Single-cell RNA sequencing reveals cellular diversity and gene expression dynamics in maize root development

Main conclusionIntegrated root phenotypes and transcriptome analysis have revealed key candidate genes responsible for maize root growth and development in potassium deficiency.Potassium (K) is a vital macronutrient for plant growth, but our understanding of its regulatory mechanisms in maize root system architecture (RSA) and K+ uptake remains limited. To address this, we conducted hydroponic and field trials at different growth stages. K+ deficiency significantly inhibited maize root growth, with metrics like total root length, primary root length, width and maximum root number reduced by 50% to 80% during early seedling stages. In the field, RSA traits exhibited maximum values at the silking stage but continued to decline thereafter. Furthermore, K deprivation had a pronounced negative impact on root morphology and RSA growth and grain yield. RNA-Seq analysis identified 5972 differentially expressed genes (DEGs), including 17 associated with K+ signaling, transcription factors, and transporters. Weighted gene co-expression network analysis revealed 23 co-expressed modules, with enrichment of transcription factors at different developmental stages under K deficiency. Several DEGs and transcription factors were predicted as potential candidate genes responsible for maize root growth and development. Interestingly, some of these genes exhibited homology to well-known regulators of root architecture or development in Arabidopsis, such as Zm00001d014467 (AtRCI3), Zm00001d011237 (AtWRKY9), and Zm00001d030862 (AtAP2/ERF). Identifying these key genes helps to provide a deeper understanding of the molecular mechanisms governing maize root growth and development under nutrient deficient conditions offering potential benefits for enhancing maize production and improving stress resistance through targeted manipulation of RSA traits in modern breeding efforts.

Background Acute type A aortic dissection (AAD) is a catastrophic disease with high mortality, but the pathogenesis has not been fully elucidated. This study is aimed at identifying hub genes and immune cells associated with the pathogenesis of AAD. Methods The datasets were downloaded from Gene Expression Omnibus (GEO). Gene Set Enrichment Analysis (GSEA), gene set variation analysis (GSVA), and differential analysis were performed. The differentially expressed genes (DEGs) were intersected with specific genes collected from MSigDB. The gene function and pathway enrichment analysis were also performed on intersecting genes. The key modules were selected by weighted gene coexpression network analysis (WGCNA). Hub genes were identified by least absolute shrinkage and selection operator (LASSO) analysis and were verified in the metadataset. The immune cell infiltration was analyzed by CIBERSORT, and the relationship between hub genes and immune cells was performed by Pearson's correlation analysis. The single-cell RNA sequencing (scRNA-seq) dataset was used to verify the differences in DNA damage and repair signaling pathways and hub genes in different cell types. Results The results of GSEA and GSVA indicated that DNA damage and repair processes were activated in the occurrence of AAD. The gene function and pathway enrichment analysis on differentially expressed DNA damage- and repair-related genes showed that these genes were mainly involved in the regulation of the cell cycle process, cellular response to DNA damage stimulus, response to wounding, p53 signaling pathway, and cellular senescence. Three key modules were identified by WGCNA. Five genes were screened as hub genes, including CDK2, EIF4A1, GLRX, NNMT, and SLCO2A1. Naive B cells and Gamma delta T cells (γδ T cells) were decreased in AAD, but monocytes and M0 macrophages were increased. scRNA-seq analysis included that DNA damage and repair processes were activated in smooth muscle cells (SMCs), tissue stem cells, and monocytes in the aortic wall of patients with AAD. Conclusions Our results suggested that DNA damage- and repair-related genes may be involved in the occurrence of AAD by regulating many biological processes. The hub genes and immune cells reported in this study also increase the understanding of AAD.

Non-alcoholic fatty liver disease (NAFLD) is the most common chronic liver disease globally, with the potential to progress to non-alcoholic steatohepatitis (NASH), cirrhosis, and even hepatocellular carcinoma. Given the absence of effective treatments to halt its progression, novel molecular approaches to the NAFLD diagnosis and treatment are of paramount importance. Firstly, we downloaded oxidative stress-related genes from the GeneCards database and retrieved NAFLD-related datasets from the GEO database. Using the Limma R package and WGCNA, we identified differentially expressed genes closely associated with NAFLD. In our study, we identified 31 intersection genes by analyzing the intersection among oxidative stress-related genes, NAFLD-related genes, and genes closely associated with NAFLD as identified through Weighted Gene Co-expression Network Analysis (WGCNA). In a study of 31 intersection genes between NAFLD and Oxidative Stress (OS), we identified three hub genes using three machine learning algorithms: Least Absolute Shrinkage and Selection Operator (LASSO) regression, Support Vector Machine - Recursive Feature Elimination (SVM-RFE), and RandomForest. Subsequently, a nomogram was utilized to predict the incidence of NAFLD. The CIBERSORT algorithm was employed for immune infiltration analysis, single sample Gene Set Enrichment Analysis (ssGSEA) for functional enrichment analysis, and Protein-Protein Interaction (PPI) networks to explore the relationships between the three hub genes and other intersecting genes of NAFLD and OS. The distribution of these three hub genes across six cell clusters was determined using single-cell RNA sequencing. Finally, utilizing relevant data from the Attie Lab Diabetes Database, and liver tissues from NASH mouse model, Western Blot (WB) and Reverse Transcription Quantitative Polymerase Chain Reaction (RT-qPCR) assays were conducted, this further validated the significant roles of CDKN1B and TFAM in NAFLD. In the course of this research, we identified 31 genes with a strong association with oxidative stress in NAFLD. Subsequent machine learning analysis and external validation pinpointed two genes: CDKN1B and TFAM, as demonstrating the closest correlation to oxidative stress in NAFLD. This investigation found two hub genes that hold potential as novel targets for the diagnosis and treatment of NAFLD, thereby offering innovative perspectives for its clinical management.

Background and aimsSystemic lupus erythematosus (SLE) is one of the most prevalent systemic autoimmune diseases, characterized by aberrant activation of the immune system that leads to diverse clinical symptoms; periodontal disease (PD) is an inflammatory oral disorder caused by immune-mediated damage against subgingival microflora. Although clinical evidence suggests a potential association between SLE and PD, their shared pathogenic mechanisms remain unclear. This study aims to explore common genetic markers in SLE and PD that hold diagnostic and therapeutic implications.MethodsMicroarray datasets for systemic lupus erythematosus (SLE) and periodontal disease (PD) were obtained from the Gene Expression Omnibus (GEO) database. Module genes between the two diseases were screened using Weighted Gene Co-expression Network Analysis (WGCNA), and module genes overlapping between the significant correlation modules of GSE61635 and GSE16134 were identified. Functional enrichment analyses of genes within overlapping modules and their significantly correlated associated modules were performed using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. Overlapping module genes underwent differential expression analysis in GSE16134. A diagnostic model was constructed using the Random Forest (RF) machine learning technique under Receiver Operating Characteristic (ROC) curve assessment, which top 10 key genes were screened and analyzed for differential expression across three datasets (GSE61635, GSE10334, and GSE50772) to identify hub genes. Protein-protein interaction (PPI) network analysis was conducted to explore relationships between hub genes. CIBERSORT and Gene Set Variation Analysis (GSVA) were used to evaluate the correlation between shared hub genes and immune infiltration patterns as well as metabolic pathways. Finally, hub genes were validated using additional datasets, single-cell RNA sequencing (scRNA-seq) data, and immunohistochemistry (IHC) experiments.ResultsUsing WGCNA, we identified significant correlation modules and overlapping module genes, which were subjected to differential expression analysis in different datasets. Further, 4 hub genes were screened and successfully used to build a prognostic model. Those shared hub genes were associated with immunological and metabolic processes in peripheral blood. The additional datasets, scRNA-seq and IHC results verified that LY96 and TMEM140, possessing the promising diagnostic and therapeutic performance.ConclusionLY96 andTMEM140 can be used as new diagnostic and therapeutic markers for SLE and PD.

BackgroundAcute rejection (AR) in kidney transplantation is an established risk factor that reduces the survival rate of allografts. Despite standard immunosuppression, molecules with regulatory control in the immune pathway of AR can be used as important targets for therapeutic operations to prevent rejection.MethodsWe downloaded the microarray data of 15 AR patients and 37 non-acute rejection (NAR) patients from Gene Expression Omnibus (GEO). Gene network was constructed, and genes were classified into different modules using weighted gene co-expression network analysis (WGCNA). Kyoto Encyclopedia of Genes and Genomes (KEGG) and Cytoscape were applied for the hub genes in the most related module to AR. Different cell types were explored by xCell online database and single-cell RNA sequencing. We also validated the SLAMF8 and TLR4 levels in Raw264.7 and human kidney tissues of TCMR.ResultsA total of 1,561 differentially expressed genes were filtered. WGCNA was constructed, and genes were classified into 12 modules. Among them, the green module was most closely associated with AR. These genes were significantly enriched in 20 pathway terms, such as cytokine–cytokine receptor interaction, chemokine signaling pathway, and other important regulatory processes. Intersection with GS > 0.4, MM > 0.9, the top 10 MCC values and DEGs in the green module, and six hub genes (DOCK2, NCKAP1L, IL2RG, SLAMF8, CD180, and PTPRE) were identified. Their expression levels were all confirmed to be significantly elevated in AR patients in GEO, Nephroseq, and quantitative real-time PCR (qRT-PCR). Single-cell RNA sequencing showed that AR patient had a higher percentage of native T, CD1C+_B DC, NKT, NK, and monocytes in peripheral blood mononuclear cells (PBMCs). Xcell enrichment scores of 20 cell types were significantly different (p<0.01), mostly immune cells, such as B cells, CD4+ Tem, CD8+ T cells, CD8+ Tcm, macrophages, M1, and monocytes. GSEA suggests that highly expressed six hub genes are correlated with allograft rejection, interferon γ response, interferon α response, and inflammatory response. In addition, SLAMF8 is highly expressed in human kidney tissues of TCMR and in M1 phenotype macrophages of Raw264.7 cell line WGCNA accompanied by high expression of TLR4.ConclusionThis study demonstrates six hub genes and functionally enriched pathways related to AR. SLAMF8 is involved in the M1 macrophages via TLR4, which contributed to AR process.

Objective in order to discover correlated gene modules and hub genes for recurrent laryngeal squamous cell carcinoma through weighted gene co-expression network analysis method. The microarray dataset of recurrent laryngeal cancer, namely GSE27020, were obtained from the gene expression Omnibus database. Weighted gene co-expression network analysis was introduced to establish a gene co-expression network, mining key clinical trait correlated hub genes. Gene ontology enrichment analyses were performed for the genes in modules related to recurrent laryngeal squamous cell carcinoma. Then, we build up a proteinprotein interaction network with the genes in interest gene module and identified hub genes through analyze such protein-protein interaction network. The hub genes were mined using cytohubba plus-in. Finally, we analyzed these hub genes overall survival using gene expression profiling interactive analysis database. Forty four gene co-expression gene modules were achieved via weighted gene co-expression network analysis analysis. We found that the orangered4 module was the most correlated module with recurrence in laryngeal squamous cell carcinoma patients. Genes in the orangered4 module were related to organelle organization, response to chemical, regulation of catalytic activity and regulation of cell differentiation. Two genes were discovered as hub genes which were related to poor prognosis, namely annexin A2 and S100 calcium binding protein A10. Here, we found several hub genes that played important roles in recurrent laryngeal squamous cell carcinoma, which may improve our understanding of the mechanisms underlying recurrence laryngeal squamous cell carcinoma.

Background and Aims The fact that activation of the innate immune system and chronic inflammation are closely involved in the pathogenesis of diabetic Kidney disease (DKD). Recent studies have suggested the inflammatory process plays a crucial role in the progression of DKD. Identifying novel inflammatory molecules closely related to the decline of renal function is of significance in diagnosing and predicting the progression of DKD. The weighted gene co-expression network analysis (WGCNA) algorithm represents a novel systems biology method that provide the approach of association between gene modules and clinical traits to find the genes involvement into the certain phenotypic trait. The goal of this study was to identify hub genes and their roles in DKD from the gene sets associated with the decline of renal function by WGCNA. Method The Gene Expression Omnibus (GEO) database and “Nephroseq” website were searched and transcriptome study from DN biopsies with well-established clinical phenotypic data were selected for analysis. Next, we constructed a weighted gene co-expression network and identified modules negatively correlated with eGFR by WGCNA in the data of glomerular tissue. Functional annotations of the genes in modules negatively correlated with eGFR were analyzed by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses. Through protein-protein interaction (PPI) analysis and hub gene screening, the hub genes were obtained. Furthermore, we compared the expression level of hub genes between DKD and normal control and drew ROC curves to determine the diagnosis value to DKD of these genes. Results The microarray-based expression datasets GSE30528 were screened out for analysis, which included glomeruli tissue of 9 cases of DKD and 13 cases of control. This microarray platform represented the transcriptome profile of 12411 well-characterized genes. Using WGCNA, a total of 19 gene modules were identified. Then module eigengene were analyzed for correlation with clinical traits of age, sex, ethnicity and eGFR and the “MEhoneydew1” module showed negative associated with eGFR (r=-0.58). GO functional annotation showed that these 551 genes in the “MEhoneydew1” module mainly enriched in the T cell activation. KEGG annotation showed mainly enriched in chemokine signaling pathway. Except for C3, top 10 hub genes, CCR5, CXCR4, CCR7, CCL5, CXCL8, CCR2, CCR1, CX3CR1, C3AR1 and C3, are all members of chemokines or chemokine receptors. Furthermore, we compared the expression level of these 9 genes between DKD and control, and found that all of these 9 genes increased in the DKD group, and the differences of 6 genes, CCR5, CCR7, CCL5, CCR2, CCR1, C3AR1, were of statistical significance. Linear correlation analysis showed that the expression of these 6 genes was negatively correlated with eGFR, and the ROC curve showed that the area under the curve could reach 0.812∼1.0. Conclusion We identified a panel of 6 hub genes focused on chemokines and chemokine receptors critical for decline of renal function of DKD using WGCNA. These genes may serve as biomarkers for diagnosis/prognosis and as putative novel therapeutic targets for DKD.

Inherited retinal diseases (IRDs) affect ∼4.5 million people worldwide. Elusive pathogenic variants in over 280 genes are associated with one or more clinical forms of IRDs. It is necessary to understand the complex interaction among retinal cell types and pathogenic genes by constructing a regulatory network. In this study, we attempt to establish a panoramic expression view of the cooperative work in retinal cells to understand the clinical manifestations and pathogenic bases underlying IRDs. Single-cell RNA sequencing (scRNA-seq) data on the retinas from 35 retina samples of 3 species (human, mouse, and zebrafish) including 259,087 cells were adopted to perform a comparative analysis across species. Bioinformatic tools were used to conduct weighted gene co-expression network analysis (WGCNA), single-cell regulatory network analysis, cell-cell communication analysis, and trajectory inference analysis. The cross-species comparison revealed shared or species-specific gene expression patterns at single-cell resolution, such as the stathmin family genes, which were highly expressed specifically in zebrafish Müller glias (MGs). Thirteen gene modules were identified, of which nine were associated with retinal cell types, and Gene Ontology (GO) enrichment of module genes was consistent with cell-specific highly expressed genes. Many IRD genes were identified as hub genes and cell-specific regulons. Most IRDs, especially the retinitis pigmentosa (RP) genes, were enriched in rod-specific regulons. Integrated expression and transcription regulatory network genes, such as congenital stationary night blindness (CSNB) genes GRK1, PDE6B, and TRPM1, showed cell-specific expression and transcription characteristics in either rods or bipolar cells (BCs). IRD genes showed evolutionary conservation (GNAT2, PDE6G, and SAG) and divergence (GNAT2, MT-ND4, and PDE6A) along the trajectory of photoreceptors (PRs) among species. In particular, the Leber congenital amaurosis (LCA) gene OTX2 showed high expression at the beginning of the trajectory of both PRs and BCs. We identified molecular pathways and cell types closely connected with IRDs, bridging the gap between gene expression, genetics, and pathogenesis. The IRD genes enriched in cell-specific modules and regulons suggest that these diseases share common etiological bases. Overall, mining of interspecies transcriptome data reveals conserved transcriptomic features of retinas across species and promising applications in both normal retina anatomy and retina pathology.

BackgroundDiabetic nephropathy (DN), a major complication of type 2 diabetes mellitus (DM), is driven by complex mechanisms involving mitochondrial dysfunction, senescence, and chronic inflammation. Despite therapeutic advances, interventions specifically targeting mitochondrial dysfunction, senescence, and inflammation remain elusive.MethodsAn integrative analysis was performed on bulk RNA-seq data from DN and normal kidney samples to identify differentially expressed genes (DEGs) associated with the disease. Weighted gene co-expression network analysis (WGCNA) was utilized to reveal gene modules linked to DN, mitochondrial dysfunction, and senescence. The key genes were determined using multiple machine learning approaches, and their diagnostic value was verified using external datasets. At single-cell resolution, the cellular landscape of DN was explored and the distinct expression patterns across different cell types were explored. Key genes and markers associated with mitochondrial dysfunction and senescence were validated through single-cell RNA sequencing (scRNA-seq) data and in vitro high-glucose-induced HK-2 cell models. Finally, functional studies were conducted using Small interfering RNA (siRNA)-mediated gene knockdown to predict the biological roles of selected targets.ResultsWe identified 2,176 DEGs between DN and normal kidney tissues, among which 259 mitochondrial-related genes (MRGs) and 273 senescence-related genes (SRGs) were significantly enriched in inflammatory and metabolic pathways. WGCNA revealed DN-associated gene modules strongly linked to mitochondrial dysfunction and senescence. Through integrated machine learning, five hub genes—CLDN1, TYROBP, HDAC9, CASP3, and RCN1—were selected, with the support vector machine (SVM) model achieving high diagnostic accuracy. ScRNA-seq revealed 13 distinct kidney cell types, with proximal tubule (PT) cells emerging as key contributors to the signaling pathway associated with mitochondrial dysfunction and senescence. These transcriptomic findings were corroborated by functional assays, in which HDAC9 upregulation in high-glucose-stimulated HK-2 cells was accompanied by mitochondrial impairment and increased levels of p53, p21, p16, and senescence associated secretory phenotype (SASP) factors. Conversely, HDAC9 knockdown mitigated these effects, underscoring its pathogenic role in DN.ConclusionMitochondrial dysfunction and senescence-associated inflammation contribute to DN progression. The five identified hub genes demonstrate strong diagnostic potential, and HDAC9 is likely to be a potential therapeutic target for reducing mitochondrial injury, senescence, and inflammation in DN.

An Arabidopsis single-nucleus atlas decodes leaf senescence and nutrient allocation.

The cardiac conduction system (CCS) consists of distinct components including the sinoatrial node, atrioventricular node, His bundle, bundle branches, and Purkinje fibers. Despite an essential role for the CCS in heart development and function, the CCS has remained challenging to interrogate because of inherent obstacles including small cell numbers, large cell-type heterogeneity, complex anatomy, and difficulty in isolation. Single-cell RNA-sequencing allows for genome-wide analysis of gene expression at single-cell resolution. Assess the transcriptional landscape of the entire CCS at single-cell resolution by single-cell RNA-sequencing within the developing mouse heart. Wild-type, embryonic day 16.5 mouse hearts (n=6 per zone) were harvested and 3 zones of microdissection were isolated, including: Zone I-sinoatrial node region; Zone II-atrioventricular node/His region; and Zone III-bundle branch/Purkinje fiber region. Tissue was digested into single-cell suspensions, cells isolated, mRNA reverse transcribed, and barcoded before high-throughput sequencing and bioinformatics analyses. Single-cell RNA-sequencing was performed on over 22 000 cells, and all major cell types of the murine heart were successfully captured including bona fide clusters of cells consistent with each major component of the CCS. Unsupervised weighted gene coexpression network analysis led to the discovery of a host of novel CCS genes, a subset of which were validated using fluorescent in situ hybridization as well as whole-mount immunolabeling with volume imaging (iDISCO+) in 3 dimensions on intact mouse hearts. Further, subcluster analysis unveiled isolation of distinct CCS cell subtypes, including the clinically relevant but poorly characterized transitional cells that bridge the CCS and surrounding myocardium. Our study represents the first comprehensive assessment of the transcriptional profiles from the entire CCS at single-cell resolution and provides a characterization in the context of development and disease.

BackgroundEsophageal cancer is associated with high incidence and mortality worldwide. Differential expression genes (DEGs) and weighted gene co-expression network analysis (WGCNA) are important methods to screen the core genes as bioinformatics methods.MethodsThe DEGs and WGCNA were combined to screen the hub genes, and pathway enrichment analyses were performed on the hub module in the WGCNA. The CCNB1 was identified as the hub gene based on the intersection between DEGs and the greenyellow module in WGCNA. Expression levels and prognostic values of CCNB1 were verified in UALCAN, GEPIA2, HCMDB, Kaplan–Meier plotter, and TIMER databases.ResultsWe identified 1,044 DEGs from dataset GSE20347, 1,904 from GSE29001, and 2,722 from GSE111044, and 32 modules were revealed by WGCNA. The greenyellow module was identified as the hub module in the WGCNA. CCNB1 gene was identified as the hub gene, which was upregulated in tumour tissues. Moreover, esophageal cancer patients with higher expression of CCNB1 showed a worse prognosis. However, CCNB1 ‘might not play an important role in immune cell infiltration.ConclusionsBased on DEGs and key modules related to esophageal cancer, CCNB1 was identified as the hub gene, which offered novel insights into the development and treatment of esophageal cancer.

BackgroundLeptin has been implicated in the prognosis of sepsis, yet its mechanistic role remains unclear. This study aimed to develop leptin-associated diagnostic and prognostic models for sepsis and identify potential biomarkers using machine learning approaches.MethodsNon-negative matrix factorization (NMF) was used to identify leptin-related molecular subtypes of sepsis. Weighted gene co-expression network analysis (WGCNA) determined relevant gene modules and hub genes. Differentially expressed genes (DEGs) between sepsis patients and controls were intersected with WGCNA results to refine key genes. Based on these analyses, a prognostic classification model predicting 28-day mortality was developed using the Least Absolute Shrinkage and Selection Operator and Random Forest algorithms, while a time-to-event prognostic model was constructed with Random Survival Forest and Gradient Boosting Machine. Single-cell RNA sequencing was performed to assess expression patterns of core genes across immune cell types. Expression validation was conducted using qPCR and Western blotting.ResultsThree leptin-associated sepsis subtypes with distinct prognoses were identified. The pink and salmon modules from WGCNA were significantly associated with sepsis. Seventy core genes were selected from the DEGs and WGCNA intersection. The prognostic classification model and the time-to-event prognostic model demonstrated strong predictive performance in both the training and external validation cohorts. TFRC and PILRA were consistently highlighted through machine learning, single-cell data, and experimental validation as potential biomarkers.ConclusionWe established leptin-related prognostic models for sepsis using integrated machine learning. TFRC and PILRA may serve as promising biomarkers, offering insights into sepsis heterogeneity and clinical management.

Heterotopic ossification (HO) is the formation of abnormal mature lamellar bone in extra-skeletal sites, including soft tissues and joints, which result in high rates of disability. The understanding of the mechanism of HO is insufficient. The aim of this study was to explore biomarkers and pathological processes in HO+ samples. The gene expression profile GSE94683 was downloaded from the Gene Expression Omnibus database. Sixteen samples from nine HO- and seven HO+ subjects were analyzed. After data preprocessing, 3,529 genes were obtained for weighted gene co-expression network analysis. Highly correlated genes were divided into 13 modules. Finally, the cyan and purple modules were selected for further study. Gene ontology functional annotation and Kyoto Encyclopedia of Genes and Genomes pathway enrichment indicated that the cyan module was enriched in a variety of components, including protein binding, membrane, nucleoplasm, cytosol, poly(A) RNA binding, biosynthesis of antibiotics, carbon metabolism, endocytosis, citrate cycle, and metabolic pathways. In addition, the purple module was enriched in cytosol, mitochondrion, protein binding, structural constituent of ribosome, rRNA processing, oxidative phosphorylation, ribosome, and non-alcoholic fatty liver disease. Finally, 10 hub genes in the cyan module [actin related protein 3 (ACTR3), ADP ribosylation factor 4 (ARF4), progesterone receptor membrane component 1 (PGRMC1), ribosomal protein S23 (RPS23), mannose-6-phosphate receptor (M6PR), WD repeat domain 12 (WDR12), synaptosome associated protein 23 (SNAP23), actin related protein 2 (ACTR2), siah E3 ubiquitin protein ligase 1 (SIAH1), and glomulin (GLMN)] and 2 hub genes in the purple module [proteasome 20S subunit alpha 3 (PSMA3) and ribosomal protein S27 like (RPS27L)] were identified. Hub genes were validated through quantitative real-time polymerase chain reaction. In summary, 12 hub genes were identified in two modules that were associated with HO. These hub genes could provide new biomarkers, therapeutic ideas, and targets in HO.

Osteoarthritis (OA) is a major cause of pain, disability, and social burden in the elderly throughout the world. Although many studies focused on the molecular mechanism of OA, its etiology remains unclear. Therefore, more biomarkers need to be explored to help early diagnosis, clinical outcome measurement, and new therapeutic target development. Our study aimed to retrieve the potential hub genes of osteoarthritis (OA) by weighted gene co-expression network analysis (WGCNA) and assess their clinical utility for predicting OA. Here, we integrated WGCNA to identify novel OA susceptibility modules and hub genes. In this study, we first selected 477 and 834 DEGs in the GSE1919 and the GSE55235 databases, respectively, from the Gene Expression Omnibus (GEO) website. Genes with p-value<0.05 and | log2FC | > 1 were included in our analysis. Then, WGCNA was conducted to build a gene co-expression network, which filtered out the most relevant modules and screened out 23 overlapping WGCNA-derived hub genes. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses elucidated that these hub genes were associated with cell adhesion molecules pathway, leukocyte activation, and inflammatory response. In addition, we conducted the protein–protein interaction (PPI) network in 23 hub genes, and the top four upregulated hub genes were sorted out (CD4, SELL, ITGB2, and CD52). Moreover, our nomogram model showed good performance in predicting the risk of OA (C-index = 0.76), and this model proved to be efficient in diagnosis by ROC curves (AUC = 0.789). After that, a single-sample gene set enrichment (ssGSEA) analysis was performed to discover immune cell infiltration in OA. Finally, human primary synoviocytes and immunohistochemistry study of synovial tissues confirmed that those candidate genes were significantly upregulated in the OA groups compared with normal groups. We successfully constructed a co-expression network based on WGCNA and found out that OA-associated susceptibility modules and hub genes, which may provide further insight into the development of pre-symptomatic diagnosis, may contribute to understanding the molecular mechanism study of OA risk genes.