Sleep deprivation marker identification method based on WGCNA and machine learning

Abstract

The invention discloses a sleep deprivation marker identification method based on WGCNA and machine learning, and relates to the technical field of sleep deprivation markers. The method specifically comprises the following steps: S1, data acquisition and test design; s2, differential gene identification; s3, function enrichment; s4, WGCNA analysis is carried out; s5, LASSO calculation is carried out; s6, performing a sleep deprivation mouse model experiment; s7, determining noradrenaline by using an enzyme-linked immunosorbent assay method; s8, a Western blot western blot test is carried out; s9, analyzing immune characteristics; and S10, statistical analysis. According to the invention, two hub genes KLRD1 and SPON2 related to sleep deprivation are screened through bioinformatics and machine learning methods, the important role of KLRD1 in SD is confirmed, a new SD biomarker SPON2 is identified, and the immune spectrum in the SD state is comprehensively analyzed for the first time on the genetic level.

Description

Technical Field

[0001] The present invention relates to the technical field of sleep deprivation markers, and in particular to a method for identifying sleep deprivation markers based on WGCNA and machine learning. Background Art

[0002] Sleep deprivation (SD) is prevalent in modern society and is significantly associated with many health problems, including metabolic disorders, cardiovascular diseases, mood disorders, and cognitive impairment. Given the major public health challenges posed by sleep deprivation, it is critical to explore new biomarkers to accurately diagnose sleep status, predict potential health risks, and effectively monitor intervention effects.

[0003] Sleep is an essential component of human physiology and is involved in many essential functions, including cell repair, energy restoration, metabolic regulation, and immune system maintenance. However, sleep deprivation (SD) is increasingly common in modern society and has serious consequences for public health. More than one-third of the global population experiences sleep deprivation (World Health Organization).

[0004] SD impairs cognitive function and compromises overall health. SD significantly affects learning ability, memory retention, and cognitive decision-making processes. Previous studies have reported that even one night of SD can lead to the accumulation of amyloid-β in the brain, an important cause of Alzheimer's disease. Just 5 hours of sleep deprivation can lead to changes in the activation state of the brain's mTOR pathway, subsequently impairing cognitive performance. In addition, SD is significantly associated with various endocrine and metabolic disorders and may increase the risk of obesity, insulin resistance, hypertension, and coronary heart disease. Reduced sleep duration significantly increases the risk of acute myocardial infarction and coronary artery disease. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and propose a method for identifying sleep deprivation markers based on WGCNA and machine learning.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] The sleep deprivation marker identification method based on WGCNA and machine learning includes the following steps:

[0008] S1: Data acquisition and experimental design; Three RNA sequencing datasets related to sleep deprivation research were obtained from the GEO database, including GSE37667, GSE56931, and GSE98564, with a total of 171 samples, including 107 sleep deprivation samples and 64 normal samples;

[0009] S2: Identification of differentially expressed genes; the “limma” package was used in R software for normalization and analysis of the exploratory cohort. During sleep deprivation, gene expression changed slightly, and |log2FC|>0.2 and P<0.05 were used as the significance thresholds for differentially expressed genes associated with sleep deprivation.

[0010] S3: Functional enrichment; Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses were performed through annotation, visualization, and integrated discovery bioinformatics resource databases to demonstrate potential effects of sleep deprivation; the top ten findings were presented in order of increasing P value; Gene set enrichment analysis was used to identify biological processes that showed significant changes after sleep deprivation;

[0011] S4: WGCNA analysis; co-expressed gene modules were identified by WGCNA analysis, and modules associated with sleep deprivation were determined;

[0012] S5: LASSO calculation; feature selection and model parameter reduction by LASSO regression;

[0013] S6: Sleep deprivation mouse model experiment; 9 healthy male C57BL / 6J mice aged 6 to 8 weeks were used; the laboratory environment was maintained at 23±1℃, with 12 hours of light and 12 hours of darkness alternating; all experimental mice had free access to food and water;

[0014] S7: Norepinephrine was measured using ELISA; mouse prefrontal cortex tissue samples were collected, processed, and lysed according to standard operating procedures, and the experimental samples were collected and the norepinephrine content of the samples was measured using an ELISA kit;

[0015] S8: Western blot; mouse prefrontal tissue was lysed using radioimmunoprecipitation buffer containing a mixture of protease and phosphatase inhibitors. After lysis, the samples were centrifuged at 12,000 rpm and 4°C for 20 minutes, and the supernatant was collected; the protein concentration in the supernatant was quantified using a BCA protein concentration assay kit; the supernatant was mixed with five times the loading buffer and heated at 95°C for 5 minutes; equal amounts of total protein were loaded and separated using 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to a 0.22 μm PVDF membrane; the membrane was blocked with skim milk and incubated with primary antibodies: KLRD1, SPON2, and β-actin overnight at 4°C; the next day, the membrane was incubated with secondary antibodies; the target protein bands were visualized by enhanced chemiluminescence and photographed;

[0016] S9: Immune signature analysis; 29 immune markers were evaluated for each sample using single-sample gene set enrichment analysis, including 16 immune cell infiltration states and 13 immune function scores, to investigate the effects of sleep deprivation on human immune markers; Pearson correlation analysis was performed to identify potential correlations between immune markers and biomarkers;

[0017] S10: Statistical analysis; R software and Graph Pad Prism software were used for data analysis; data are presented as mean ± SD; differences among multiple groups were analyzed by one-way analysis of variance; Wilcoxon rank sum test was used to evaluate the differences between two groups.

[0018] Preferably: the data acquisition and experimental design merge the GSE37667 and GSE56931 data sets to create an exploratory cohort for hub gene screening, and the GSE98564 samples serve as an external validation cohort.

[0019] Further: the WGCNA analysis parameters were set as follows: merge cut height = 0.3, depth segmentation = 4, minimum module size = 30; as a result, 24 different modules were identified.

[0020] As a preferred embodiment of the present invention: in the sleep deprivation mouse model experiment, after the experimental mice have undergone a one-week adaptation period, the C57BL / 6J mice are randomly allocated to a normal control group, a 1-day sleep deprivation group and a 3-day sleep deprivation group, with three mice in each group; sleep deprivation is induced by using a horizontal rod-type automated sleep deprivation device.

[0021] Based on the above scheme: the mouse prefrontal cortex tissue sample for the determination of norepinephrine by enzyme-linked immunosorbent assay was rinsed with pre-cooled phosphate buffer to remove residual blood; the tissue sample was excised and mixed with PBS; the tissue was lysed by ultrasonic disruption to obtain a homogenate; the homogenate was centrifuged at 5000 rpm for 10 minutes, and the supernatant was collected to determine the norepinephrine content.

[0022] Preferably: before the experiment of measuring norepinephrine by enzyme-linked immunosorbent assay, the kit is equilibrated to room temperature; samples or standards of different concentrations are added to the ELISA plate in an amount of 50 μl per well, and then 50 μl of HRP-antigen working solution is added to each well and incubated at 37°C for 1 hour; after washing the plate five times with washing solution, 90 μl of TMB substrate is added to each well and incubated at 37°C in the dark for 15 minutes, and then 50 μl of stop solution is added to each well, and the optical density is immediately measured at 450 nm.

[0023] The beneficial effects of the present invention are:

[0024] This WGCNA and machine learning-based method for identifying sleep deprivation markers screened two hub genes (KLRD1 and SPON2) related to sleep deprivation (SD) by using bioinformatics and machine learning methods, confirmed the important role of KLRD1 in SD, identified the new SD biomarker SPON2, and conducted the first comprehensive analysis of the immune spectrum in the SD state at the genetic level. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a table of three RNA sequencing datasets related to the GEO dataset and sleep deprivation (SD) study of the present invention;

[0026] Figure 2 This is a schematic diagram of the research workflow of the present invention;

[0027] Figure 3 Schematic diagram of functional enrichment analysis of differentially expressed genes (DEGs) between the control group and sleep deprivation (SD) samples of the present invention and differentially expressed genes (DEGs) of the exploratory cohort;

[0028] Figure 4 Schematic diagram of weighted gene co-expression network analysis (WGCNA) of the present invention;

[0029] Figure 5 A schematic diagram of the hub gene associated with sleep deprivation (SD) identified by the present invention;

[0030] Figure 6 A schematic diagram of the present invention for verifying hub gene expression in a sleep deprivation (SD) mouse model;

[0031] Figure 7 A schematic diagram showing the results of immune cell infiltration by single sample gene set enrichment analysis (ssGSEA) of the present invention;

[0032] Figure 8 Schematic diagram of the gene set enrichment analysis (GSEA) of the present invention.

[0033] Legend:

[0034] Figure 1 : Details of the GEO dataset, with sample sizes indicated in brackets;

[0035] Figure 2:DEGs: differentially expressed genes; GO: gene ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; LASSO: least absolute shrinkage and selection operator; NC: normal control; ROC curve: receiver operating characteristic curve; SD-1d: one-day SD group; SD-3d: three-day SD group; ssGSVA: single sample gene set variation analysis; WGCNA: weighted gene co-expression network analysis;

[0036] Figure 3 : (A) Volcano plot for differential expression analysis; (BC) GO and KEGG analysis of differentially expressed necrosis-related genes; GO: Gene Ontology; BP: Biological Process; CC: Cellular Component; MF: Molecular Function; KEGG: Kyoto Encyclopedia of Genes and Genomes;

[0037] Figure 4 :(A) Clustering tree of samples based on groups;(B) Soft threshold (β=9) confirmed using scale independence and average connectivity calculations;(C) Clustering genes with different similarities using topological overlap and module colors;(D) Module-trait association table, which contains a correlation table and a feature table, and each cell contains a correlation coefficient and P value;(E) Correlation between module membership value and gene significance in sky blue modules;

[0038] Figure 5 :(A) Venn diagram showing the intersection of WGCNA key genes and DEGs in the exploratory cohort;(B) LASSO regression analysis of the seven intersection genes;(C) Clustering of genes with different similarities using topological overlap and module colors, seven lines of different colors represent seven intersection genes, and each curve represents the trajectory of each independent variable coefficient;(DG) Expression levels of hub genes in the exploratory cohort and validation cohort;

[0039] Figure 6 :(A) Nomogram of KLRD1 and SPON2;(B) ROC curves of KLRD1 and SPON2 in the exploratory cohort and external validation cohort;(C) ELISA was used to evaluate the concentration of norepinephrine in the prefrontal cortex (N=3);(D) Western Blot (WB) results of KLRD1 and SPON2 in the mouse SD model and control group (N=3);NC: normal control group;SD-1d: one-day SD group;SD-3d: three-day SD group;***P<0.0001;**P<0.01;*P<0.05;

[0040] Figure 7:(A) Comparison of ssGSEA scores of 13 immune-related functions and 16 immune cells between the SD group and the normal group;(B) Relationship between hub genes and immune features; aDCs: activated dendritic cells; CCR: cytokine and cytokine receptor; DCs: dendritic cells; HLA: human leukocyte antigen; iDCs: immature dendritic cells; MHC: major histocompatibility complex; pDCs: plasmacytoid dendritic cells; Tfh cells: follicular helper T cells; TIL: tumor infiltrating lymphocytes; Treg: regulatory T cells; INF: interferon;

[0041] Figure 8 : (A) GSEA enrichment analysis of down-regulated genes; (B) GSEA enrichment analysis of up-regulated genes. DETAILED DESCRIPTION

[0042] The technical solution of the present invention is further described in detail below in conjunction with specific implementation methods.

[0043] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0044] The identification method of sleep deprivation markers based on WGCNA and machine learning, such as Figure 2 As shown, the following steps are included:

[0045] S1: Data acquisition and experimental design; Three RNA sequencing datasets related to sleep deprivation (SD) research were obtained from the GEO database, including GSE37667, GSE56931, and GSE98564, with a total of 171 samples, including 107 sleep deprivation samples and 64 normal samples, such as Figure 1 A comprehensive overview of the GEO dataset of the sample is shown in Figure .

[0046] like Figure 2 Flowchart of the overall analysis process shown, data acquisition and experimental design. GSE37667 and GSE56931 datasets were merged to create an exploratory cohort for hub gene screening, and GSE98564 samples were used as an external validation cohort; bioinformatics analysis techniques were used to elucidate molecular markers associated with SD; differentially expressed genes (DEGs) were identified in the exploratory cohort, and functional enrichment and gene set enrichment analysis (GSEA) were performed on these genes to reveal their potential biological functions.

[0047] WGCNA was used to identify co-expression modules, and the LASSO algorithm was applied to identify hub genes associated with SD; the diagnostic efficacy of these genes was confirmed in exploratory and external validation cohorts; the association between SD and immune infiltration was examined using single-sample gene set enrichment analysis (ssGSEA).

[0048] The sample inclusion criteria are as follows:

[0049] (1) Datasets in which participants in the SD group experienced at least 24 hours of total sleep deprivation;

[0050] (2) a data set for the control sample collected at 8 a.m. under standard working and resting conditions;

[0051] (3) The samples were from human peripheral blood and had corresponding transcriptome data;

[0052] (4) Be a publicly available and accessible research dataset.

[0053] The website of the GEO database is: https: / / www.ncbi.nlm.nih.gov / geo / .

[0054] S2: Identification of differentially expressed genes; the “limma” package was used in R software (version 4.3.2) for normalization and analysis of the exploratory cohort. Gene expression changed slightly during sleep deprivation, using |log2FC|>0.2 and P<0.05 as the significance thresholds for differentially expressed genes (DEGs) associated with sleep deprivation; in R software, “Pheatmap” and “ggplot2” were used to visualize the results.

[0055] S3: Functional enrichment; Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed using the Database for Annotation, Visualization, and Integrated Discovery Bioinformatics Resource (DAVID) to demonstrate the potential effects of sleep deprivation (SD); the top ten findings were presented in increasing order of P value (P<0.05); Gene set enrichment analysis (GSEA) was used to identify biological processes that showed significant changes after sleep deprivation (SD); the gene sets associated with this analysis (c5.go.bp.v2023.2.Hs.symbols) were obtained from the MSigDB database.

[0056] MSigDB database URL:

[0057] http: / / www.gsea-msigdb.org / gsea / msigdb / .

[0058] S4: WGCNA analysis; co-expressed gene modules were identified by WGCNA analysis, and modules associated with sleep deprivation (SD) were determined; the GEO dataset provided more than 17,000 sequenced genes, most of which did not show differential expression between samples; the top 75% of genes with the highest variance were selected for WGCNA analysis; cluster analysis was used to evaluate sample correlation, and the abnormal sample GSM1371618 (healthy control) was excluded to ensure the reliability of the results.

[0059] The soft threshold power was set to 9 to create a scale-free network, at which point the scale independence exceeded 0.85 for the first time (R2=0.83) and showed relatively high average connectivity; genes with similar expression patterns were clustered and analyzed; WGCNA analysis parameters were set as follows: merge cut height=0.3, deep split=4, minimum module size=30; 24 different modules were identified as a result; the sky blue module showed the highest correlation with sleep deprivation (r=0.3, P=0.003), where genes with gene significance values>0.2 and module membership values>0.8 were designated as key genes within the module.

[0060] S5: LASSO calculation; feature selection and model parameter reduction by LASSO regression; LASSO exploits the correlation between independent variables and effectively sets the coefficients of the correlated independent variables to zero, thereby mitigating the potential impact of multicollinearity on regression results; The LASSO algorithm is used to precisely narrow the range of genes associated with sleep deprivation (SD), thereby improving the accuracy and reliability of the model.

[0061] S6: Sleep deprivation mouse model experiment; 9 healthy male C57BL / 6J mice aged 6 to 8 weeks were used; the mice were purchased from Beijing Sibeifu Biotechnology Co., Ltd.; during the experiment, the laboratory environment was maintained at 23±1℃, with 12 hours of light and 12 hours of darkness alternating; all experimental mice had free access to food and water;

[0062] After a one-week adaptation period, the C57BL / 6J mice were randomly assigned to a normal control group, a 1-day sleep deprivation group, and a 3-day sleep deprivation group, with three mice in each group.

[0063] Sleep deprivation was induced by using a horizontal-bar automated sleep deprivation apparatus (model: Kw-bd, KEW, China), which had a stainless steel rotating rod (37 cm in diameter) fixed in the center of a plastic cage (40 cm in height, 38.4 cm in inner diameter) that rotated continuously at 3 rpm to prevent mice from falling asleep.

[0064] S7: Determination of norepinephrine using enzyme-linked immunosorbent assay; mouse prefrontal cortex tissue samples were collected according to standard operating procedures and immediately rinsed with pre-cooled phosphate buffered saline (PBS) (0.01 M, pH = 7.4) to remove residual blood; tissue samples were excised and mixed with an appropriate volume of PBS (10 ml per gram of tissue); the tissue was lysed using ultrasonic disruption to obtain a homogenate; the homogenate was centrifuged at 5000 rpm for 10 min, and the supernatant was collected to determine the norepinephrine (NE) content; NE content was detected using a NE enzyme-linked immunosorbent assay (ELISA) kit (JL13969, Jonln, Shanghai, China).

[0065] Before the NE content quantification experiment, the kit was equilibrated to room temperature; samples or standards of different concentrations were added to the ELISA plate at 50 μl per well, and then 50 μl of HRP-antigen working solution was added to each well and incubated at 37°C for 1 hour; after washing the plate five times with washing solution, 90 μl of TMB substrate was added to each well and incubated at 37°C in the dark for 15 minutes, and then 50 μl of stop solution was added to each well, and the optical density was immediately measured at 450 nm.

[0066] S8: Western blot; mouse prefrontal tissues were lysed using radioimmunoprecipitation assay (RIPA) buffer containing a cocktail of protease and phosphatase inhibitors; after lysis, samples were centrifuged at 12,000 rpm and 4°C for 20 min, and the supernatant was collected; the protein concentration in the supernatant was quantified using a BCA protein concentration assay kit (Beyotime, Shanghai, China); the supernatant was mixed with five times the loading buffer and heated at 95°C for 5 min; equal amounts of total protein (30 μg) were loaded and analyzed using 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE, ACE, Jiangsu, China). China) for separation and then transferred to a 0.22 μm PVDF membrane (Millipore, Germany); the membrane was blocked with skim milk and incubated with primary antibodies: KLRD1 (#DF6773, Affinity, Jiangsu, China), SPON2 (#DF4682, Affinity, Jiangsu, China) and β-actin (#81115-1-RR, Proteintech, Wuhan, China) at 4°C overnight; the next day, the membrane was incubated with secondary antibodies (#SA00001-2, Proteintech, Wuhan, China); the target protein bands were visualized by enhanced chemiluminescence (Merck Millipore, Germany) and photographed.

[0067] S9: Immune signature analysis; Single-sample gene set enrichment analysis (ssGSEA) was used to evaluate 29 immune markers for each sample, including 16 immune cell infiltration states and 13 immune function scores, to investigate the effects of sleep deprivation (SD) on human immune markers; Pearson correlation analysis was performed to identify potential correlations between immune markers and biomarkers;

[0068] S10: Statistical analysis; R software (version 4.32) and Graph Pad Prism software (version 9) were used for data analysis; data are presented as mean ± standard deviation; differences between multiple groups were analyzed by one-way ANOVA; Wilcoxon rank sum test was used to evaluate the differences between two groups; P values ​​less than 0.05 were considered statistically significant.

[0069] The experimental results are as follows:

[0070] 1. Identification of sleep deprivation-related differentially expressed genes: The GSE37667 and GSE56931 datasets were preprocessed using the R packages “GEOquery” and “limma”, resulting in 100 samples (63 sleep deprived and 37 normal samples); 143 differentially expressed genes (DEGs) between the two groups were identified, of which 70 genes were upregulated and 73 genes were downregulated in sleep deprived participants ( Figure 3 A).

[0071] 2. Differential gene function enrichment: Gene ontology (GO) biological process analysis revealed that the main biological processes were related to cytotoxicity to other organisms, inflammatory response, positive regulation of apoptotic processes, and regulation of immune system processes; the plasma membrane and extracellular region were identified as the most relevant cellular components; the analysis revealed transmembrane signaling receptor activity and carbohydrate binding as the most significant molecular functions ( Figure 3 B); Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis revealed that natural killer (NK) cell-mediated cytotoxicity was the most significantly enriched pathway (P = 0.02; Figure 3 C), which suggests a strong correlation between sleep deprivation and inflammatory / immune responses.

[0072] 3. WGCNA analysis and identification of sleep deprivation-related modules: WGCNA (weighted gene co-expression network analysis) identified potential regulatory networks associated with sleep deprivation (SD); cluster analysis was used to examine the gene expression profiles of the participants ( Figure 4 A), and use a soft threshold power (β) of 9 (unscaled R 2 =0.83) to fit a scale-free network ( Figure 4 B).

[0073] A hierarchical clustering tree containing 24 significant modules was created ( Figure 4 C); the sky blue module showed the highest correlation with the SD feature (r = 0.3, P = 0.0003; Figure 4 D), containing 116 genes; based on gene significance values ​​> 0.2 and module membership values ​​> 0.8, 10 module key genes (PRF1, KLRD1, FASLG, NKG7, SPON2, TBX21, NMUR1, S1PR5, OSBPL5, and FGFBP2) were identified; Figure 4 E).

[0074] 4. Identification and evaluation of hub genes: These key genes were analyzed with 143 differentially expressed genes (DEGs) and seven intersection genes were found ( Figure 5 A); LASSO regression analysis was used to screen these seven genes ( Figure 5 BC) to identify key biomarkers associated with sleep deprivation (SD); this process ultimately identified two hub genes: KLRD1 and SPON2 ( Figure 5 DG); a nomogram predicting sleep status was created based on these two hub genes ( Figure 6 A); The predictive accuracy of the nomogram was verified using ROC curves of internal and external datasets, showing satisfactory performance ( Figure 6 B).

[0075] 5. Animal model verification of hub gene: Based on previous studies on the effect of SD on the noradrenergic system, we paid special attention to the effect of sleep deprivation (SD) on the noradrenergic system. The experimental results showed that the level of noradrenalin (NE) increased significantly after a single day of sleep deprivation, while the level of NE decreased significantly after three consecutive days of SD ( Figure 6 C); These results reveal the effectiveness of the horizontal bar automated sleep deprivation system.

[0076] Western Blot (WB) analysis was used to verify the protein levels of hub genes; the study found that the protein secretion levels of hub genes in sleep-deprived mice were significantly different from those in healthy mice ( Figure 6 D).

[0077] 6. Immune signature analysis: Single sample gene set enrichment analysis (ssGSEA) was used to evaluate 29 immune signatures in 63 sleep deprivation samples and 37 healthy control samples; compared with the control group, SD mice showed lower levels of cytotoxic activity, pro-inflammatory elements, NK cells, pDCs (plasmacytoid dendritic cells), and T cell co-inhibition ( Figure 7A); Subsequently, we calculated the Pearson correlation coefficients between the seven intersection genes and the 29 immune signatures; Figure 7 B shows that these seven intersection genes are significantly positively correlated with cytotoxic activity and NK cell infiltration; therefore, the hypothesis is proposed: SD significantly affects the immune microenvironment.

[0078] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for identifying sleep deprivation markers based on WGCNA and machine learning, characterized in that: The steps include: S1: Data acquisition and experimental design; Three RNA sequencing datasets related to sleep deprivation research were obtained from the GEO database, including GSE37667, GSE56931, and GSE98564, with a total of 171 samples, including 107 sleep deprivation samples and 64 normal samples; S2: Identification of differentially expressed genes; The "limma" package was used in R software to normalize and analyze the exploratory cohort. During sleep deprivation, gene expression changed slightly, and |log2FC|>0.2 and P<0.05 were used as the significance thresholds for differentially expressed genes associated with sleep deprivation. S3: Functional enrichment; Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses were performed through annotation, visualization, and integrated discovery bioinformatics resource databases to demonstrate potential effects of sleep deprivation; the top ten findings were presented in order of increasing P value; Gene set enrichment analysis was used to identify biological processes that showed significant changes after sleep deprivation; S4: WGCNA analysis; co-expressed gene modules were identified by WGCNA analysis, and modules associated with sleep deprivation were determined; S5: LASSO calculation; feature selection and model parameter reduction by LASSO regression; S6: Sleep deprivation mouse model experiment; 9 healthy male C57BL / 6J mice aged 6 to 8 weeks were used; the laboratory environment was maintained at 23±1℃, with 12 hours of light and 12 hours of darkness alternating; all experimental mice had free access to food and water; S7: Norepinephrine was measured using ELISA; mouse prefrontal cortex tissue samples were collected, processed, and lysed according to standard operating procedures, and the experimental samples were collected and the norepinephrine content of the samples was measured using an ELISA kit; S8: Western blot; mouse prefrontal tissue was lysed using radioimmunoprecipitation buffer containing a cocktail of protease and phosphatase inhibitors. After lysis, the samples were centrifuged at 12,000 rpm and 4°C for 20 minutes, and the supernatant was collected; the protein concentration in the supernatant was quantified using a BCA protein concentration assay kit; the supernatant was mixed with five times the loading buffer and heated at 95°C for 5 minutes; equal amounts of total protein were loaded and separated using 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to a 0.22 μm PVDF membrane; the membrane was blocked with skim milk and incubated with primary antibodies: KLRD1, SPON2, and β-actin overnight at 4°C; the next day, the membrane was incubated with secondary antibodies; the target protein bands were visualized by enhanced chemiluminescence and photographed; S9: Immune profiling; Single-sample gene set enrichment analysis was used to evaluate 29 immune markers per sample, including 16 immune cell infiltration states and 13 immune function scores, to investigate the effects of sleep deprivation on human immune markers; Pearson correlation analysis was performed to identify potential correlations between immune markers and biomarkers; S10: Statistical analysis; R software and Graph Pad Prism software were used for data analysis; data are presented as mean ± SD; differences among multiple groups were analyzed by one-way analysis of variance; Wilcoxon rank sum test was used to evaluate the differences between two groups.

2. The method for identifying sleep deprivation markers based on WGCNA and machine learning according to claim 1, characterized in that: The data acquisition and experimental design merged the GSE37667 and GSE56931 datasets to create an exploratory cohort for hub gene screening, and the GSE98564 samples were used as an external validation cohort.

3. The method for identifying sleep deprivation markers based on WGCNA and machine learning according to claim 1, characterized in that: The WGCNA analysis parameters were set as follows: merge cut height = 0.3, depth segmentation = 4, minimum module size = 30; as a result, 24 different modules were identified.

4. The method for identifying sleep deprivation markers based on WGCNA and machine learning according to claim 1, characterized in that: In the sleep deprivation mouse model experiment, after a one-week adaptation period, the C57BL / 6J mice were randomly assigned to a normal control group, a 1-day sleep deprivation group, and a 3-day sleep deprivation group, with three mice in each group; sleep deprivation was induced by using a horizontal bar-type automated sleep deprivation device.

5. The method for identifying sleep deprivation markers based on WGCNA and machine learning according to claim 1, characterized in that: The mouse prefrontal cortex tissue samples for the determination of norepinephrine using the enzyme-linked immunosorbent assay were rinsed with pre-cooled phosphate buffered saline to remove residual blood; the tissue samples were excised and mixed with PBS; The tissue was lysed by ultrasonic disruption to obtain a homogenate; the homogenate was centrifuged at 5000 rpm for 10 minutes, and the supernatant was collected to determine the norepinephrine content.

6. The method for identifying sleep deprivation markers based on WGCNA and machine learning according to claim 1, characterized in that: Before the experiment of measuring norepinephrine by enzyme-linked immunosorbent assay, the kit was equilibrated to room temperature; samples or standards of different concentrations were added to the ELISA plate at 50 μl per well, and then 50 μl of HRP-antigen working solution was added to each well and incubated at 37°C for 1 hour; the plate was washed five times with washing solution. After that, 90 μl of TMB substrate was added to each well and incubated at 37°C in the dark for 15 min. Then 50 μl of stop solution was added to each well, and the optical density was immediately measured at 450 nm.

Images