ROC curve drawn based on the evaluation results of the COVID-19 susceptibility prediction model;

[0052] Figure 2 This is a schematic diagram of the construction and training process of a COVID-19 susceptibility prediction model;

[0053] Figure 3 A flowchart of the steps of a method for predicting susceptibility to the new coronavirus;

[0054] Figure 4 This is a structural block diagram of a device for predicting susceptibility to the new coronavirus. DETAILED DESCRIPTION

[0055] The embodiments of the present invention provide a predictive marker for novel coronavirus susceptibility, a method for predicting novel coronavirus susceptibility, a device for predicting novel coronavirus susceptibility, an electronic device and a storage medium, which are used to solve or partially solve the technical problems existing in the existing novel coronavirus susceptibility risk prediction methods, such as unstable biomarker discovery, large limitations in risk prediction and low feasibility.

[0056] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] As an example, in the field of virus research, omics is a discipline that uses a variety of high-throughput technologies such as genomics, transcriptomics, proteomics, metabolomics and intestinal microbiome to study the molecular components and interactions in organisms. In the study of COVID-19 susceptibility, omics technology can play an important role and provide more comprehensive and in-depth information at the molecular level. In the study of predicting COVID-19 susceptibility and whether symptoms appear after infection, researchers usually collect a large amount of blood sample data, including genomic data, gene expression data, protein expression data, metabolite data, etc. These data are then analyzed using machine learning algorithms to determine biomarkers associated with COVID-19 infection. These biomarkers can be susceptibility genes, expression levels of specific genes, expression levels of specific proteins, or the content of specific metabolites.

[0058] Based on this, the study of susceptibility to the new coronavirus currently mainly adopts prediction methods based on susceptibility genes, biomarker discovery methods based on metabolomics data, or TCR / BCR sequencing methods.

[0059] However, although the susceptibility gene prediction method can discover susceptibility genes, it can only reveal some genetic factors and cannot fully cover all factors affecting susceptibility, which lacks comprehensiveness. Moreover, this method is based on static genotype data for prediction, which cannot reflect the dynamic changes of individuals at different time points and environments. It is easily affected by the interaction between genetics and the environment and lacks dynamism.

[0060] When using biomarker discovery methods based on metabolomics data for prediction, it is easily affected by factors such as individual metabolic status and dietary habits, and sample heterogeneity may also lead to unstable biomarker discovery. In addition, although metabolomics data can provide biomarkers, they often lack in-depth explanations of the biological mechanisms behind them, limiting the understanding of susceptibility to COVID-19 infection.

[0061] The TCR / BCR sequencing method is limited to the immune system and cannot fully consider other factors that affect the risk of COVID-19 infection. Moreover, this method often requires complex experimental operations and data processing, which invisibly increases the experimental cost and time cost, limiting its feasibility in large-scale applications.

[0062] Therefore, one of the core inventive points of the embodiment of the present invention is to construct a transcriptome prediction model in the population to predict the susceptibility to the new coronavirus. By analyzing the gene expression data of the population, the differences in host susceptibility and biological mechanisms under the new coronavirus infection environment are revealed, providing more accurate information for the prediction and intervention of individual susceptibility, while better reflecting the genetic background and immune status of the population. First, a predictive marker for new coronavirus susceptibility is provided, and the predictive marker is a differentially expressed gene set extracted based on a human blood sample. Secondly, a corresponding method for predicting susceptibility to the new coronavirus is provided, obtaining a blood sample to be tested, and extracting a PBMC sample to be tested from the blood sample to be tested; extracting the gene expression data of the predictive marker in the PBMC sample to be tested from the gene expression spectrum; inputting the gene expression data to be tested into the pre-trained new coronavirus susceptibility prediction model for susceptibility risk judgment, and obtaining the new coronavirus susceptibility risk assessment result of the human body corresponding to the blood sample to be tested, thereby simply, quickly and accurately predicting the human body's new coronavirus susceptibility risk. By adopting the technical solution of the present invention, firstly, by conducting research under the condition of "natural" control of virus and environmental factor variables during the pandemic, the differences in host susceptibility of the population can be better analyzed, the interference of external environmental factors can be eliminated, and the accuracy and reliability of the prediction results can be improved; secondly, the transcriptome can detect the expression levels of almost all genes at the same time, and can detect slight changes in gene expression, which can provide a more comprehensive understanding of the gene expression of cells or tissues, thereby further improving the comprehensiveness and accuracy of the prediction, and it also has a high sensitivity for discovering potential genes related to susceptibility; in addition, the blood transcriptome used in the present invention requires blood samples, and compared with proteomics and metabolomics that require complex sample processing and analysis processes, the samples of the present invention are easy to obtain and process, which is more convenient. Finally, the application of the new crown susceptibility prediction model provided by the present invention may involve fields such as medical diagnosis and personalized treatment, and is expected to become an important technology and product in the future medical and health field.

[0063] Based on the above examples and core invention points, the present invention mainly includes two aspects: first, providing a risk marker for predicting susceptibility to the new coronavirus; second, providing a prediction method for accurately predicting the risk of susceptibility to the new coronavirus through risk markers.

[0064] Regarding the first aspect, in a specific implementation, an embodiment of the present invention provides a predictive marker for susceptibility to the new coronavirus. The predictive marker is a differentially expressed genome extracted from a human blood sample. The predictive marker is used to predict the risk of susceptibility to the new coronavirus in the human body.

[0065] Among them, the differentially expressed gene set referred to in the embodiment of the present invention includes up-regulated genes consisting of 6 genes and down-regulated genes consisting of 6 genes. The up-regulated genes include ZSWIM3, EGR2, EGR3, FFAR2, LIMA1 and APOL1. The down-regulated genes include TNFRSF21, CDYL, RELL1, USP3, CTSK and FAM156B.

[0066] The predictive markers for novel coronavirus susceptibility provided in the embodiments of the present invention will be determined and verified by machine learning algorithms based on novel coronavirus-related experimental data in the following embodiments.

[0067] Regarding the second aspect, in the specific implementation, we first need to build a new coronavirus susceptibility prediction model for judging the risk of new coronavirus susceptibility.

[0068] Furthermore, in combination with specific experiments, the construction and training process of the COVID-19 susceptibility prediction model can be achieved by executing the following sub-steps S01 to S05:

[0069] Step S01: Obtain a pre-constructed COVID-19 susceptible risk dataset, and randomly divide the COVID-19 susceptible risk dataset into a COVID-19 susceptible training set and a COVID-19 susceptible test set according to a preset division ratio;

[0070] Furthermore, the COVID-19 susceptibility risk dataset provided in the embodiment of the present invention can be constructed by executing the following sub-steps S01-1 to S01-10:

[0071] Step S01-1: Obtain blood sample data of the test population, as well as the symptoms of the test population during the COVID-19 pandemic;

[0072] The present invention includes 173 volunteers as the test population for the novel coronavirus susceptibility experiment by setting the following standards, and collects blood sample data from the test population, wherein each volunteer can be regarded as a test individual:

[0073] (1) Good health. In particular, no history of underlying diseases such as respiratory diseases (such as asthma, chronic bronchitis, chronic rhinitis, etc.), malignant tumors, cardiovascular and cerebrovascular diseases, metabolic diseases, blood diseases, autoimmune diseases or immunodeficiency diseases. No history of infectious diseases such as HIV (Human Immunodeficiency Virus), hepatitis B and hepatitis C. Also, no special conditions such as pregnancy or lactation. No history of long-term medication.

[0074] (2) No smoking history.

[0075] (3) Has not been vaccinated with relevant vaccines (except COVID-19 vaccines) and has not been diagnosed with infectious diseases (except COVID-19 infection) in the past three years.

[0076] In addition, the present invention also collected information from these 173 volunteers through a questionnaire survey to determine whether they had experienced cold-like symptoms during the COVID-19 pandemic, including fever, fatigue, loss of smell / taste, dry cough, nasal congestion, runny nose, sore throat, muscle aches, diarrhea, conjunctivitis, etc. Based on the collected results, the specific symptoms can be asymptomatic or symptomatic.

[0077] It should be noted that all test data sources used in the relevant experiments in the embodiments of the present invention are legal.

[0078] Step S01-2: Perform multi-omics sequencing analysis based on the blood sample data to obtain antibody level data;

[0079] Multi-omics sequencing analysis of the blood sample data collected from volunteers can obtain antibody level data.

[0080] Step S01-3: Divide the test population into an uninfected group, an asymptomatic group, and a symptomatic group based on antibody level data and symptom conditions;

[0081] In addition, the present invention also collected the vaccination status and interval time of volunteers through questionnaire surveys. By consulting the literature, the law of changes in the general antibody levels of the human body after vaccination and infection was understood. At the same time, the antibody levels of the volunteers were counted. Based on the above experimental preparation, the IgG antibody level = 100 (AU / mL) was set as the judgment threshold to determine the new crown infected and uninfected. The infected can be further subdivided into symptomatic and asymptomatic infected according to the symptoms during the pandemic (symptomatic / asymptomatic).

[0082] In the specific implementation, according to the antibody level data and symptoms, the test population is divided into an uninfected group, an asymptomatic group, and a symptomatic group, which can be divided into the following three situations:

[0083] In the first case, the test subjects with IgG antibody levels ≤ 100 (AU / mL) were classified into the uninfected group;

[0084] The second scenario: the test subjects with IgG antibody levels >100 (AU / mL) and asymptomatic symptoms (i.e., asymptomatic during the COVID-19 pandemic) were classified as the asymptomatic group;

[0085] The third situation: test subjects with IgG antibody levels >100 (AU / mL) and symptomatic symptoms (i.e., symptoms during the COVID-19 pandemic) are classified as the symptomatic group.

[0086] Since the general antibody level of symptomatic infected people is much higher than 100 (AU / mL), the binary classification used in the embodiment of the present invention is less affected by the antibody threshold. And the physical fitness of the included people is relatively similar. By counting the antibody levels of volunteers in different situations, it also provides guidance for guiding more reasonable and scientific classification, reducing the impact of individual differences or physical fitness on grouping.

[0087] Step S01-4: extracting a test PBMC sample of each test blood sample in the blood sample data, and performing sequencing processing on each test PBMC sample to obtain transcriptome sequencing data;

[0088] Among them, the experimental PBMC (Peripheral Blood Mononuclear Cells) samples refer to a type of cell population with a single nucleus separated from human peripheral blood. They are an important component of the immune system and participate in the body's immune response and defense mechanism. PBMC can be separated and extracted from blood samples by density gradient centrifugation. The main principle is to use lymphocyte separation fluid, such as Ficoll (a commonly used density gradient centrifugation medium), to remove polymorphonuclear cells such as red blood cells and granulocytes in the blood, thereby obtaining a relatively pure mononuclear cell population.

[0089] After the PBMC samples are extracted from the blood samples of the volunteers, they can be sequenced to obtain the pre-processed transcriptome sequencing data. At the same time, transcriptome quality control (such as FastQC version 0.12.1) can be used to evaluate the quality of sequence reads in transcriptome sequencing data to ensure data quality and comparability.

[0090] Step S01-5: aligning each transcriptome sequencing data to the reference genome, obtaining gene alignment information, and converting the gene alignment information into an original gene expression matrix;

[0091] Next, each transcriptome sequencing data is aligned to the reference genome using HiSat (Hierarchical Indexing for SplicedAlignment of Transcripts, a sequence alignment tool), such as HiSat22.2.1, to obtain gene alignment information such as the position and matching quality of reads on the aligned genome or other reference sequences. The gene × sample original gene expression matrix (hereinafter referred to as the original gene expression matrix) is obtained through the featurecount command in subread2.0.6 (a tool for quantitative expression calculation and expression matrix generation).

[0092] In this step, by comparing the measured RNA transcriptome sequencing data with the reference genome, it is possible to identify which genes are expressed under specific conditions and further understand the expression levels and differential expression of genes.

[0093] Step S01-6: filtering the original gene expression matrix according to a preset filtering condition to obtain a filtered gene expression matrix;

[0094] The low-expression genes in the original gene expression matrix were filtered out using the expression level of 0 in more than 50% of the samples as the filtering condition to obtain the filtered gene expression matrix.

[0095] Step S01-7: Screening protein-coding genes from the filtered gene expression matrix, removing duplicate genes, obtaining a target gene expression matrix, and performing standardization on the target gene expression matrix to obtain a TPM value, which is used to compare gene expression levels between different experimental PBMC samples;

[0096] Then, protein-coding genes were screened from the filtered gene expression matrix, and duplicate genes were removed to obtain a target gene expression matrix of 15,580 genes in 173 experimental PBMC samples.

[0097] The target gene expression matrix was standardized to obtain the TPM (Transcripts Per Million, the number of transcripts per million RNA molecules) value. The calculation formula of the TPM value is as follows:

[0098]

[0099] Among them, Counts of gene / transcript represents the number of measurements of a gene or transcript (counts); Effective transcript length represents the effective transcript length, which is usually the actual length of the gene minus the length of the untranslated region (UTR); Σ represents the sum symbol, which is used to calculate the sum of the measured values ​​of all genes or transcripts.

[0100] In practical applications, TPM is particularly suitable for: (1) comparing the relative expression levels of different genes or transcripts; and (2) comparing gene expression levels between samples, especially when taking into account the measurement depth and differences between samples.

[0101] Step S01-8: Statistical analysis is performed on the target gene expression matrix, and the symptomatic group is used as the first susceptible group, and the asymptomatic group and the uninfected group are used as the second susceptible group;

[0102] The target gene expression matrix obtained after pretreatment was used for statistical analysis, and the symptomatic group was taken as the first susceptible group (i.e., high susceptible group), totaling 70 people. The asymptomatic group and the uninfected group were taken as the second susceptible group (i.e., low susceptible group), totaling 103 people.

[0103] Step S01-9: Perform gene differential expression analysis on the first susceptible group and the second susceptible group, and screen out differentially expressed gene sets that meet the preset significant expression change conditions as predictive markers for human susceptibility to COVID-19;

[0104] In the embodiment of the present invention, DEseq2 software (a gene expression differential analysis toolkit) is used to perform differential expression analysis. DEseq2, a conventional differential expression analysis software, is used for analysis, and the type and quantity of biomarkers obtained can be affected by controlling the threshold.

[0105] Further, the implementation process of sub-step S01-9 may specifically be:

[0106] Gene differential expression analysis was performed on the first susceptible group and the second susceptible group, and padj≤0.05 and log2FC>0 were used as the significant up-regulation criteria for genes to screen out up-regulated genes, including ZSWIM3, EGR2, EGR3, FFAR2, LIMA1, and APOL1;

[0107] Using log2FC<0 as the significant down-regulation standard of genes, down-regulated genes were screened out, including TNFRSF21, CDYL, RELL1, USP3, CTSK and FAM156B;

[0108] The differentially expressed gene set consisting of 12 genes, including ZSWIM3, EGR2, EGR3, FFAR2, LIMA1, APOL1, TNFRSF21, CDYL, RELL1, USP3, CTSK and FAM156B, will be used as a predictive marker for human susceptibility to COVID-19.

[0109] FC (Fold Change) indicates the fold difference of gene expression; the P value is used to determine whether the gene difference is significant; and the padj value is used to determine whether the P value is credible.

[0110] Based on the above biomarker screening principles, those skilled in the art can also screen out more representative features by adjusting the differential expression analysis threshold, using other differential expression analysis software, or adopting a joint weighted gene co-expression network method.

[0111] For the adjustment of the threshold of differential expression analysis, you can try to adjust the threshold, such as adjusting log2FC and adjusting P value, to screen out more representative features.

[0112] In addition to DEseq2 software, other differential expression analysis software can also be used, such as limma or edgeR. Among them, limma is suitable for differential expression analysis of small sample sizes and has high accuracy and robustness. EdgeR is suitable for differential expression analysis of RNA-seq data and can take into account technical differences between samples. These software may have different algorithms and parameter settings. Those skilled in the art can select appropriate software for analysis based on actual conditions and specific data conditions.

[0113] A joint weighted gene co-expression network can be constructed to use the relationship information between genes for feature screening. This method can take into account the interactions between genes and help discover potential functional modules.

[0114] Therefore, by optimizing statistical methods or combining different high-throughput data toolkits, the accuracy and reliability of differential expression analysis can be improved, thereby discovering biomarkers associated with specific biological processes or diseases.

[0115] Step S01-10: Extract the test gene expression data of predictive markers in each test PBMC sample from the gene expression spectrum, and integrate the test gene expression data into a new coronavirus susceptibility risk dataset.

[0116] The gene expression data of 12 genes, including ZSWIM3, EGR2, EGR3, FFAR2, LIMA1, APOL1 and TNFRSF21, CDYL, RELL1, USP3, CTSK, and FAM156B, in each experimental PBMC sample (i.e. all experimental PBMC samples) were extracted from the gene expression spectrum, and these gene expression data were integrated into a new coronavirus susceptibility risk dataset.

[0117] After constructing the COVID-19 susceptibility risk dataset, you can use train_test_split to randomly divide the COVID-19 susceptibility risk dataset into a COVID-19 susceptibility training set (80%) and a COVID-19 susceptibility test set (20%).

[0118] Step S02: constructing several different types of classification prediction models;

[0119] Based on the sklearn package (a machine learning library) in Python 3, different machine learning models are used to build classification prediction models. For example, different types of classification prediction models can be built, such as Random Forest Classifier (RFC), Support Vector Machine (SVM), eXtreme Gradient Boosting (XGboost), Naive Bayesian Model (NBM), etc.

[0120] Step S03: Based on the COVID-19 susceptible training set, each classification prediction model is trained for susceptible risk judgment, and parameters of each classification prediction model are adjusted and optimized during the training process;

[0121] For example, in the model training process, the feature selection method in machine learning can be implemented mainly in the following ways:

[0122] SelectKBest (a feature selection function): Selects the K best features, which can be selected based on statistical tests or information theory methods.

[0123] Recursive Feature Elimination: Select the best subset of features by recursively training the model and eliminating features that have less influence on the model.

[0124] SelectFromModel (a method of feature selection through models): Select the most important features through supervised learning models. For example, feature selection methods based on tree models. These methods can further select features based on differential expression analysis to improve the generalization ability and interpretability of the model.

[0125] Step S04: The classification prediction model with the highest prediction accuracy obtained after training is used as the COVID-19 susceptibility prediction model;

[0126] In addition to the above-mentioned method of selecting the best model from multiple models, you can also directly select a classification prediction model as the basic model based on empirical knowledge, then train the basic model by combining the adjustment parameter method, and finally select the model with the highest prediction accuracy from the models with different parameter settings as the final target model.

[0127] For example, you can choose to use the random forest classification model RFC to build a random forest classification prediction model on the training set. The model uses 5-fold cross validation, and optimizes the model adjustment parameters through optuna (an open source hyperparameter optimization framework) to find the model with the highest prediction accuracy. In this example, the model setting parameters are: n_estimators: 57, max_depth: 10, min_samples_split: 19, min_samples_leaf: 6.

[0128] Similarly, you can also choose other classification prediction models such as support vector machine SVM, extreme gradient boosting model XGboost or naive Bayes model NBM as the basic model according to actual application needs, adjust and optimize the model parameters according to the above method, and select the final target model.

[0129] Step S05: Use the new coronavirus susceptibility test set to evaluate the model prediction effect of the new coronavirus susceptibility prediction model, obtain multiple model evaluation indicators, draw the ROC curve corresponding to the new coronavirus susceptibility risk data set, and calculate the feature importance.

[0130] The test set is used to evaluate the model effect, and multiple model evaluation indicators such as accuracy, precision, recall, F1 score (F1 Score, an indicator used to measure the accuracy of binary classification models in statistics) and AUC (Area Under Curve, the area under the ROC (Receiver Operating Characteristic curve) curve and the coordinate axis) are obtained. Based on AUC, the ROC curve corresponding to the new crown susceptibility risk data set can be drawn, which can reflect the results of this model training. At the same time, the feature importance can be calculated based on the model output results.

[0131] In order to enable those skilled in the art to better understand the technical solution of the present invention, four classification prediction models, namely, random forest classification model RFC, support vector machine SVM, extreme gradient boosting model XGboost and naive Bayes model NBM, are used as examples for model training verification, and the specific model effect evaluation comparison results are shown in Table 1 below:

[0132]

[0133] Table 1: Comparison of model effect evaluation of different COVID-19 susceptibility prediction models

[0134] Combining the final evaluation results of multiple classification prediction models, it can be seen that among the multiple new crown risk prediction models constructed based on the biomarkers ZSWIM3, EGR2, EGR3, FFAR2, LIMA1, APOL1 and TNFRSF21, CDYL, RELL1, USP3, CTSK, and FAM156B of the present invention, the random forest classification model RFC has the best comprehensive evaluation effect, with an AUC of 0.86, an accuracy of 0.80, a precision of 0.87, a recall rate of 0.72, and an F1 score of 0.78, with high accuracy and high sensitivity.

[0135] Furthermore, taking the above-mentioned random forest classification model RFC as an example, the ROC curve diagram drawn based on the evaluation results of the new crown susceptibility prediction model is as follows: Figure 1 As shown. The horizontal axis is the false positive rate (False Positive Rate), and the vertical axis is the true positive rate (True Positive Rate, also known as sensitivity). It can be seen that based on the biomarkers for predicting the risk of COVID-19 susceptibility provided by the present invention, combined with the COVID-19 susceptibility prediction method proposed by the present invention, a prediction result with higher accuracy and higher sensitivity can be obtained, thereby verifying the feasibility of the prediction method provided by the embodiment of the present invention.

[0136] To better explain, Figure 2 A schematic diagram of the construction and training process of a COVID-19 susceptibility prediction model provided by an embodiment of the present invention is shown.

[0137] Combination Figure 2 The general process of constructing and training the COVID-19 susceptibility prediction model proposed in the embodiment of the present invention is as follows:

[0138] (1) Blood sample collection and grouping

[0139] Obtain blood sample data of the trial population, as well as the symptoms of the trial population during the COVID-19 pandemic; perform multi-omics sequencing analysis based on the blood sample data to obtain antibody level data; and divide the trial population into an uninfected group, an asymptomatic group, and a symptomatic group based on the antibody level data and symptoms.

[0140] (2) Transcriptome sequencing and upstream analysis

[0141] The experimental PBMC sample of each experimental blood sample in the blood sample data is extracted, and each experimental PBMC sample is sequenced to obtain transcriptome sequencing data.

[0142] (3) Gene expression matrix and preprocessing

[0143] Each transcriptome sequencing data is aligned to the reference genome to obtain gene alignment information, and the gene alignment information is converted into the original gene expression matrix; the original gene expression matrix is ​​filtered according to the preset filtering conditions to obtain the filtered gene expression matrix; protein-coding genes are screened out from the filtered gene expression matrix, duplicate genes are removed, and the target gene expression matrix is ​​obtained, and the TPM value is obtained after standardization of the target gene expression matrix.

[0144] (4) Differential expression analysis

[0145] Statistical analysis was performed on the target gene expression matrix, and the symptomatic group was taken as the first susceptible group, and the asymptomatic group and the uninfected group as the second susceptible group; gene differential expression analysis was performed on the first susceptible group and the second susceptible group, and the differentially expressed gene groups that met the preset significant expression change conditions were screened out as predictive markers for human susceptibility to the new coronavirus.

[0146] (5) Selecting feature gene sets

[0147] The experimental gene expression data of predictive markers in each experimental PBMC sample were extracted from the gene expression spectrum, and the expression data of each experimental gene were integrated into a new coronavirus susceptibility risk dataset.

[0148] (6) Divide the training set into the test set

[0149] According to the preset division ratio (such as 8:2), the COVID-19 susceptibility risk dataset is randomly divided into a COVID-19 susceptible training set and a COVID-19 susceptible test set.

[0150] (7) Machine learning model construction and selection

[0151] Construct several different types of classification prediction models; based on the new coronavirus susceptibility training set, conduct susceptibility risk judgment training for each classification prediction model.

[0152] (8) Optuna adjustment parameters

[0153] During the training process, the parameters of each classification prediction model are adjusted and optimized; the classification prediction model with the highest prediction accuracy obtained after training is used as the new coronavirus susceptibility prediction model.

[0154] (9) Susceptibility prediction and effect evaluation

[0155] The COVID-19 susceptibility test set was used to evaluate the model prediction effect of the COVID-19 susceptibility prediction model, obtain multiple model evaluation indicators, draw the ROC curve corresponding to the COVID-19 susceptibility risk data set, and calculate the feature importance.

[0156] Based on the COVID-19 susceptibility prediction model constructed in the aforementioned embodiment, refer to Figure 3, showing a flowchart of the steps of a method for predicting novel coronavirus susceptibility provided by an embodiment of the present invention. The method uses the prediction markers provided in the aforementioned embodiments to predict the risk of novel coronavirus susceptibility in the human body, and the method may specifically include the following steps:

[0157] Step 301, obtaining a blood sample to be tested, and extracting a PBMC sample to be tested from the blood sample to be tested;

[0158] The method of extracting PBMC samples can be performed with reference to the above-mentioned embodiment, and this step will not be repeated here.

[0159] Step 302, extracting the gene expression data of the prediction marker in the PBMC sample to be tested from the gene expression profile;

[0160] Then, the gene expression data of 12 genes, namely, ZSWIM3, EGR2, EGR3, FFAR2, LIMA1, APOL1, TNFRSF21, CDYL, RELL1, USP3, CTSK, and FAM156B, in the PBMC samples to be tested can be extracted from the gene expression profile.

[0161] Step 303: Input the gene expression data to be tested into a pre-trained COVID-19 susceptibility prediction model to judge the susceptibility risk, and obtain the COVID-19 susceptibility risk assessment result of the human body corresponding to the blood sample to be tested.

[0162] According to the new coronavirus susceptibility prediction model constructed and trained in the above-mentioned embodiment, based on the prediction markers including 12 genes, it is possible to make a susceptibility risk judgment on the gene expression data to be tested, predict whether the human body corresponding to the blood sample to be tested belongs to the susceptible population or the non-susceptible population, and output the corresponding judgment result.

[0163] It should be noted that, in order to enable those skilled in the art to better distinguish data of the same type but with different actual meanings, some technical features are distinguished by the first and the second in the embodiments of the present invention. The first and the second are only used for data distinction and have no other special meanings. It can be understood that the present invention is not limited to this.

[0164] In an embodiment of the present invention, by analyzing the gene expression data of a population, the differences in host susceptibility and biological mechanisms under the new coronavirus infection environment are revealed, providing more accurate information for the prediction and intervention of individual susceptibility, while better reflecting the genetic background and immune status of the population. First, a predictive marker for new coronavirus susceptibility is provided, and the predictive marker is a differentially expressed gene set extracted based on a human blood sample. Secondly, a corresponding method for predicting new coronavirus susceptibility is provided, obtaining a blood sample to be tested, and extracting a PBMC sample to be tested from the blood sample to be tested; extracting the gene expression data of the predictive marker in the PBMC sample to be tested from the gene expression spectrum; inputting the gene expression data to be tested into the pre-trained new coronavirus susceptibility prediction model for susceptibility risk judgment, and obtaining the new coronavirus susceptibility risk assessment result of the human body corresponding to the blood sample to be tested, thereby simply, quickly and accurately predicting the human body's new coronavirus susceptibility risk. By adopting the technical solution of the present invention, firstly, by conducting research under the condition of "natural" control of virus and environmental factor variables during the pandemic, the differences in host susceptibility of the population can be better analyzed, the interference of external environmental factors can be eliminated, and the accuracy and reliability of the prediction results can be improved; secondly, the transcriptome can detect the expression levels of almost all genes at the same time, and can detect slight changes in gene expression, which can provide a more comprehensive understanding of the gene expression of cells or tissues, thereby further improving the comprehensiveness and accuracy of the prediction, and it also has a high sensitivity for discovering potential genes related to susceptibility; in addition, the blood transcriptome used in the present invention requires blood samples, and compared with proteomics and metabolomics that require complex sample processing and analysis processes, the samples of the present invention are easy to obtain and process, which is more convenient. Finally, the application of the new crown susceptibility prediction model provided by the present invention may involve fields such as medical diagnosis and personalized treatment, and is expected to become an important technology and product in the future medical and health field.

[0165] Reference Figure 4 , showing a block diagram of a device for predicting COVID-19 susceptibility provided by an embodiment of the present invention. The device is applied to a method for predicting COVID-19 susceptibility, and the method uses the prediction markers provided in the aforementioned embodiments to predict the COVID-19 susceptibility risk of a human body. The device may specifically include:

[0166] The PBMC sample extraction module 401 is used to obtain a blood sample to be tested and extract a PBMC sample to be tested from the blood sample to be tested;

[0167] A gene expression data extraction module 402 is used to extract the gene expression data of the prediction marker in the PBMC sample to be tested from the gene expression spectrum;

[0168] The COVID-19 susceptibility risk judgment module 403 is used to input the gene expression data to be tested into a pre-trained COVID-19 susceptibility prediction model to perform susceptibility risk judgment, and obtain the COVID-19 susceptibility risk assessment result of the human body corresponding to the blood sample to be tested.

[0169] In an optional embodiment, the device further includes a model building and training module, and the model building and training module specifically includes:

[0170] A data set partitioning module is used to obtain a pre-constructed COVID-19 susceptibility risk data set, and randomly divide the COVID-19 susceptibility risk data set into a COVID-19 susceptibility training set and a COVID-19 susceptibility test set according to a preset partitioning ratio;

[0171] A classification prediction model building module is used to build several different types of classification prediction models;

[0172] A model training module, for performing susceptibility risk judgment training on each of the classification prediction models based on the COVID-19 susceptibility training set, and adjusting and optimizing parameters of each of the classification prediction models during the training process;

[0173] A COVID-19 susceptibility prediction model determination module, which is used to use the classification prediction model with the highest prediction accuracy obtained after training as the COVID-19 susceptibility prediction model;

[0174] The model evaluation module is used to evaluate the model prediction effect of the new coronavirus susceptibility prediction model using the new coronavirus susceptibility test set, obtain multiple model evaluation indicators, draw the ROC curve corresponding to the new coronavirus susceptibility risk data set, and calculate feature importance.

[0175] In an optional embodiment, the device further includes a COVID-19 susceptible risk data set construction module, and the COVID-19 susceptible risk data set construction module specifically includes:

[0176] A blood sample data acquisition module, used to obtain blood sample data of the test population and the symptoms of the test population during the COVID-19 pandemic;

[0177] A multi-omics sequencing analysis module, used to perform multi-omics sequencing analysis based on the blood sample data to obtain antibody level data;

[0178] A test population division module, used to divide the test population into an uninfected group, an asymptomatic group and a symptomatic group according to the antibody level data and the symptom conditions;

[0179] A transcriptome sequencing data generation module, used for extracting a test PBMC sample of each test blood sample in the blood sample data, and performing sequencing processing on each of the test PBMC samples to obtain transcriptome sequencing data;

[0180] An original gene expression matrix construction module is used to align each of the transcriptome sequencing data to a reference genome, obtain gene alignment information, and convert the gene alignment information into an original gene expression matrix;

[0181] A gene filtering module, used to filter the original gene expression matrix according to a preset filtering condition to obtain a filtered gene expression matrix;

[0182] A gene expression matrix processing module, used to screen protein-coding genes from the filtered gene expression matrix, remove duplicate genes, obtain a target gene expression matrix, and obtain a TPM value after standardization of the target gene expression matrix, and the TPM value is used to compare gene expression levels between different experimental PBMC samples;

[0183] A statistical analysis module, used for performing statistical analysis on the target gene expression matrix, taking the symptomatic group as a first susceptible group, and taking the asymptomatic group and the uninfected group as a second susceptible group;

[0184] A predictive marker screening module is used to perform gene differential expression analysis on the first susceptible group and the second susceptible group, and screen out differentially expressed gene sets that meet preset significant expression change conditions as predictive markers for human susceptibility to COVID-19;

[0185] The new coronavirus susceptibility risk data set integration module is used to extract the test gene expression data of the predictive marker in each of the test PBMC samples from the gene expression spectrum, and integrate the various test gene expression data into a new coronavirus susceptibility risk data set.

[0186] In an optional embodiment, the test population includes a plurality of test individuals, the symptom condition is asymptomatic or symptomatic, and the test population division module includes:

[0187] The uninfected group classification module is used to classify the test subjects with IgG antibody levels ≤ 100 (AU / mL) into the uninfected group;

[0188] The asymptomatic group classification module is used to classify the test subjects with IgG antibody levels >100 (AU / mL) and asymptomatic symptoms into the asymptomatic group;

[0189] The symptomatic group classification module is used to classify test subjects with IgG antibody levels >100 (AU / mL) and symptomatic symptoms into the symptomatic group.

[0190] In an optional embodiment, the predictive marker screening module comprises:

[0191] An up-regulated gene screening module is used to perform gene differential expression analysis on the first susceptible group and the second susceptible group, and screen out up-regulated genes using padj≤0.05 and log2FC>0 as significant up-regulated gene standards. The up-regulated genes include ZSWIM3, EGR2, EGR3, FFAR2, LIMA1, and APOL1;

[0192] A down-regulated gene screening module is used to screen out down-regulated genes using log2FC<0 as a significant down-regulated gene standard, wherein the down-regulated genes include TNFRSF21, CDYL, RELL1, USP3, CTSK and FAM156B;

[0193] The predictive marker determination module is used to use the differentially expressed gene set consisting of 12 genes, namely, ZSWIM3, EGR2, EGR3, FFAR2, LIMA1, APOL1, TNFRSF21, CDYL, RELL1, USP3, CTSK and FAM156B, as a predictive marker for human susceptibility to COVID-19.

[0194] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the aforementioned method embodiment.

[0195] An embodiment of the present invention further provides an electronic device, the device comprising a processor and a memory:

[0196] The memory is used to store the program code and transmit the program code to the processor;

[0197] The processor is used to execute the method for predicting susceptibility to the new coronavirus of any embodiment of the present invention according to the instructions in the program code.

[0198] An embodiment of the present invention also provides a computer-readable storage medium, which is used to store program code, and the program code is used to execute the method for predicting susceptibility to the new coronavirus of any embodiment of the present invention.

[0199] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0200] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0201] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0202] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0203] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0204] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting susceptibility to novel coronavirus, characterized in that: The method uses predictive markers to predict the human body's susceptibility risk to COVID-19, wherein the predictive markers are differentially expressed gene sets extracted from human blood samples; the differentially expressed gene sets include up-regulated genes and down-regulated genes, the up-regulated genes include ZSWIM3, EGR2, EGR3, FFAR2, LIMA1 and APOL1, and the down-regulated genes include TNFRSF21, CDYL, RELL1, USP3, CTSK and FAM156B; the method comprises: Obtaining a blood sample to be tested, and extracting a PBMC sample to be tested from the blood sample to be tested; Extracting the gene expression data of the prediction marker in the PBMC sample to be tested from the gene expression profile; The gene expression data to be tested is input into a pre-trained COVID-19 susceptibility prediction model to determine the susceptibility risk, and the COVID-19 susceptibility risk assessment result of the human body corresponding to the blood sample to be tested is obtained; The construction and training process of the COVID-19 susceptibility prediction model includes: Obtain a pre-constructed COVID-19 susceptibility risk dataset, and randomly divide the COVID-19 susceptibility risk dataset into a COVID-19 susceptibility training set and a COVID-19 susceptibility test set according to a preset division ratio; Build several different types of classification prediction models; Based on the COVID-19 susceptible training set, each of the classification prediction models is trained for susceptible risk judgment, and parameters of each of the classification prediction models are adjusted and optimized during the training process; The classification prediction model with the highest prediction accuracy obtained after training is used as the COVID-19 susceptibility prediction model; The COVID-19 susceptibility test set is used to evaluate the model prediction effect of the COVID-19 susceptibility prediction model, obtain multiple model evaluation indicators, draw the ROC curve corresponding to the COVID-19 susceptibility risk data set, and calculate the feature importance; The construction process of the COVID-19 susceptibility risk dataset includes: Obtain blood sample data from the trial population and their symptoms during the COVID-19 pandemic; Perform multi-omics sequencing analysis based on the blood sample data to obtain antibody level data; According to the antibody level data and the symptom conditions, the test population is divided into an uninfected group, an asymptomatic group, and a symptomatic group; Extracting a test PBMC sample of each test blood sample in the blood sample data, and performing sequencing processing on each of the test PBMC samples to obtain transcriptome sequencing data; Aligning each of the transcriptome sequencing data to a reference genome to obtain gene alignment information, and converting the gene alignment information into an original gene expression matrix; Filtering the original gene expression matrix according to a preset filtering condition to obtain a filtered gene expression matrix; Screening protein-coding genes from the filtered gene expression matrix, removing duplicate genes, obtaining a target gene expression matrix, and performing standardization on the target gene expression matrix to obtain a TPM value, wherein the TPM value is used to compare gene expression levels between different experimental PBMC samples; Performing statistical analysis on the target gene expression matrix, taking the symptomatic group as a first susceptible group, and taking the asymptomatic group and the uninfected group as a second susceptible group; Perform gene differential expression analysis on the first susceptible group and the second susceptible group to screen out differentially expressed gene sets that meet the preset significant expression change conditions as predictive markers for human susceptibility to COVID-19; The test gene expression data of the predictive marker in each of the test PBMC samples are extracted from the gene expression spectrum, and the test gene expression data of each are integrated into a new coronavirus susceptibility risk data set.

2. The method for predicting susceptibility to novel coronavirus according to claim 1, characterized in that: The test population includes multiple test individuals, the symptom status is asymptomatic or symptomatic, and the test population is divided into an uninfected group, an asymptomatic group, and a symptomatic group according to the antibody level data and the symptom status, including: Test subjects with IgG antibody levels ≤100 (AU / mL) were classified into the uninfected group; The test subjects with IgG antibody level>100 (AU / mL) and asymptomatic condition were classified as asymptomatic group; The test subjects with IgG antibody levels>100 (AU / mL) and symptomatic conditions were classified as the symptomatic group.

3. The method for predicting susceptibility to novel coronavirus according to claim 1, characterized in that: The method of performing differential gene expression analysis on the first susceptible group and the second susceptible group to screen out differentially expressed gene sets that meet preset significant expression change conditions as predictive markers for human susceptibility to COVID-19 includes: The first susceptible group and the second susceptible group were subjected to gene differential expression analysis, and padj≤0.05 and log2FC>0 were used as the significant up-regulation criteria of genes to screen out up-regulated genes, which included ZSWIM3, EGR2, EGR3, FFAR2, LIMA1 and APOL1; Using log2FC<0 as the significant down-regulation standard of genes, down-regulated genes were screened out, including TNFRSF21, CDYL, RELL1, USP3, CTSK and FAM156B; The differentially expressed gene set consisting of 12 genes, including ZSWIM3, EGR2, EGR3, FFAR2, LIMA1, APOL1, TNFRSF21, CDYL, RELL1, USP3, CTSK and FAM156B, will be used as a predictive marker for human susceptibility to COVID-19.

4. A device for predicting susceptibility to novel coronavirus, characterized in that: The device is applied to a method for predicting susceptibility to the novel coronavirus, wherein the method uses a predictive marker to predict the susceptibility risk of the novel coronavirus to the human body, wherein the predictive marker is a differentially expressed gene set extracted from a human blood sample; the differentially expressed gene set includes up-regulated genes and down-regulated genes, wherein the up-regulated genes include ZSWIM3, EGR2, EGR3, FFAR2, LIMA1, and APOL1, and the down-regulated genes include TNFRSF21, CDYL, RELL1, USP3, CTSK, and FAM156B; the device includes: A PBMC sample extraction module to be tested, used to obtain a blood sample to be tested, and extract a PBMC sample to be tested from the blood sample to be tested; A gene expression data extraction module to be tested, used to extract the gene expression data to be tested of the prediction marker in the PBMC sample to be tested from the gene expression spectrum; A COVID-19 susceptibility risk judgment module, which is used to input the gene expression data to be tested into a pre-trained COVID-19 susceptibility prediction model to perform susceptibility risk judgment, and obtain the COVID-19 susceptibility risk assessment result of the human body corresponding to the blood sample to be tested; The device also includes a model building and training module, which specifically includes: A data set partitioning module is used to obtain a pre-constructed COVID-19 susceptibility risk data set, and randomly divide the COVID-19 susceptibility risk data set into a COVID-19 susceptibility training set and a COVID-19 susceptibility test set according to a preset partitioning ratio; A classification prediction model building module is used to build several different types of classification prediction models; A model training module, for performing susceptibility risk judgment training on each of the classification prediction models based on the COVID-19 susceptibility training set, and adjusting and optimizing parameters of each of the classification prediction models during the training process; A COVID-19 susceptibility prediction model determination module, which is used to use the classification prediction model with the highest prediction accuracy obtained after training as the COVID-19 susceptibility prediction model; A model evaluation module, used to evaluate the model prediction effect of the new coronavirus susceptibility prediction model using the new coronavirus susceptibility test set, obtain multiple model evaluation indicators, draw the ROC curve corresponding to the new coronavirus susceptibility risk data set, and calculate feature importance; The device also includes a COVID-19 susceptible risk data set construction module, and the COVID-19 susceptible risk data set construction module specifically includes: A blood sample data acquisition module, used to obtain blood sample data of the test population and the symptoms of the test population during the COVID-19 pandemic; A multi-omics sequencing analysis module, used to perform multi-omics sequencing analysis based on the blood sample data to obtain antibody level data; A test population division module, used to divide the test population into an uninfected group, an asymptomatic group and a symptomatic group according to the antibody level data and the symptom conditions; A transcriptome sequencing data generation module, used for extracting a test PBMC sample of each test blood sample in the blood sample data, and performing sequencing processing on each of the test PBMC samples to obtain transcriptome sequencing data; An original gene expression matrix construction module is used to align each of the transcriptome sequencing data to a reference genome, obtain gene alignment information, and convert the gene alignment information into an original gene expression matrix; A gene filtering module, used to filter the original gene expression matrix according to a preset filtering condition to obtain a filtered gene expression matrix; A gene expression matrix processing module, used to screen protein-coding genes from the filtered gene expression matrix, remove duplicate genes, obtain a target gene expression matrix, and obtain a TPM value after standardization of the target gene expression matrix, and the TPM value is used to compare gene expression levels between different experimental PBMC samples; A statistical analysis module, used for performing statistical analysis on the target gene expression matrix, taking the symptomatic group as a first susceptible group, and taking the asymptomatic group and the uninfected group as a second susceptible group; A predictive marker screening module is used to perform gene differential expression analysis on the first susceptible group and the second susceptible group, and screen out differentially expressed gene sets that meet preset significant expression change conditions as predictive markers for human susceptibility to COVID-19; The new coronavirus susceptibility risk data set integration module is used to extract the test gene expression data of the predictive marker in each of the test PBMC samples from the gene expression spectrum, and integrate the various test gene expression data into a new coronavirus susceptibility risk data set.

5. An electronic device, characterized in that: The device comprises a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the method for predicting susceptibility to the new coronavirus as described in any one of claims 1-3 according to the instructions in the program code.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store program code, and the program code is used to execute the method for predicting susceptibility to the new coronavirus as described in any one of claims 1-3.

Images