Single-cell WGCNA combined with transcriptome sequencing to study the molecular mechanisms of TRPM4-related genes in obesity with experimental validation

Single-cell WGCNA combined with transcriptome sequencing to study the molecular mechanisms of TRPM4-related genes in obesity with experimental validation | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Single-cell WGCNA combined with transcriptome sequencing to study the molecular mechanisms of TRPM4-related genes in obesity with experimental validation Yang Shen, Yanfen Li, Cijing Cai, Shenghong Lei This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8754472/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract While TRPM4 is crucial in immune and metabolic diseases, its role in obesity remains unclear. This study aimed to identify TRPM4-associated key genes in obesity. Obesity-related transcriptomic data were analyzed using single-cell RNA sequencing (scRNA-seq) to identify critical cell types, high-Dimensional Weighted Gene Co-expression Network Analysis (hdWGCNA) to determine module genes, and machine learning to screen key genes. A diagnostic nomogram was constructed, potential therapeutics were predicted, and reverse transcription quantitative PCR (RT-qPCR) validated gene expression in clinical samples. Endothelial cells were identified as critical, with ETV6, LAPTM5, and MCOLN3 highlighted as key genes. The nomogram demonstrated strong diagnostic efficacy, and compounds including cyclosporin A, retinoic acid, arsenious acid, and trichostatin A were predicted as potential modulators. RT-qPCR confirmed elevated ETV6 and LAPTM5 and reduced MCOLN3 expression in obesity (P < 0.001 and P < 0.01, respectively). Integrating bulk and scRNA-seq analyses identified endothelial cells and genes ETV6, LAPTM5, and MCOLN3 as key players, offering potential targets for TRPM4-related obesity interventions. Health sciences/Biomarkers Biological sciences/Computational biology and bioinformatics Health sciences/Diseases Biological sciences/Genetics Biological sciences/Molecular biology Obesity TRPM4 WGCNA Single-cell sequencing analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 1. Introduction Obesity is a long-lasting metabolic disorder marked by excessive fat buildup and abnormal weight gain, usually measured by BMI. In recent years, the worldwide rate of obesity has steadily risen, becoming a major public health concern globally[ 1 – 4 ]. At present, treatment strategies for obesity primarily encompass lifestyle modifications—including dietary regulation and physical activity—pharmacotherapy, and surgical interventions such as bariatric procedures. Although these approaches can partially enhance metabolic parameters and promote weight loss, they are often associated with such problems as inconsistent efficacy, poor patient adherence, high rates of recurrence, and potential complications. For instance, weight reduction medications may induce adverse effects, while surgical options, despite their effectiveness, carry risks of complications and entail substantial financial costs. Consequently, the current therapeutic modalities possess notable limitations in addressing this multifaceted disease[ 5 – 8 ]. The advent of molecular biology, genomics, transcriptomics, and metabolomics has enabled researchers to gain a deeper understanding of the mechanisms underlying obesity. Several regulatory factors have been identified, including adipokines, inflammatory factors, microbial metabolites, and non-coding RNAs. These biomarkers facilitate early diagnosis, classification, and personalized risk assessment, providing a basis for identifying therapeutic targets and improving individualized treatments. In recent years, a novel form of cell death, sodium overload-induced necrosis (NECSO), has gained research interest. It involves Na⁺ accumulation, causing osmotic imbalance, cell swelling, membrane rupture, and necrosis. This releases inflammatory factors that activate immune cells. NECSO is closely linked to TRPM4, a transient receptor potential channel family[ 9 ]. TRPM4 is a calcium-activated, non-selective cation channel that primarily transmits sodium ions, rather than calcium. It activates in response to increased intracellular calcium, leading to sodium influx, membrane depolarization, and indirectly influencing the opening of calcium channels and signaling[ 10 ]. TRPM4 is widely expressed in various tissues, including the myocardium, nervous system, and immune system, as well as in stem cells. It regulates proliferation, differentiation, migration, and secretion in various physiological and pathological contexts[ 11 , 12 ]. Recent studies have demonstrated that TRPM4 is present in tissues involved in metabolism, such as adipose tissue, liver, and immune cells. Its increased levels correlate with chronic inflammation and metabolic disruptions caused by obesity[ 13 – 15 ]. Research has shown that TRPM4 plays a crucial role in regulating calcium signaling and adipocyte differentiation in human adipose-derived stem cells (hASCs). TRPM4 controls adipogenesis in human adipose-derived stem cells (hASCs) by mediating histamine-triggered Ca²⁺ signaling. The process involves releasing Ca²⁺ from the endoplasmic reticulum and increasing Ca²⁺ influx by promoting the opening of L-type voltage-dependent Ca²⁺ channels (Cav1.2), which then activates adipogenic transcription factors and enzymes. Conversely, knocking down this gene decreases lipid droplet formation and downregulates adipocyte marker genes (e.g., C/EBPα), confirming its vital role in hASC adipogenesis. Furthermore, it may hold significance in the pathogenesis of obesity-related diseases[ 16 ]. Therefore, re-examining obesity's pathogenesis through cellular ion homeostasis and programmed cell death offers a new understanding and foundation for research on TRPM4-mediated sodium-dependent cell death in the context of obesity pathophysiology. Recently, single-cell analysis has become a pioneering tool in the life sciences, enabling the detection of cellular heterogeneity, rare cell populations, and changes in cell states with single-cell precision. Its potential is evident in tumour immunity, neurodevelopment, and inflammatory diseases[ 17 , 18 ]. Integrating bulk RNA-seq with single-cell RNA-seq helps understand gene expression trends and disease-related pathways at the systemic level. It also enables detailed analysis of the expression and regulation of key genes within cell subpopulations. This approach improves result stability and interpretability, identifying core cell types and driver genes involved at key disease nodes[ 19 – 21 ]. Our study analyzes obesity-related transcriptomic data from public repositories, using single-cell sequencing, hierarchical Weighted Gene Co-expression Network Analysis (hdWGCNA), and machine learning to identify and assess important genes linked to TRPM4. It explores their regulatory roles in obese tissue through gene-gene interaction networks and drug target prediction, aiming to uncover the molecular functions of TRPM4 in obesity. This integrated bioinformatics approach enables the identification of diagnostic and therapeutic targets, thereby supporting personalized interventions and precision treatment strategies. 2 Results 2.1 TRPM4 as a potential contributor to obesity: expression, pathways, and regulation In the GSE156906 dataset, TRPM4 demonstrated significantly elevated expression levels in the obese group relative to the control group (P < 0.01) (Fig. 1 a), suggesting its potential crucial role in the pathogenesis of obesity. Among the top 10 pathways identified through GSEA analysis of TRPM4, pathways with a clear link to obesity were observed, including "response of Eif2ak4/Gcn2 to amino acid deficiency," "SRP-dependent cotranslational protein targeting to membrane," and "cytoplasmic ribosomal proteins (Fig. 1 b)." Furthermore, TRPM4 was subject to regulation by numerous regulatory factors. Specifically, TRPM4 was predicted to be regulated by five miRNAs, such as hsa-miR-877-5p (Fig. 1 c). In addition, TRPM4 was found to be regulated by 31 TFs, such as SP1 (Fig. 1 d). 2.2 Endothelial cells were identified as a crucial cell type Before the implementation of QC procedures, GSE155960 comprised 83,537 cells and 18,498 genes. Following the QC procedures, after quality control, the dataset contained 83,149 cells and 18,498 genes ( Supplementary Fig. 1a-b ). After standard data processing, 2,000 HVGs were identified (Fig. 2 a). PCA analysis demonstrated that distinct clusters, visibly segregated in the principal component space, emerged from different samples within GSE155960 ( Supplementary Fig. 1c ), highlighting significant expression disparities among the samples. The top 15 PCs (P < 0.05) were subsequently chosen (Fig. 2 b-c). After that, UMAP was employed to visualize 20 distinct cell clusters (Fig. 2 d). A total of 10 cell types were identified based on specific marker genes: T cells (marked by CD3D, CD3E, and CD3G), preadipocytes (marked by PDGFRA, F3, and CEBPA), natural killer (NK) cells (marked by FGFBP2, FCGR3A, and GNLY), APCs (marked by FN1, CD34, and CD55), monocytes (marked by S100A8, S100A9, and FCN1), conventional dendritic cells (cDCs, marked by XCR1, FLT3, CCR7, and CD1E), macrophages (marked by C1QA, C1QB, and CD14), endothelial cells (marked by PECAM1, VWF, and IFI27), smooth muscle cells (marked by ACTA2, TAGLN, and MYH11), and B cells (marked by MS4A1, CD79B, and CD79A) (Fig. 2 e-f). Notably, in the CD45-positive cell samples, both the obese and control groups exhibited the highest proportion of T cells ( Supplementary Fig. 1d ). Additionally, among all annotated cell types, TRPM4 demonstrated a significant difference in expression exclusively in endothelial cells between the obese and control groups (P < 0.05), thereby designating endothelial cells as the crucial cell type (Fig. 2 g). 2.3 Analysis of cellular communication, pseudo-temporal trajectory, functional enrichment, and TF activity Further analysis was conducted on the intricate communication networks among the annotated cell types. Endothelial cells were found to exhibit a relatively high number of interactions with other annotated cells, while T cells and NK cells demonstrated stronger interaction weights (Fig. 3 a-b, Supplementary Fig. 1e ). Notably, the VEGFB-VEGFR1 and ANGPTL2-(ITGA5 + ITGB1) interaction pairs exhibited a high likelihood of occurrence during the pathway from preadipocytes to endothelial cells (Fig. 3 c). Subsequently, pseudo-temporal trajectory analysis was performed on endothelial cells. The diverging temporal trajectories of endothelial cells and their subclusters are depicted in Fig. 3 d, revealing that endothelial cells were categorized into five distinct states over time. Further investigation revealed that the genes contained in endothelial cells were clustered into three groups (Fig. 3 e). The genes in cluster 1 exhibited a trend of initially decreasing and then increasing during endothelial cell differentiation. The genes in cluster 2 showed an opposite pattern, rising first and then declining. The genes in cluster 3 demonstrated a more complex trend, decreasing initially, then increasing, and subsequently decreasing again throughout endothelial cell differentiation. Additionally, a pseudo-temporal trajectory analysis was performed on APCs, which revealed that APCs were differentiated into three distinct states over time (Fig. 3 f). The expression level of TRPM4 remained unchanged during APC differentiation (Fig. 3 g). Moreover, GSVA was employed to evaluate variations in gene enrichment pathways among annotated cells in obese and control samples. The obese group demonstrated the activation of six pathways, including "vesicle tethering to Golgi" and "chiasma assembly" (t > 2, P < 0.05), whereas 38 pathways, such as "glycerol transport" and "regulation of synaptic vesicle transport," were suppressed (t < 2, P < 0.05) (Fig. 3 h). Furthermore, the enrichment patterns of the 10 annotated cell types demonstrated considerable variation across different biological pathways (Fig. 3 i). For example, the process of "inner ear receptor cell fate commitment" showed significant enrichment in endothelial cells but lower enrichment in other cell types. Subsequent predictions of TF activities revealed distinct differences among the annotated cell types (Fig. 3 j). For instance, the analysis indicated that ATOH8 displayed high expression levels in endothelial cells and comparatively low expression in other annotated cell types. 2.4 Identification of 384 key module genes via hdWGCNA Key molecular characteristics of endothelial cells were discerned through hdWGCNA. Specifically, during co-expression network construction, a soft-thresholding power of 6 was selected ( Fig. 4 a ) to construct an unweighted endothelial cell network, as this value achieved a scale-free topology fit index above 0.8 and ensured optimal connectivity. This was followed by the identification of 11 distinct co-expression modules through hierarchical clustering combined with dynamic tree cuttinge (Fig. 4 b). The correlation among the 11 modules is shown in Fig. 4 c and Supplementary Table 3 . Moreover, genes within the black module demonstrated significant upregulation in the obesity group, whereas genes within the purple and magenta module demonstrated significant downregulation in the obesity group (adj. P < 0.05) (Fig. 4 d). Consequently, these three modules were designated as key modules, comprising 186 (black), 78 (purple), and 120 (magenta) genes, for a total of 384 genes. 2.5 Recognition and related functional exploration of candidate genes linked to TRPM4 in obesity Analysis of GSE156906 revealed 2,189 DEGs when comparing the obese and control groups. Among these DEGs, 1,569 genes exhibited up-regulation, while 620 genes showed down-regulation in the obese group (Fig. 5 a-b). An intersection analysis between the DEGs and key module genes yielded 17 candidate genes (Fig. 5 c). Functional analysis revealed that the 17 candidate genes were significantly enriched for 223 GO terms (P < 0.05), comprising 165 BPs, 20 CCs, and 38 MFs ( Supplementary Table 4 ). Specifically, the top 10 significantly enriched pathways (P < 0.05) for BPs, CCs and MFs included "cellular response to tumor necrosis factor", "late endosome membrane", and "DNA-binding transcription activator activity" (Fig. 5 d). After excluding four candidate genes encoding discrete proteins, a PPI network was constructed with the remaining 13 proteins (e.g., ETV6) (Fig. 5 e). 2.6 ETV6, LAPTM5, and MCOLN3 were identified as key genes Among the 17 candidate genes, LASSO analysis yielded the lowest prediction error when lambda was set to lambda. min = 0.03331 (Fig. 6 a). At this specific lambda value, seven genes exhibited non-zero regression coefficients, and these genes (LAPTM5, SMIM3, GJA1, LRMDA, ETV6, SLC16A3, and MCOLN3) were subsequently identified as LASSO characteristic genes. Moreover, the SVM-RFE model achieved peak predictive accuracy when 16 variables were selected (Fig. 6 b). The following 16 genes were consequently identified as SVM-RFE characteristic genes: MCOLN3, NR4A1, SLC16A3, FABP4, ETV6, LAPTM5, RGS19, CDKN1C, GJA1, PRDX6, SMIM3, LRMDA, KLF9, CEBPD, CXCL3, and RHOB. Furthermore, the Boruta algorithm pinpointed 12 genes that exhibited importance scores exceeding those of ShadowMax, thereby designating them as Boruta characteristic genes (Fig. 6 c). These genes included FABP4, ASAH1, PRDX6, RHOB, LAPTM5, SMIM3, GJA1, KLF9, LRMDA, ETV6, MCOLN3, and RGS19. Subsequently, the overlap among the characteristic genes identified by the three aforementioned machine learning algorithms was determined (Fig. 6 d), yielding six common characteristic genes: LAPTM5, SMIM3, GJA1, LRMDA, ETV6, and MCOLN3. In datasets GSE156906 and GSE25401, ETV6, LAPTM5, and MCOLN3 exhibited AUC values exceeding 0.7 (Fig. 6 e-f). Moreover, within GSE156906 and GSE25401, ETV6 and LAPTM5 demonstrated markedly elevated expression levels in the obese group (P < 0.0001), while MCOLN3 levels were significantly reduced (P < 0.0001) ( Figure. 6g-h ). The results indicated that ETV6, LAPTM5, and MCOLN3 exhibited good stability and high diagnostic potential, rendering them promising candidates for further exploration as key genes associated with obesity. 2.7 Nomogram demonstrated favorable performance in assessing the diagnosis of obesity A nomogram was constructed to evaluate the diagnostic effectiveness of ETV6, LAPTM5, and MCOLN3 for identifying obesity. According to the nomogram, a higher total score derived from these three genes was associated with a greater likelihood of obesity onset (Fig. 7 a). The calibration curve demonstrated satisfactory performance by passing the HL goodness-of-fit test, yielding a P value of 0.252 (Fig. 7 b). This result suggested no significant discrepancy between predicted and observed outcomes, thus validating the nomogram's high accuracy. The DCA findings revealed that the nomogram model demonstrated a notably higher net benefit than utilizing individual key genes alone (Fig. 7 c), suggesting that it could help clinicians to make more accurate early diagnoses of obesity. 2.8 Subcellular localization, GGI network establishment, and drug/compound interactions associated with ETV6, LAPTM5, and MCOLN3 Subsequent analysis was performed on the structures of ETV6, LAPTM5, and MCOLN3. In particular, ETV6 and MCOLN3 were mainly localized within the nucleus, whereas LAPTM5 exhibited predominant cytoplasmic localization (Fig. 8 a). The GGI network comprised 23 genes, encompassing ETV6, LAPTM5, and MCOLN3, in addition to 20 other related genes (Fig. 8 b). MCOLN3, along with MCOLN1 and MCOLN2, collectively participated in multiple biological pathways, such as "calcium ion transmembrane transporter activity" and "divalent inorganic cation transmembrane transporter activity." As predicted by DsigDB, ETV6, LAPTM5, and MCOLN3 targeted 51 drugs/compounds (Fig. 8 c). Of these, cyclosporin A, retinoic acid and aflatoxin B1 were targeted by both LAPTM5 and MCOLN3, arsenious acid and formaldehyde were targeted by both ETV6 and LAPTM5, and trichostatin A was targeted by both ETV6 and MCOLN3. 2.9 RT-qPCR analysis of ETV6, LAPTM5, and MCOLN3 expression Consistent with our bioinformatic predictions, RT-qPCR validation confirmed that mRNA levels of ETV6 and LAPTM5 were significantly upregulated in the obese group (P < 0.001; Fig. 9 a-b), while MCOLN3 expression was notably downregulated (P < 0.01; Fig. 9 c). The RT-qPCR results for ETV6, LAPTM5, and MCOLN3 were in agreement with their predicted expression patterns from the GSE156906 and GSE25401 datasets, collectively confirming the accuracy of our bioinformatic predictions. 3. Discussion Obesity is a complex metabolic disorder closely connected to chronic inflammation, immune imbalance, and multiple molecular signaling pathways. Research indicates that TRPM4, a member of the transient receptor potential channel family, plays a crucial regulatory role in various immunological and metabolic diseases, suggesting that it may also contribute to the development and progression of obesity[ 13 , 22 , 23 ]. To investigate the role of TRPM4 in obesity, this study combined bulk and single-cell transcriptomics data, revealing significant differences in TRPM4 expression between endothelial cells in obese and control groups. Endothelial cells were identified as a key cell type. hdWGCNA was employed to identify crucial module genes associated with endothelial cells, followed by screening using machine learning, ROC analysis, and validation. This process identified three essential genes: ETV6, LAPTM5, and MCOLN3. These genes could enhance understanding of TRPM4 regulatory networks in obesity and serve as potential targets for diagnosis and treatment. This study, for the first time, indicates an association between TRPM4 and the occurrence and development of obesity, offering a new perspective for research on TRPM4-related factors in obesity. This discovery deepens our understanding of the molecular mechanisms behind obesity and offers valuable theoretical insights and practical hints for creating new diagnostic markers and intervention strategies focused on these key genes. scRNA-seq analysis revealed significantly altered TRPM4 expression in endothelial cells of obese subjects compared to controls. Endothelial cells, identified as key regulators, provide insights into their role in obesity and TRPM4-related functions. Endothelial cell dysfunction is a crucial factor in the pathophysiology of obesity. In obesity, inflammatory factors (TNF-α, IL-6) and adipokines secreted by adipose tissue suppress the expression of insulin receptor substrate (Irs) in endothelial cells, thereby impairing insulin-induced vasodilation and glucose uptake, which leads to systemic insulin resistance[ 24 , 25 ]. Meanwhile, endothelial cell dysfunction is characterized by reduced vasodilatory capacity (decreased eNOS activity, lower NO production), decreased vascular density, and abnormal remodeling. These changes actively contribute to the development and persistence of metabolic disorders by impairing nutrient transport and oxygen supply to tissues[ 26 ]. Studies have shown that TRPM4 regulates calcium homeostasis, vascular barrier integrity, vascular tension, and inflammatory signaling pathways in endothelial cells[ 27 – 29 ]. Additionally, TRPM4 also affects vascular permeability and blood perfusion, indirectly regulating the metabolic supply of adjacent tissues[ 30 , 31 ]. Therefore, the results of this study suggest that RPM4 may participate in the process of obesity-related endothelial dysfunction and metabolic disorders by regulating the calcium signal transduction and barrier function of endothelial cells. This provides an important theoretical basis for understanding the mechanism of sodium overload-induced necrosis (NECSO) mediated by TRPM4 in the pathophysiology of obesity. This study thoroughly explores the potential mechanistic roles of three key genes—ETV6, LAPTM5, and MCOLN3. First, it clarifies the biological functions of these three genes under normal physiological conditions. Then, by integrating previous related research and analyzing the expression differences between obese individuals and the control group in this study, the expression patterns and possible roles of these genes in obesity and related metabolic diseases are thoroughly explored. Notably, some of these genes have not been previously reported in obesity. Additionally, the results of the RT-qPCR experiment can be combined for discussion: ETV6 and LAPTM5 were significantly upregulated in the obese group, while MCOLN3 was downregulated considerably (P < 0.01). This trend in expressions was consistent with the database's prediction results. The expression patterns shown in the experimental data match those in the database analysis, indicating these genes have stable expression changes under specific conditions. This finding not only supports the reliability of bioinformatics predictions at the experimental level but also further strengthens the credibility and scientific basis of this study’s conclusions[ 32 – 34 ]. ETV6 (ETS variant transcription factor 6) is a type of transcription repressor within the ETS transcription factor family. It plays a crucial regulatory role in the hematopoietic system[ 35 , 36 ].Additionally, ETV6 plays a vital role in the maintenance and differentiation of hematopoietic stem cells. Its mutation or abnormal expression has been widely linked to various hematologic malignancies and myelodysplastic syndromes (MDS)[ 37 , 38 ]. In addition to its role in the blood system, recent studies have indicated that ETV6 may also serve a significant regulatory function in metabolic disorders. Notably, a study investigating the epigenetic landscape of human subcutaneous and visceral adipose tissue observed an inverse correlation between ETV6 promoter methylation and gene expression. This phenomenon was consistently validated across two independent population cohorts[ 39 ]. These findings suggest that ETV6 may be involved in adipogenesis, lipid metabolism, and the growth and function of adipose tissue through epigenetic mechanisms, potentially contributing to the development of obesity. In summary, ETV6 plays a crucial role in blood development and also regulates fat tissue metabolism. This suggests that it could serve as a potential link between metabolic disorders and hematopoietic system diseases, warranting further research. LAPTM5 (Lysosomal-Associated Protein Transmembrane 5) is a protein associated with lysosomes, featuring multiple transmembrane domains. It is primarily expressed in hematopoietic cells, especially in B cells and macrophages. LAPTM5 is situated on the lysosomal membrane and is involved in lysosome-mediated protein degradation, signal transduction, and immune regulation. It plays a vital role in maintaining cellular homeostasis and immune responses[ 40 – 43 ]. Recent studies have demonstrated that the role of LAPTM5 has expanded from traditional immune regulation to the area of metabolic diseases. Gene expression analysis revealed that in conditions of obesity and insulin resistance, LAPTM5 was upregulated, alongside several genes related to lysosomal function, suggesting that it may play a role in regulating lipid metabolism and the abnormal lysosomal functions associated with obesity[ 44 ]. The bioinformatics research further identified LAPTM5 as a key hub gene in the obesity-related chronic inflammation network. Its expression level is closely connected to the infiltration of M1 macrophages, suggesting its potential involvement in regulating the immune response following bariatric surgery[ 45 ]. Studies on childhood obesity have also shown that LAPTM5 is involved in the process of polyamine metabolism disorder, leading to an imbalance in the immune microenvironment through the lysosome pathway, further supporting its significant role in the development of obesity[ 46 ]. Additionally, animal experiments have provided indirect evidence for the metabolic regulatory function of LAPTM5: liver cell-specific knockout of Laptm5 worsens non-alcoholic fatty liver disease (NAFLD), while overexpression of LAPTM5 can significantly reduce liver inflammation and lipid buildup by promoting the lysosomal degradation of CDC42 and inhibiting the MAPK signaling pathway, emphasizing its protective role in lipid metabolism[ 47 ]. In conclusion, LAPTM5, as a key molecule linking the immune system and metabolic disorders, not only performs the traditional lysosomal regulatory function in immune cells but also shows extensive regulatory potential in obesity and related metabolic diseases. Its value as a diagnostic or therapeutic target is worth in-depth exploration. MCOLN3 (Mucolipin 3), also known as TRPML3 (Transient Receptor Potential Mucolipin 3), belongs to the mucolipin subfamily of the TRP (transient receptor potential) ion channel family. MCOLN3 is a cation channel situated in the inner membrane system of cells, widely present on the membranes of early endosomes, late endosomes, lysosomes, and other organelles. It primarily regulates Ca²⁺ homeostasis within cells and participates in various biological processes, such as vesicle transport, regulation of the endocytosis pathway, and maintenance of lysosomal function[ 48 , 49 ]. The activity of MCOLN3 in the lysosome-endosome system is essential for the cell's response to osmotic pressure changes, regulation of vesicle acidification, and facilitation of membrane fusion and other cellular stress and transport processes[ 50 , 51 ]. Although MCOLN3 is crucial for organelle function and calcium signaling regulation, there has been no systematic research on its specific roles in obesity or metabolism-related conditions, such as abnormal fat processing or energy balance issues. Our PCR results demonstrated significantly lower MCOLN3 expression in the obese group compared to controls. This downregulation suggests a potential role for MCOLN3 in obesity pathogenesis, possibly mediated through its known functions in regulating intracellular calcium levels and vesicle transport. The reduced expression of MCOLN3 could impair the lysosome-endosome system, which would affect lipid metabolism and cellular energy regulation. However, further experimental validation is necessary to confirm this finding mechanism. This finding opens up new avenues for investigating the functions of MCOLN3 within the energy metabolism network. It may broaden the scope of research into the roles of TRPML family channels in metabolic diseases. It also suggests potential new interaction mechanisms between calcium channels and lipid metabolism in the endosomal system, providing valuable theoretical insights and promising research opportunities. Based on the prediction results of drugs and compounds, explore potential intervention targets and mechanisms of key genes: Both LAPTM5 and MCOLN3 target cyclosporin A and retinoic acid. Additionally, ETV6 and MCOLN3 target trichostatin A. These findings suggest that these compounds may participate in regulating obesity-related immune and metabolic pathways through a multi-target synergistic mechanism and could have notable intervention potential. Cyclosporin A (Cyclosporin A) is a well-known calcineurin inhibitor widely used in immunosuppressive therapy after organ transplantation. Recent research has shown that it also plays a significant role in non-immune systems[ 52 ]. Cyclosporine A influences the inflammatory response and metabolic balance of adipose tissue by disrupting mitochondrial function[ 53 , 54 ]. In this study, the drug was predicted to be a targeted molecule that acts on both LAPTM5 and MCOLN3, suggesting that cyclosporine A may participate in regulating adipose tissue inflammation and metabolic homeostasis through pathways involving LAPTM5 and MCOLN3. Retinoic acid (RA) is a metabolite of vitamin A that can regulate gene expression by activating nuclear receptors RAR/RXR. Retinoic acid has been extensively studied in the processes of adipocyte fate determination, brown fat activation, and adipose tissue remodeling[ 55 – 57 ]. We found that LAPTM5 and MCOLN3 are also potential targets for retinoic acid, indicating that these two genes may influence retinoic acid-driven lipid metabolism processes. Trichostatin A (TSA) is a widely used histone deacetylase inhibitor (HDACi) that significantly influences chromatin structure and transcriptional activity by inhibiting the functions of Class I and Class II HDACs[ 58 ]. TSA may also influence the expression of genes involved in adipogenesis and metabolism[ 59 , 60 ]. In this study, TSA was predicted to target both ETV6 and MCOLN3 simultaneously, suggesting that these two may be involved in regulating the influence of epigenetic mechanisms on obesity. LAPTM5 is the gene among the three key genes that has the most drug interactions. It is also associated with metabolic regulatory functions, such as cyclosporine A and retinoic acid. This suggests that it may act as a central hub in the obesity-related immune and metabolic regulation network. This finding highlights the importance of LAPTM5 as a potential target and further confirms its role in regulating multi-pathway and multi-drug response networks. It is worth noting that these drug-gene interaction predictions are primarily based on analyzing gene expression correlations and may involve various mechanisms, including direct binding, transcriptional regulation, and regulation of signaling pathways. The specific molecular interaction patterns still need further experimental verification to be confirmed. Moving forward, a drug-gene-phenotype functional verification system could be developed around LAPTM5 to evaluate its translational potential in the treatment of obesity thoroughly. This study investigates the potential molecular mechanisms behind obesity. It starts with single-cell transcriptomic analysis to identify endothelial cells as key regulatory types in adipose tissue, guiding future efforts to target specific cell populations. Further research, combining various machine learning algorithms with bioinformatics methods, carefully screens and identifies three critical genes—ETV6, LAPTM5, and MCOLN3. These genes support and extend the functional regulatory network centered on TRPM4, offering new insights into the complex interactions of obesity-related signaling pathways. Although this research provides innovative bioinformatic evidence for the development of obesity, some limitations remain. For example, some analyses rely on data from public databases, where sample size and data quality may limit the applicability of the findings. Additionally, these key genes need validation through larger clinical samples, animal models, or in vitro functional tests to better understand their functions at the tissue level and within the broader metabolic framework. Moving forward, we will continue to explore the expression patterns and regulatory mechanisms of ETV6, LAPTM5, and MCOLN3 across various obesity types. Our goal is to thoroughly understand their roles in TRPM4-related signaling pathways and the immunometabolism disruptions associated with obesity. This research aims to strengthen the scientific foundation and offer new directions for early diagnosis, biomarker identification, and targeted therapies for obesity. 4. Methods 4.1 Data acquisition The obese datasets employed in the study included GSE156906 (platform: GPL24676), GSE25401 (platform: GPL6244), and GSE155960 (platform: GPL24676), all of which were retrieved from the Gene Expression Omnibus (GEO) database ( https://www.ncbi.nlm.nih.gov/geo/ ). The analysis utilized subcutaneous adipose tissue samples from two cohorts: the training set (GSE156906) included 25 obese patients and 14 healthy controls, and the validation set (GSE25401) comprised 30 obese patients and 26 healthy controls. The single-cell RNA sequencing (scRNA-seq) dataset (GSE155960) included matrix files of CD45-positive and CD45-negative cells derived from subcutaneous adipose tissue, with six samples each from the obese and control groups. 4.2 Evaluation of TRPM4 expression and gene set enrichment analysis (GSEA) TRPM4 expression was assessed between obese and control samples in the GSE156906 dataset, employing the Wilcoxon test (P < 0.05). GSEA was performed to investigate TRPM4-associated signaling pathways in obesity, utilizing the "c2.cp.v2024.1.Hs.symbols.gmt" collection from the Molecular Signatures Database (MSigDB) ( https://www.gsea-msigdb.org/ ) as the gene set background. Using the cor function from the R stats package (v 4.3.3; https://www.R-project.org/ ), we computed Spearman correlations between TRPM4 and all other genes in GSE156906, then ranked the genes in descending order to create a TRPM4-associated gene list. Thereafter, we performed GSEA for TRPM4 using the clusterProfiler package (v 4.8.3)[ 61 ], employing significance criteria of |NES| > 1 and adjusted P-value < 0.05. 4.3 construction of molecular regulatory networks The identification of microRNAs (miRNAs) and transcription factors (TFs) regulating TRPM4 enhanced the understanding of the gene's regulatory logic and facilitated the discovery of potential key regulators. Specifically, the StarBase v3.0 database ( https://starbase.sysu.edu.cn/ ) was searched for miRNAs associated with TRPM4, and these were filtered based on the criterion of "pancancerNum > 6". Additionally, TFs regulating TRPM4 were retrieved from the Cistrome database ( http://cistrome.org/ ) based on the criterion of a "regulatory potential score" exceeding 0.7. The miRNA-TRPM4 and TF-TRPM4 regulatory networks were constructed and visualized using Cytoscape software (v 3.8.0)[ 62 ]. 4.4 Processing of scRNA-seq data and identification of crucial cell types Quality control (QC) procedures for dataset GSE155960 were carried out utilizing the "PercentageFeatureSet function" from the Seurat package (v 5.0.1)[ 63 ]. Following this filtering, all downstream analyses were performed exclusively with the functions provided in this package. Cells with an nFeature_RNA (number of detected genes) below 200 or above 4,000 were excluded. Cells with an nCount_RNA (total number of RNA counts per cell) below 500 or above 20,000 were also filtered out. Subsequently, the "NormalizeData function" was employed to standardize the data. Following this, the top 2,000 highly variable genes (HVGs) were identified utilizing the variance-stabilizing transformation (vst), which was implemented in the "FindVariableFeatures function". The top 10 genes were labeled utilizing the "LabelPoints function" for visualization. After that, the data were normalized via the "Scale Data function". Subsequently, principal component analysis (PCA) was applied to reduce data dimensionality, and the Harmony method was applied for batch effect correction. PCA facilitated the embedding of cells into a low-dimensional space, whereas Harmony efficiently removed batch-related discrepancies through an iterative optimization algorithm conducted within the PCA-transformed space. The "JackStraw function" was used to identify the most significant principal components (PCs) by selecting those with a higher number of genes with low P values (P < 0.05). The "ElbowPlot function" was employed to generate a scree plot, where inflection points served as indicators to identify the optimal number of PCs for subsequent analysis. Thereafter, unsupervised clustering was conducted with a resolution parameter of 0.5 to identify the distinct cell populations. The Uniform Manifold Approximation and Projection (UMAP) method allowed for unsupervised clustering and provided an unbiased visualization of cell clusters in two dimensions. Additionally, to identify cluster-specific marker genes, we employed the "FindAllMarkers" function, applying thresholds of |log2 fold change (FC)| > 1 and an adjusted P-value < 0.05. The SingleR package (v 2.2.0)[ 64 ] and the CellMarker database ( http://xteam.xbio.top/CellMarker/ ) were used as the primary references for cell type annotation, while the HumanPrimaryCellAtlasData, BlueprintEncodeData, and ImmuneCellExpressionData were used as supplementary references. Based on the clustering results and concerning relevant literature[ 65 – 67 ], cell subclusters were annotated with corresponding cell types. The distribution of annotated cell types in the GSE155960 dataset was visualized using stacked plots generated with ggplot2 (v 3.5.1)[ 68 ]. Subsequently, crucial cell types were identified as those exhibiting significant differences (P < 0.05) in TRPM4 expression between obese and control groups, as determined by the Wilcoxon test. 4.5 Cellular communication and pseudo-temporal trajectory analyses The CellChat package (v 1.6.1)[ 69 ] was employed to analyse patterns of cell-to-cell communication, facilitating the quantification and characterization of interactions between crucial cell types and other annotated cell types. The reference human ligand-receptor database, CellChatDB ( http://www.cellchat.org/ ), was employed in the analysis to evaluate human intercellular communication. Additionally, the same package was employed to analyse ligand-receptor interaction pairs across these annotated cell types. Furthermore, dimensionality reduction and clustering were performed on the crucial cell types again, utilizing the same methodologies as those previously employed for processing the scRNA-seq data. Moreover, the differentiation trajectories of the crucial cell types were reconstructed using the Monocle package (v 2.28.0)[ 70 ], and the trajectories for each crucial cell type were depicted utilizing the "DDRTree function" within the same package. To visualize the dynamic expression patterns of significant genes in crucial cell types, we generated an expression heatmap across pseudotime using the plot_pseudo_time_heatmap function. Adipocyte progenitor cells (APCs) play a pivotal role in obesity. Consequently, dimensionality reduction and clustering, along with pseudo-temporal trajectory analyses, were conducted on APCs. The dynamic expression pattern of TRPM4 during APC differentiation was delineated utilizing the monocle package (v 2.28.0). 4.6 Gene set variation analysis (GSVA) and prediction of TF activity A comprehensive assessment of the signaling pathways implicated in both obese and control groups was carried out by employing GSVA on all samples from GSE155960. The GSVA package (v 1.53.28)[ 71 ] was utilized to analyze all samples in GSE155960, calculating enrichment scores for each sample across gene sets, which subsequently facilitated Gene Ontology (GO) enrichment analysis. Subsequently, we performed differential analysis between the obese and control groups with the limma package (v 3.56.2)[ 72 ] (|t| > 2, P < 0.05). Additionally, the enrichment of these pathways within each annotated cell type was analyzed. To predict the transcription factors activated under obese conditions, the activities of gene regulatory networks (GRNs) and TFs were identified utilizing the SCENIC Python workflow (v 0.9.1)[ 73 ] with default parameters ( https://github.com/aertslab/pySCENIC ). The list of human TF genes was compiled from the same source. The binary matrix was used to identify activated TFs, and the specificity of key TFs in different cell types was analysed alongside the regulon specificity score (RSS) values. 4.7 High-dimensional weighted gene co-expression network analysis (hdWGCNA) The hdWGCNA was conducted to identify significant genes linked to critical cell types in the samples derived from GSE155960. The "MetacellsByGroups function" of the hdWGCNA package (v 0.4.0)[ 74 ] was employed to construct metacells for critical cell types within each sample from GSE155960. During this process, the k-nearest neighbour parameter was set to 25, and the maximum number of cells that could be shared between any two metacells was limited to 10. Subsequently, following the standard hdWGCNA analysis pipeline, the "TestSoftPowers function" was employed to evaluate and select an appropriate soft power threshold. The minimum soft-thresholding power yielding a scale-free topology fit index ≧ 0.8 was selected. Following this determination, a weighted co-expression network was built using the "ConstructNetwork" function. This function was then applied to perform topological overlap matrix (TOM)-based hierarchical clustering on the clustered samples, enabling a clustering dendrogram to be established for module identification. The "ModuleCorrelogram function" was employed to calculate correlations between different modules (|correlation coefficient (cor)| > 0.3 and P < 0.05). Furthermore, the "FindDMEs function" was employed to conduct the Wilcoxon test on genes within each module in both the obese and control groups (adj.P < 0.05). We defined key modules as those exhibiting significantly distinct expression profiles, and termed their constituent genes as key module genes. 4.8 Analysis of differential gene expression Differentially expressed genes (DEGs) between obese and control groups in GSE156906 were identified utilizing the DESeq2 package (v 1.40.2)[ 75 ], with the following criteria applied: |log₂ FC| > 0.5 and P < 0.05. Furthermore, A volcano plot visualizing the differentially expressed genes (DEGs) was generated using the ggplot2 package (v 3.5.1), and a heatmap illustrating DEGs was generated with the ComplexHeatmap package (v 2.16.0)[ 76 ]. 4.9 Identification and related functional analysis of candidate genes The ggvenn package (v 0.1.10)[ 77 ] was used to visualize the intersection between DEGs and key module genes, aiming to identify candidate genes potentially linked to TRPM4 in crucial obesity-related cell types. Following this, the clusterProfiler package (v 4.8.3) was used to perform Gene Ontology (GO) enrichment analysis on the candidate genes (P < 0.05), thereby clarifying their molecular functions (MFs), cellular components (CCs), and biological processes (BPs). Subsequently, the genes enriched in each BP, CC, and MF category were ranked in descending order according to their counts. We displayed the top 10 most significantly enriched pathways for BP, CC, and MF, respectively. A protein-protein interaction (PPI) network was constructed using the Retrieval of Interacting Genes (STRING) database ( https://www.string-db.org ; score > 0.15) to explore interactions among candidate genes and was then visualized in Cytoscape (v 3.8.0). 4.10 Discernment of key genes through machine learning, receiver operating characteristic (ROC) analysis, and expression validation All samples within GSE156906 were subjected to analysis utilizing three different machine learning approaches, which leveraged candidate genes to pinpoint characteristic genes linked to obesity. Initially, Least Absolute Shrinkage and Selection Operator (LASSO) regression was employed to minimize redundant information and eliminate irrelevant variables, ultimately shrinking the coefficients of unimportant variables to zero while retaining only a small number of significant variables with non-zero coefficients. We applied LASSO regression to the candidate genes via the "cv.glmnet" function in the glmnet package (v 4.1.8)[ 78 ], with five-fold cross-validation applied. Genes with non-zero coefficients at the optimal lambda (lambda.min), which yielded the lowest cross-validation error, were selected as the LASSO feature set. Concurrently, we employed Support Vector Machine-Recursive Feature Elimination (SVM-RFE), a feature selection technique that sequentially removes features with the least importance based on the SVM's maximum margin principle. It trains the model iteratively, ranks the features based on their scores, and progressively eliminates the lowest-scoring features until the desired set is selected. The "svmRFE function" from the e1071 package (v 1.7.16)[ 79 ] was employed to construct an SVM-RFE model. This model utilized candidate genes as feature variables and implemented five-fold cross-validation. Additionally, the parameter halve.above was set to 100 to evaluate the feature importance and ranking of each candidate gene. During the model iteration process, the classification error rate and accuracy for each round of feature combinations were recorded. The feature combination corresponding to the minimum error rate was selected as the final optimal gene set, which was defined as the SVM-RFE characteristic genes. The core of the Boruta feature selection method can be distilled into two primary steps: constructing shadow features and employing a random forest-based voting mechanism. If an original feature is significantly more important than the shadow features, it is classified as important; otherwise, it is regarded as unimportant. Specifically, the Boruta algorithm was employed for feature selection using the Boruta package (v 8.0.0)[ 80 ], was employed for feature selection from candidate genes. Features with importance scores exceeding shadowMax were categorized as "important" and subsequently labeled as Boruta characteristic genes. Subsequently, the ggvenn package (v 0.1.10) was utilized to identify genes that overlapped among those selected by the LASSO, SVM-RFE, and Boruta algorithms. These overlapping genes were then designated as common characteristic genes. Furthermore, the pROC package (v 1.18.5)[ 81 ] was employed to conduct ROC analysis on common characteristic genes, and the area under the curve (AUC) values were calculated. We defined genes with AUC values exceeding 0.7, indicating good obese-control discriminative ability, as candidate key genes. Additionally, the expression profiles of the candidate key genes were compared between obese and control samples in the GSE156906 and GSE25401 datasets using the Wilcoxon test. The candidate key genes showing notably differential expression between obese and control samples (P < 0.05) with a consistent trend in both GSE156906 and GSE25401 were identified as key genes. 4.11 Establishment and assessment of a nomogram A nomogram was constructed utilizing the rms package (v 6.8.1)[ 82 ] to evaluate the predictive probability of obesity onset based on key genes in GSE156906. A nomogram was developed where each key gene contributed a specific point value, and the summed score predicted obesity risk. The model's calibration was quantified by a curve generated with the regplot package (v 1.1)[ 83 ] and confirmed by a non-significant Hosmer-Lemeshow test statistic (P > 0.05). We employed decision curve analysis (DCA) with the ggDCA package (v 1.1)[ 84 ] to evaluate the clinical net benefit of the nomogram. 4.12 Subcellular localization and gene-gene interaction (GGI) network construction Subcellular localization analysis was carried out to acquire a deeper understanding of the roles played by the key genes. In particular, FASTA-formatted sequence files corresponding to the key genes were retrieved from the National Center for Biotechnology Information (NCBI) database ( https://www.ncbi.nlm.nih.gov/gene/ ). The subcellular localization of the key genes was predicted using the mRNALocater database ( http://bio-bigdata.cn/mRNALocater/ ). To further elucidate their functional relationships and shared biological roles, a GGI network was constructed via the GeneMANIA database ( http://www.genemania.org/ ). 4.13 Drugs/compounds prediction The identification of potential drugs and compounds capable of targeting specific key genes was facilitated by utilizing the Drug Signatures Database (DSigDB) ( https://dsigdb.tanlab.org/DSigDBv1.0/ ). Subsequently, a drug/compound-key gene interaction network was generated using Cytoscape software (v 3.8.0). 4.14 Reverse transcription-quantitative PCR (RT-qPCR) To experimentally validate the differential expression of key genes between obese and control samples, RT-qPCR analysis was performed using subcutaneous adipose tissue samples obtained from five obese patients, alongside five samples from control subjects, sourced from shenzhen qianHai taikang hospital. This research received approval from the medical ethics committee of Shenzhen Qianhai Taikang Hospital and all participating patients provided their informed consent by signing the relevant form. Total RNA was isolated from whole blood samples employing the TRIzol kit (Vazyme Biotech Co., Ltd., Catalog No. R401-01, Nanjing, China), Total RNA was reverse-transcribed into cDNA using the HP All-in-one qRT Master Mix II (Yungen Biotechnology, Cat. No. 24Y0124) according to the manufacturer's protocol. The resulting cDNA was diluted 5- to 20-fold with RNase/DNase-free ddH 2 O, and a 10 µL qPCR reaction mixture was prepared containing 3 µL of diluted cDNA, 5 µL of 2× Universal Blue SYBR Green qPCR Master Mix, and 1 µL each of forward and reverse primers (10 µM each). Amplification was performed for 40 cycles on a CFX Connect real-time PCR system (BIO-RAD, Cat. No. XLFZ006) using a program that omitted the pre-denaturation step. The program details are outlined in Supplementary Table 1 . The primer sequences for key genes are listed in Table S2. GAPDH served as the reference gene, and relative expression levels were determined using the 2 −△△CT method. 4.15 Statistical analysis Statistical analyses were primarily conducted in R (v 4.3.3), employing the Wilcoxon test for group comparisons, while RT-qPCR Ct values were analyzed by unpaired two-tailed t-tests in GraphPad Prism (v 5.0). A uniform significance threshold of P < 0.05 was applied to all tests. Declarations Acknowledgements (optional) We would like to express our sincere gratitude to all individuals and organizations who supported and assisted us throughout this research. Special thanks to the following Yanfen Li.In conclusion, we extend our thanks to everyone who has supported and assisted us along the way. Without your support, this research would not have been possible. Author contributions Yang Shen:Conceptualization, Data curation, Validation, Visualization, Writing–original draft, Writing–review & editing. Yanfen Li:Data curation, Validation, Visualization, Writing–review & editing. Cijing Cai:Visualization, Writing–review & editing. Shenghong Lei: Conceptualization, Supervision, Writing–review & editing. This work currently described has not been published, is not being considered for publication elsewhere, and its publication was approved by all authors. Data availability statement The datasets ANALYZED for this study can be found in the [Gene Expression Omnibus (GEO) database] [http://www.ncbi.nlm.nih.gov/geo/]. Ethics approval This study was conducted in accordance with the Declaration of Helsinki and received approval from the Medical Ethics Committee of Shenzhen Qianhai Taikang Hospital (Approval No.: AF-LLPJ-20251022-02; Date: October 22, 2025). All patients provided written informed consent when clinical samples were collected for RT-qPCR experiments to ensure that the research process was in accordance with ethical norms and that the patients' rights and wishes were fully respected. Consent to publish Not Applicable Conflict of Interest The authors declare no competing interests. Funding This research received no specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References Rubino, F. et al. Definition and diagnostic criteria of clinical obesity. Lancet Diabetes Endocrinol. 13 , 221–262 (2025). Novelli, G., Cassadonte, C., Sbraccia, P. & Biancolella, M. Genetics: A Starting Point for the Prevention and the Treatment of Obesity. Nutrients 15 (2023). Jebeile, H., Kelly, A. S., O'Malley, G. & Baur, L. A. Obesity in children and adolescents: epidemiology, causes, assessment, and management. Lancet Diabetes Endocrinol. 10 , 351–365 (2022). Koenen, M., Hill, M. A., Cohen, P., Sowers, J. R. & Obesity Adipose Tissue and Vascular Dysfunction. Circ. Res. 128 , 951–968 (2021). Perdomo, C. M., Cohen, R. V., Sumithran, P., Clément, K. & Frühbeck, G. Contemporary medical, device, and surgical therapies for obesity in adults. Lancet 401 , 1116–1130 (2023). Yanovski, S. Z. & Yanovski, J. A. Approach to Obesity Treatment in Primary Care: A Review. JAMA Intern. Med. 184 , 818–829 (2024). Elmaleh-Sachs, A. et al. Obesity Management in Adults: A Review. Jama 330 , 2000–2015 (2023). Müller, T. D., Blüher, M., Tschöp, M. H. & DiMarchi, R. D. Anti-obesity drug discovery: advances and challenges. Nat. Rev. Drug Discov . 21 , 201–223 (2022). Fu, W. et al. Persistent activation of TRPM4 triggers necrotic cell death characterized by sodium overload. Nat. Chem. Biol. 21 , 1238–1249 (2025). Launay, P. et al. TRPM4 is a Ca2+-activated nonselective cation channel mediating cell membrane depolarization. Cell 109 , 397–407 (2002). Kappel, S., Ross-Kaschitza, D., Hauert, B., Rother, K. & Peinelt, C. p53 alters intracellular Ca(2+) signaling through regulation of TRPM4. Cell. Calcium . 104 , 102591 (2022). Csípő, T. et al. A Central Role for TRPM4 in Ca(2+)-Signal Amplification and Vasoconstriction. Int J. Mol. Sci 23 (2022). Bishnoi, M., Khare, P., Brown, L. & Panchal, S. K. Transient receptor potential (TRP) channels: a metabolic TR(i)P to obesity prevention and therapy. Obes. Rev. 19 , 1269–1292 (2018). Zsombok, A. & Derbenev, A. V. TRP Channels as Therapeutic Targets in Diabetes and Obesity. Pharmaceuticals (Basel) 9 (2016). Marigo, V., Courville, K., Hsu, W. H., Feng, J. M. & Cheng, H. TRPM4 impacts on Ca2 + signals during agonist-induced insulin secretion in pancreatic beta-cells. Mol. Cell. Endocrinol. 299 , 194–203 (2009). Tran, T. D. et al. Histamine-induced Ca²⁺ signalling is mediated by TRPM4 channels in human adipose-derived stem cells. Biochem. J. 463 , 123–134 (2014). Slovin, S. et al. Single-Cell RNA Sequencing Analysis: A Step-by-Step Overview. Methods Mol. Biol. 2284 , 343–365 (2021). Jovic, D. et al. Single-cell RNA sequencing technologies and applications: A brief overview. Clin. Transl Med. 12 , e694 (2022). So, J. et al. Robust single-nucleus RNA sequencing reveals depot-specific cell population dynamics in adipose tissue remodeling during obesity. Elife 13 (2025). Acharjee, A., Wijesinghe, S. N., Russ, D., Gkoutos, G. & Jones, S. W. Cross-species transcriptomics identifies obesity associated genes between human and mouse studies. J. Transl Med. 22 , 592 (2024). Liu, D. et al. Employing single-cell RNA sequencing coupled with an array of bioinformatics approaches to ascertain the shared genetic characteristics between osteoporosis and obesity. Bone Joint Res. 13 , 573–587 (2024). Ghosh, S. et al. Plasma Membrane Channel TRPM4 Mediates Immunogenic Therapy-Induced Necrosis. Cancer Res. 83 , 3115–3130 (2023). Yu, F. et al. Loss of the TRPM4 channel in humans causes immune dysregulation with defective monocyte migration. J. Allergy Clin. Immunol. 154 , 792–806 (2024). Sabaratnam, R. & Svenningsen, P. Adipocyte-Endothelium Crosstalk in Obesity. Front. Endocrinol. (Lausanne) . 12 , 681290 (2021). Li, M., Qian, M., Kyler, K. & Xu, J. Adipose Tissue-Endothelial Cell Interactions in Obesity-Induced Endothelial Dysfunction. Front. Cardiovasc. Med. 8 , 681581 (2021). Graupera, M. & Claret, M. Endothelial Cells: New Players in Obesity and Related Metabolic Disorders. Trends Endocrinol. Metab. 29 , 781–794 (2018). Garland, C. J. et al. TRPM4 inhibitor 9-phenanthrol activates endothelial cell intermediate conductance calcium-activated potassium channels in rat isolated mesenteric artery. Br. J. Pharmacol. 172 , 1114–1123 (2015). Wei, S. et al. Comparison of Anti-oncotic Effect of TRPM4 Blocking Antibody in Neuron, Astrocyte and Vascular Endothelial Cell Under Hypoxia. Front. Cell. Dev. Biol. 8 , 562584 (2020). Becerra, A. et al. Transient receptor potential melastatin 4 inhibition prevents lipopolysaccharide-induced endothelial cell death. Cardiovasc. Res. 91 , 677–684 (2011). Wang, C., Naruse, K. & Takahashi, K. Role of the TRPM4 Channel in Cardiovascular Physiology and Pathophysiology. Cells . 7 (2018). Abriel, H., Syam, N., Sottas, V., Amarouch, M. Y. & Rougier, J. S. TRPM4 channels in the cardiovascular system: physiology, pathophysiology, and pharmacology. Biochem. Pharmacol. 84 , 873–881 (2012). Miao, L. et al. Circulating miR-3659 may be a potential biomarker of dyslipidemia in patients with obesity. J. Transl Med. 17 , 25 (2019). Wang, Y. et al. Identification of metabolism related biomarkers in obesity based on adipose bioinformatics and machine learning. J. Transl Med. 22 , 986 (2024). Shi, S. et al. Identification of hub genes in the crosstalk between type 2 diabetic nephropathy and obesity according to bioinformatics analysis. Adipocyte 13 , 2423723 (2024). De, S. et al. Steric mechanism of auto-inhibitory regulation of specific and non-specific DNA binding by the ETS transcriptional repressor ETV6. J. Mol. Biol. 426 , 1390–1406 (2014). Bohlander, S. K. ETV6: a versatile player in leukemogenesis. Semin Cancer Biol. 15 , 162–174 (2005). Eldeeb, M. et al. Ontogeny shapes the ability of ETV6::RUNX1 to enhance hematopoietic stem cell self-renewal and disrupt early lymphopoiesis. Leukemia 38 , 455–459 (2024). Hock, H. et al. Tel/Etv6 is an essential and selective regulator of adult hematopoietic stem cell survival. Genes Dev. 18 , 2336–2341 (2004). Keller, M. et al. Genome-wide DNA promoter methylation and transcriptome analysis in human adipose tissue unravels novel candidate genes for obesity. Mol. Metab. 6 , 86–100 (2017). Wang, Y. et al. LAPTM5 mediates immature B cell apoptosis and B cell tolerance by regulating the WWP2-PTEN-AKT pathway. Proc. Natl. Acad. Sci. U S A . 119 , e2205629119 (2022). Glowacka, W. K., Alberts, P., Ouchida, R., Wang, J. Y. & Rotin, D. LAPTM5 protein is a positive regulator of proinflammatory signaling pathways in macrophages. J. Biol. Chem. 287 , 27691–27702 (2012). Jiang, B. et al. Lysosomal protein transmembrane 5 promotes lung-specific metastasis by regulating BMPR1A lysosomal degradation. Nat. Commun. 13 , 4141 (2022). Zhang, M. M. et al. The function and mechanism of LAPTM5 in diseases. Biomed. Pharmacother . 178 , 117237 (2024). Xu, X. et al. Obesity activates a program of lysosomal-dependent lipid metabolism in adipose tissue macrophages independently of classic activation. Cell. Metab. 18 , 816–830 (2013). Yan, K. et al. Integrative analyses of hub genes and their association with immune infiltration in adipose tissue, liver tissue and skeletal muscle of obese patients after bariatric surgery. Adipocyte 11 , 190–201 (2022). Sun, M. et al. Transcriptome combined with Mendelian randomization to identify key genes related to polyamine metabolism in childhood obesity and elucidate their molecular regulatory mechanisms. Sci. Rep. 15 , 17799 (2025). Jiang, L. et al. Lysosomal-associated protein transmembrane 5 ameliorates non-alcoholic steatohepatitis by promoting the degradation of CDC42 in mice. Nat. Commun. 14 , 2654 (2023). Kim, S. W. et al. Palmitoylation controls trafficking of the intracellular Ca(2+) channel MCOLN3/TRPML3 to regulate autophagy. Autophagy 15 , 327–340 (2019). Lei, Y. & Klionsky, D. J. MCOLN3/TRPML3 bridges the regulation of autophagosome biogenesis by PtdIns3P and the calcium channel. Autophagy 19 , 377–378 (2023). Spix, B. et al. Lung emphysema and impaired macrophage elastase clearance in mucolipin 3 deficient mice. Nat. Commun. 13 , 318 (2022). Martina, J. A., Lelouvier, B. & Puertollano, R. The calcium channel mucolipin-3 is a novel regulator of trafficking along the endosomal pathway. Traffic 10 , 1143–1156 (2009). Qadir, M. et al. Cyclosporin A is a broad-spectrum multidrug resistance modulator. Clin. Cancer Res. 11 , 2320–2326 (2005). Fuhrmann, A. et al. Molecular mechanisms underlying the effects of cyclosporin A and sirolimus on glucose and lipid metabolism in liver, skeletal muscle and adipose tissue in an in vivo rat model. Biochem. Pharmacol. 88 , 216–228 (2014). Jiang, M. et al. Cyclosporin A attenuates weight gain and improves glucose tolerance in diet-induced obese mice. Mol. Cell. Endocrinol. 370 , 96–102 (2013). Wang, B. et al. Nutrigenomic regulation of adipose tissue development - role of retinoic acid: A review. Meat Sci. 120 , 100–106 (2016). Zhu, S. et al. Adipose tissue plays a major role in retinoic acid-mediated metabolic homoeostasis. Adipocyte 11 , 47–55 (2022). Wang, B. et al. Retinoic acid induces white adipose tissue browning by increasing adipose vascularity and inducing beige adipogenesis of PDGFRα(+) adipose progenitors. Cell. Discov . 3 , 17036 (2017). Mogal, A. & Abdulkadir, S. A. Effects of Histone Deacetylase Inhibitor (HDACi); Trichostatin-A (TSA) on the expression of housekeeping genes. Mol. Cell. Probes . 20 , 81–86 (2006). Lu, J. C. et al. Trichostatin A modulates thiazolidinedione-mediated suppression of tumor necrosis factor α-induced lipolysis in 3T3-L1 adipocytes. PLoS One . 8 , e71517 (2013). Liu, X. et al. Trichostatin A regulates fibro/adipogenic progenitor adipogenesis epigenetically and reduces rotator cuff muscle fatty infiltration. J. Orthop. Res. 39 , 1452–1462 (2021). Wu, T. et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innov. (Camb) . 2 , 100141 (2021). Shannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 13 , 2498–2504 (2003). Hao, Y. et al. Dictionary learning for integrative, multimodal and scalable single-cell analysis. Nat. Biotechnol. 42 , 293–304 (2024). Aran, D. et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nat. Immunol. 20 , 163–172 (2019). Li, Y. et al. Development and validation of a hypoxia- and mitochondrial dysfunction- related prognostic model based on integrated single-cell and bulk RNA sequencing analyses in gastric cancer. Front. Immunol. 15 , 1419133 (2024). Lin, S. et al. Single-nucleus RNA sequencing reveals heterogenous microenvironments and specific drug response between cervical squamous cell carcinoma and adenocarcinoma. EBioMedicine . 97, 104846 (2023). Chen, H. et al. Integrated immunological analysis of single-cell and bulky tissue transcriptomes reveals the role of interactions between M0 macrophages and naïve CD4(+) T cells in the immunosuppressive microenvironment of cervical cancer. Comput. Biol. Med. 163 , 107151 (2023). Gustavsson, E. K., Zhang, D., Reynolds, R. H., Garcia-Ruiz, S. & Ryten, M. ggtranscript: an R package for the visualization and interpretation of transcript isoforms using ggplot2. Bioinformatics 38 , 3844–3846 (2022). Jin, S., Plikus, M. V. & Nie, Q. CellChat for systematic analysis of cell-cell communication from single-cell transcriptomics. Nat. Protoc. 20 , 180–219 (2025). Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32 , 381–386 (2014). Hänzelmann, S., Castelo, R. & Guinney, J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinform. 14 , 7 (2013). Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43 , e47 (2015). Van de Sande, B. et al. A scalable SCENIC workflow for single-cell gene regulatory network analysis. Nat. Protoc. 15 , 2247–2276 (2020). Morabito, S., Reese, F., Rahimzadeh, N., Miyoshi, E. & Swarup, V. hdWGCNA identifies co-expression networks in high-dimensional transcriptomics data. Cell. Rep. Methods . 3 , 100498 (2023). Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15 , 550 (2014). Gu, Z. Complex heatmap visualization. Imeta 1 , e43 (2022). Ortúzar, M., Riesco, R., Criado, M., Alonso, M. D. P. & Trujillo, M. E. Unraveling the dynamic interplay of microbial communities associated to Lupinus angustifolius in response to environmental and cultivation conditions. Sci. Total Environ. 946 , 174277 (2024). Friedman, J., Hastie, T. & Tibshirani, R. Regularization Paths for Generalized Linear Models via Coordinate Descent. J. Stat. Softw. 33 , 1–22 (2010). Guan, S. et al. Identifying potential targets for preventing cancer progression through the PLA2G1B recombinant protein using bioinformatics and machine learning methods. Int. J. Biol. Macromol. 276 , 133918 (2024). Mahmoudian, M., Venäläinen, M. S., Klén, R. & Elo, L. L. Stable Iterative Variable Selection Bioinformatics 37 , 4810–4817 (2021). Robin, X. et al. pROC: an open-source package for R and S + to analyze and compare ROC curves. BMC Bioinform. 12 , 77 (2011). Núñez, E., Steyerberg, E. W. & Núñez, J. [Regression modeling strategies]. Rev. Esp. Cardiol. 64 , 501–507 (2011). Cui, L. et al. Development of a nomogram for predicting 90-day mortality in patients with sepsis-associated liver injury. Sci. Rep. 13 , 3662 (2023). Vickers, A. J. & Elkin, E. B. Decision curve analysis: a novel method for evaluating prediction models. Med. Decis. Mak. 26 , 565–574 (2006). Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 03 Mar, 2026 Reviews received at journal 24 Feb, 2026 Reviewers agreed at journal 11 Feb, 2026 Reviewers agreed at journal 11 Feb, 2026 Reviewers agreed at journal 09 Feb, 2026 Reviewers invited by journal 09 Feb, 2026 Editor assigned by journal 09 Feb, 2026 Editor invited by journal 06 Feb, 2026 Submission checks completed at journal 05 Feb, 2026 First submitted to journal 05 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8754472","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":588484113,"identity":"ac125867-d320-4295-937d-c190dee7cd51","order_by":0,"name":"Yang Shen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxUlEQVRIiWNgGAWjYFACHgaGhAqb+n5m5oMPiNfy4Uwa48x2tmQDorUwzmw7zLjhPI+ZAFEazGfkHpPmbWNmNj7MYMbAUGMTTVCLzI28NGmec2xsZocZ0h4wHEvLbSCkRUIix0yap4yHB6jluAFjw2FitbBJSBg3M7ZJEK1FckabgYEBMzMbkVp43iVbfDiTkCBxmI3ZIIEov7DnHryRUPE/gb///McHH2psCGsBAhYJODOBCOUgwPyBSIWjYBSMglEwUgEA3+46NHBFuJUAAAAASUVORK5CYII=","orcid":"","institution":"Shenzhen Qianhai Tankang Hospital","correspondingAuthor":true,"prefix":"","firstName":"Yang","middleName":"","lastName":"Shen","suffix":""},{"id":588484114,"identity":"3dddb607-92aa-42dd-b18b-66897a8d728a","order_by":1,"name":"Yanfen Li","email":"","orcid":"","institution":"Bao’an Central Hospital of Shenzhen","correspondingAuthor":false,"prefix":"","firstName":"Yanfen","middleName":"","lastName":"Li","suffix":""},{"id":588484115,"identity":"41013aa4-4549-4fc2-99ea-f407e6e043b4","order_by":2,"name":"Cijing Cai","email":"","orcid":"","institution":"Bao’an Central Hospital of Shenzhen","correspondingAuthor":false,"prefix":"","firstName":"Cijing","middleName":"","lastName":"Cai","suffix":""},{"id":588484116,"identity":"e546c94b-344c-4328-a9a7-6d2ad71a0725","order_by":3,"name":"Shenghong Lei","email":"","orcid":"","institution":"Bao’an Central Hospital of Shenzhen","correspondingAuthor":false,"prefix":"","firstName":"Shenghong","middleName":"","lastName":"Lei","suffix":""}],"badges":[],"createdAt":"2026-02-01 07:38:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8754472/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8754472/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102717943,"identity":"c1cfe6d3-9cd5-443f-9d6b-895b3a1eab94","added_by":"auto","created_at":"2026-02-15 18:39:10","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":6397263,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eTRPM4\u003c/em\u003e expression, GSEA and regulatory networks. (a) Expression of \u003cem\u003eTRPM4\u003c/em\u003e in GSE155960dataset. Red represents the Obese group, blue represents the Control group. **: P \u0026lt; 0.01. (b) GSEA enrichment plot. The plot is divided into three parts from top to bottom: the first part shows the Enrichment Score (ES); the second part shows the hit, indicating genes within the gene set; the third part shows the rank distribution of all genes, using the Signal2Noise algorithm by default. (c) miRNA-mRNA regulatory network. Yellow nodes represent the key gene, green nodes represent predicted miRNAs. (d) TF-mRNA regulatory network. Yellow nodes represent the key gene, purple nodes represent predicted TFs.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8754472/v1/21b6697fb90122b47a700de6.png"},{"id":102717949,"identity":"3c9dfcd4-ba1b-44b7-b930-55862e573bdb","added_by":"auto","created_at":"2026-02-15 18:39:14","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":6397263,"visible":true,"origin":"","legend":"\u003cp\u003escRNA-seq analysis and identification of crucial cell types. (a) Identification of highly variable genes (HVGs). Red dots represent the top 2000 HVGs, black dots represent genes with low variation. The top 10 HVGs are labeled. (b) JackStraw plot. P-values were calculated for PCs 1:20, threshold P \u0026lt; 0.05. (c) Elbow plot. The inflection point tending to be flat occurs at PC=15. (d) UMAP plot of cell clusters. UMAP clustering identified 20 distinct cell clusters, with PCs 1:15 and resolution=0.5. (e) Cell type annotation. Different colors represent different cell types. (f) Bubble plot of marker gene expression across cell types. Color intensity represents average expression level in the cell type; dot size represents the percentage of cells expressing the gene. (g) Expression differences of \u003cem\u003eTRPM4\u003c/em\u003e in obese vs. control tissues.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8754472/v1/d7ec546043d974135607ea7f.png"},{"id":102717958,"identity":"32810abc-af26-4b99-b401-219a1b088972","added_by":"auto","created_at":"2026-02-15 18:39:18","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":962189,"visible":true,"origin":"","legend":"\u003cp\u003eCell communication, pseudotime trajectory, functional enrichment and TF activity analysis. (a-b) Cell-cell communication network. Network diagram showing the number of interactions (\u003cstrong\u003ea\u003c/strong\u003e) and interaction weights (\u003cstrong\u003eb\u003c/strong\u003e) between different cell types. Dot size represents cell number; line thickness represents communication strength. (c) Dot plot of key ligand-receptor interactions. Color from red to blue represents communication strength from strong to weak; dot size represents P-value. (d) Pseudotime trajectory analysis of endothelial cell subtypes. From left to right: cell differentiation pseudotime trajectory, cell state trajectory, cell type trajectory. (e) Heatmap of pseudotime-related gene expression during endothelial cell development. Top to bottom: pseudotime from early to late. Genes are clustered into 3 clusters. (f) Pseudotime trajectory analysis of APCs. From left to right: cell differentiation pseudotime trajectory, cell state trajectory, cell type trajectory. (g) Expression of \u003cem\u003eTRPM4\u003c/em\u003e during APC differentiation. Top to bottom: pseudotime from early to late. Y-axis represents gene expression level. (h-i) GSVA score comparison of differential GO pathways between obese and control samples. (H) Differential GO pathways between obese and control samples (P \u0026lt; 0.05). Green: downregulated pathways; Blue: upregulated pathways. (I) Enrichment of differential pathways across cell types. (j) SCENIC analysis of highly activated TFs across cell types. Color intensity represents RSS score.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8754472/v1/9f4d42cbb44be61368d4cffe.png"},{"id":102717942,"identity":"51b74a8c-7726-41a6-aa0c-a44e9511c5e8","added_by":"auto","created_at":"2026-02-15 18:39:10","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":436732,"visible":true,"origin":"","legend":"\u003cp\u003eIdentification of key module genes via hdWGCNA. (a) Scale-free topology fit index and mean connectivity for soft-thresholding powers. A soft-thresholding power of 6 was selected. (b) Hierarchical clustering dendrogram. The dendrogram shows 11 co-expression modules (excluding the grey module for unassigned genes). (c) Correlation among different modules. (d) Volcano plot of gene expression level differences for modules between obese and control groups. Dot size represents the number of genes in the module. X-axis: log2FC; Y-axis: -log10(P-value).\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8754472/v1/a713295d25720ca242b430e0.png"},{"id":102717956,"identity":"d1626ca2-4d3f-4592-b2f0-5d8353e636a0","added_by":"auto","created_at":"2026-02-15 18:39:17","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":542979,"visible":true,"origin":"","legend":"\u003cp\u003eIdentification and functional analysis of candidate genes. (a) Volcano plot of DEGs. X-axis: log2FoldChange; Y-axis: -log10(P-value). Red dots: upregulated DEGs; Blue dots: downregulated DEGs. Genes are labeled as the top 10 up/down-regulated by log2FC. (b) Heatmap of DEGs. Top: density distribution heatmap of DEGs. Bottom: expression heatmap of the top 10 up/down-regulated DEGs by log2FC. Color bar indicates relative gene expression from blue (low) to red (high). (c) Identification of candidate genes. Green circle: DEGs; Purple circle: key module genes from hdWGCNA; Overlap: candidate genes. (d) GO enrichment analysis. Dot size represents the number of enriched genes; color represents P-value. (e) PPI network. Nodes represent candidate genes; connecting lines represent interactions.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8754472/v1/7285283b503971ab391aa812.png"},{"id":102717887,"identity":"4177e25e-fc06-412b-97af-9b038b389af3","added_by":"auto","created_at":"2026-02-15 18:38:59","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":870141,"visible":true,"origin":"","legend":"\u003cp\u003eIdentification of key genes using machine learning and ROC analysis. (a) LASSO regression coefficient distribution (left) and cross-validation curve (right). The vertical dashed line (right) corresponds to lambda.min where the error is minimal. (b) Accuracy and error rate curves of the SVM-RFE algorithm. The red circle indicates the point with the lowest error rate. (c) Boruta feature selection. Importance scores of variables during Boruta algorithm runs. Green: important features; Red: unimportant features; Blue: shadow features; Yellow: tentative features. (d) Venn diagram of candidate key genes. Red circle: LASSO genes; Yellow circle: SVM-RFE genes; Orange circle: Boruta genes; Overlap: common characteristic genes. (e) ROC curves in GSE156906 dataset. AUC values for each gene are labeled. (f) ROC curves in GSE25401 dataset. (g) Expression of candidate biomarkers in GSE156906 dataset. Box plots show expression differences of key genes in the training set. Red: Obese group; Blue: Control group. ***: P \u0026lt; 0.001, ****: P \u0026lt; 0.0001. (h) Expression of candidate biomarkers in GSE25401 dataset.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8754472/v1/6a9862b842893c160772f164.png"},{"id":102717883,"identity":"9a2f0fe9-c07b-4f0c-9972-aaa8d77cb115","added_by":"auto","created_at":"2026-02-15 18:38:58","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":400416,"visible":true,"origin":"","legend":"\u003cp\u003eConstruction and evaluation of the diagnostic nomogram. (a) Nomogram. The total points predict the probability of obesity onset. (b) Calibration curve. A slope closer to 1 indicates higher prediction accuracy. (c) DCA curve. The net benefit value of the DCA curve being greater than 0 indicates good predictive performance.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8754472/v1/17f4d1fa795095464ef756a3.png"},{"id":102717954,"identity":"fede7d49-e702-491f-8cf1-cce48a623f70","added_by":"auto","created_at":"2026-02-15 18:39:17","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":369681,"visible":true,"origin":"","legend":"\u003cp\u003eSubcellular localization, GGI network and drug/compound interactions. (a) Subcellular localization. Radar plot shows the predicted probability distribution of subcellular localization for key genes. (b) GeneMANIA analysis of biomarkers. The center shows the 3 key genes, surrounded by the top 20 related genes, displaying the top 7 significant functional pathways and their interactions via 7 different methods. (c) Drug-biomarker interaction network. Orange nodes: predicted drugs/compounds; Blue nodes: biomarkers.\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8754472/v1/1a393d4a51abc513584954db.png"},{"id":102717952,"identity":"5c242646-c579-464c-8042-04044d32c2c0","added_by":"auto","created_at":"2026-02-15 18:39:15","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":183494,"visible":true,"origin":"","legend":"\u003cp\u003eRT-qPCR validation of key gene expression. (a-c) mRNA levels of \u003cem\u003eETV6\u003c/em\u003e (\u003cstrong\u003ea\u003c/strong\u003e), \u003cem\u003eLAPTM5\u003c/em\u003e (\u003cstrong\u003eb\u003c/strong\u003e), and \u003cem\u003eMCOLN3\u003c/em\u003e (\u003cstrong\u003ec\u003c/strong\u003e) in clinical samples. **: P \u0026lt; 0.01, ***: P \u0026lt; 0.001, ****: P \u0026lt; 0.0001.\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-8754472/v1/42db09c1e38da48353cfd622.png"},{"id":102717992,"identity":"0fc0591b-aea5-4f68-b08e-013d9d20dbf5","added_by":"auto","created_at":"2026-02-15 18:39:30","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":18146988,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8754472/v1/1dc732ec-ea03-451e-8555-6b4820505905.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Single-cell WGCNA combined with transcriptome sequencing to study the molecular mechanisms of TRPM4-related genes in obesity with experimental validation","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eObesity is a long-lasting metabolic disorder marked by excessive fat buildup and abnormal weight gain, usually measured by BMI. In recent years, the worldwide rate of obesity has steadily risen, becoming a major public health concern globally[\u003cspan additionalcitationids=\"CR2 CR3\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. At present, treatment strategies for obesity primarily encompass lifestyle modifications\u0026mdash;including dietary regulation and physical activity\u0026mdash;pharmacotherapy, and surgical interventions such as bariatric procedures. Although these approaches can partially enhance metabolic parameters and promote weight loss, they are often associated with such problems as inconsistent efficacy, poor patient adherence, high rates of recurrence, and potential complications. For instance, weight reduction medications may induce adverse effects, while surgical options, despite their effectiveness, carry risks of complications and entail substantial financial costs. Consequently, the current therapeutic modalities possess notable limitations in addressing this multifaceted disease[\u003cspan additionalcitationids=\"CR6 CR7\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. The advent of molecular biology, genomics, transcriptomics, and metabolomics has enabled researchers to gain a deeper understanding of the mechanisms underlying obesity. Several regulatory factors have been identified, including adipokines, inflammatory factors, microbial metabolites, and non-coding RNAs. These biomarkers facilitate early diagnosis, classification, and personalized risk assessment, providing a basis for identifying therapeutic targets and improving individualized treatments.\u003c/p\u003e \u003cp\u003eIn recent years, a novel form of cell death, sodium overload-induced necrosis (NECSO), has gained research interest. It involves Na⁺ accumulation, causing osmotic imbalance, cell swelling, membrane rupture, and necrosis. This releases inflammatory factors that activate immune cells. NECSO is closely linked to TRPM4, a transient receptor potential channel family[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. TRPM4 is a calcium-activated, non-selective cation channel that primarily transmits sodium ions, rather than calcium. It activates in response to increased intracellular calcium, leading to sodium influx, membrane depolarization, and indirectly influencing the opening of calcium channels and signaling[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. TRPM4 is widely expressed in various tissues, including the myocardium, nervous system, and immune system, as well as in stem cells. It regulates proliferation, differentiation, migration, and secretion in various physiological and pathological contexts[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Recent studies have demonstrated that TRPM4 is present in tissues involved in metabolism, such as adipose tissue, liver, and immune cells. Its increased levels correlate with chronic inflammation and metabolic disruptions caused by obesity[\u003cspan additionalcitationids=\"CR14\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Research has shown that TRPM4 plays a crucial role in regulating calcium signaling and adipocyte differentiation in human adipose-derived stem cells (hASCs). TRPM4 controls adipogenesis in human adipose-derived stem cells (hASCs) by mediating histamine-triggered Ca\u0026sup2;⁺ signaling. The process involves releasing Ca\u0026sup2;⁺ from the endoplasmic reticulum and increasing Ca\u0026sup2;⁺ influx by promoting the opening of L-type voltage-dependent Ca\u0026sup2;⁺ channels (Cav1.2), which then activates adipogenic transcription factors and enzymes. Conversely, knocking down this gene decreases lipid droplet formation and downregulates adipocyte marker genes (e.g., C/EBPα), confirming its vital role in hASC adipogenesis. Furthermore, it may hold significance in the pathogenesis of obesity-related diseases[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Therefore, re-examining obesity's pathogenesis through cellular ion homeostasis and programmed cell death offers a new understanding and foundation for research on TRPM4-mediated sodium-dependent cell death in the context of obesity pathophysiology.\u003c/p\u003e \u003cp\u003eRecently, single-cell analysis has become a pioneering tool in the life sciences, enabling the detection of cellular heterogeneity, rare cell populations, and changes in cell states with single-cell precision. Its potential is evident in tumour immunity, neurodevelopment, and inflammatory diseases[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Integrating bulk RNA-seq with single-cell RNA-seq helps understand gene expression trends and disease-related pathways at the systemic level. It also enables detailed analysis of the expression and regulation of key genes within cell subpopulations. This approach improves result stability and interpretability, identifying core cell types and driver genes involved at key disease nodes[\u003cspan additionalcitationids=\"CR20\" citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eOur study analyzes obesity-related transcriptomic data from public repositories, using single-cell sequencing, hierarchical Weighted Gene Co-expression Network Analysis (hdWGCNA), and machine learning to identify and assess important genes linked to TRPM4. It explores their regulatory roles in obese tissue through gene-gene interaction networks and drug target prediction, aiming to uncover the molecular functions of TRPM4 in obesity. This integrated bioinformatics approach enables the identification of diagnostic and therapeutic targets, thereby supporting personalized interventions and precision treatment strategies.\u003c/p\u003e"},{"header":"2 Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 TRPM4 as a potential contributor to obesity: expression, pathways, and regulation\u003c/h2\u003e \u003cp\u003eIn the GSE156906 dataset, TRPM4 demonstrated significantly elevated expression levels in the obese group relative to the control group (P\u0026thinsp;\u0026lt;\u0026thinsp;0.01) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea), suggesting its potential crucial role in the pathogenesis of obesity. Among the top 10 pathways identified through GSEA analysis of TRPM4, pathways with a clear link to obesity were observed, including \"response of Eif2ak4/Gcn2 to amino acid deficiency,\" \"SRP-dependent cotranslational protein targeting to membrane,\" and \"cytoplasmic ribosomal proteins (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb).\" Furthermore, TRPM4 was subject to regulation by numerous regulatory factors. Specifically, TRPM4 was predicted to be regulated by five miRNAs, such as hsa-miR-877-5p (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec). In addition, TRPM4 was found to be regulated by 31 TFs, such as SP1 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Endothelial cells were identified as a crucial cell type\u003c/h2\u003e \u003cp\u003eBefore the implementation of QC procedures, GSE155960 comprised 83,537 cells and 18,498 genes. Following the QC procedures, after quality control, the dataset contained 83,149 cells and 18,498 genes (\u003cb\u003eSupplementary Fig.\u0026nbsp;1a-b\u003c/b\u003e). After standard data processing, 2,000 HVGs were identified (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). PCA analysis demonstrated that distinct clusters, visibly segregated in the principal component space, emerged from different samples within GSE155960 (\u003cb\u003eSupplementary Fig.\u0026nbsp;1c\u003c/b\u003e), highlighting significant expression disparities among the samples. The top 15 PCs (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) were subsequently chosen (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb-c). After that, UMAP was employed to visualize 20 distinct cell clusters (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed). A total of 10 cell types were identified based on specific marker genes: T cells (marked by CD3D, CD3E, and CD3G), preadipocytes (marked by PDGFRA, F3, and CEBPA), natural killer (NK) cells (marked by FGFBP2, FCGR3A, and GNLY), APCs (marked by FN1, CD34, and CD55), monocytes (marked by S100A8, S100A9, and FCN1), conventional dendritic cells (cDCs, marked by XCR1, FLT3, CCR7, and CD1E), macrophages (marked by C1QA, C1QB, and CD14), endothelial cells (marked by PECAM1, VWF, and IFI27), smooth muscle cells (marked by ACTA2, TAGLN, and MYH11), and B cells (marked by MS4A1, CD79B, and CD79A) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ee-f). Notably, in the CD45-positive cell samples, both the obese and control groups exhibited the highest proportion of T cells (\u003cb\u003eSupplementary Fig.\u0026nbsp;1d\u003c/b\u003e). Additionally, among all annotated cell types, TRPM4 demonstrated a significant difference in expression exclusively in endothelial cells between the obese and control groups (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05), thereby designating endothelial cells as the crucial cell type (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eg).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Analysis of cellular communication, pseudo-temporal trajectory, functional enrichment, and TF activity\u003c/h2\u003e \u003cp\u003eFurther analysis was conducted on the intricate communication networks among the annotated cell types. Endothelial cells were found to exhibit a relatively high number of interactions with other annotated cells, while T cells and NK cells demonstrated stronger interaction weights (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea-b, \u003cb\u003eSupplementary Fig.\u0026nbsp;1e\u003c/b\u003e). Notably, the VEGFB-VEGFR1 and ANGPTL2-(ITGA5\u0026thinsp;+\u0026thinsp;ITGB1) interaction pairs exhibited a high likelihood of occurrence during the pathway from preadipocytes to endothelial cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec). Subsequently, pseudo-temporal trajectory analysis was performed on endothelial cells. The diverging temporal trajectories of endothelial cells and their subclusters are depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed, revealing that endothelial cells were categorized into five distinct states over time. Further investigation revealed that the genes contained in endothelial cells were clustered into three groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ee). The genes in cluster 1 exhibited a trend of initially decreasing and then increasing during endothelial cell differentiation. The genes in cluster 2 showed an opposite pattern, rising first and then declining. The genes in cluster 3 demonstrated a more complex trend, decreasing initially, then increasing, and subsequently decreasing again throughout endothelial cell differentiation. Additionally, a pseudo-temporal trajectory analysis was performed on APCs, which revealed that APCs were differentiated into three distinct states over time (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ef). The expression level of TRPM4 remained unchanged during APC differentiation (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eg). Moreover, GSVA was employed to evaluate variations in gene enrichment pathways among annotated cells in obese and control samples. The obese group demonstrated the activation of six pathways, including \"vesicle tethering to Golgi\" and \"chiasma assembly\" (t\u0026thinsp;\u0026gt;\u0026thinsp;2, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05), whereas 38 pathways, such as \"glycerol transport\" and \"regulation of synaptic vesicle transport,\" were suppressed (t\u0026thinsp;\u0026lt;\u0026thinsp;2, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eh). Furthermore, the enrichment patterns of the 10 annotated cell types demonstrated considerable variation across different biological pathways (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ei). For example, the process of \"inner ear receptor cell fate commitment\" showed significant enrichment in endothelial cells but lower enrichment in other cell types. Subsequent predictions of TF activities revealed distinct differences among the annotated cell types (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ej). For instance, the analysis indicated that ATOH8 displayed high expression levels in endothelial cells and comparatively low expression in other annotated cell types.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Identification of 384 key module genes via hdWGCNA\u003c/h2\u003e \u003cp\u003eKey molecular characteristics of endothelial cells were discerned through hdWGCNA. Specifically, during co-expression network construction, a soft-thresholding power of 6 was selected \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea\u003cb\u003e)\u003c/b\u003e to construct an unweighted endothelial cell network, as this value achieved a scale-free topology fit index above 0.8 and ensured optimal connectivity. This was followed by the identification of 11 distinct co-expression modules through hierarchical clustering combined with dynamic tree cuttinge (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb). The correlation among the 11 modules is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec and \u003cb\u003eSupplementary Table\u0026nbsp;3\u003c/b\u003e. Moreover, genes within the black module demonstrated significant upregulation in the obesity group, whereas genes within the purple and magenta module demonstrated significant downregulation in the obesity group (adj. P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed). Consequently, these three modules were designated as key modules, comprising 186 (black), 78 (purple), and 120 (magenta) genes, for a total of 384 genes.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Recognition and related functional exploration of candidate genes linked to TRPM4 in obesity\u003c/h2\u003e \u003cp\u003eAnalysis of GSE156906 revealed 2,189 DEGs when comparing the obese and control groups. Among these DEGs, 1,569 genes exhibited up-regulation, while 620 genes showed down-regulation in the obese group (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea-b). An intersection analysis between the DEGs and key module genes yielded 17 candidate genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec). Functional analysis revealed that the 17 candidate genes were significantly enriched for 223 GO terms (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05), comprising 165 BPs, 20 CCs, and 38 MFs (\u003cb\u003eSupplementary Table\u0026nbsp;4\u003c/b\u003e). Specifically, the top 10 significantly enriched pathways (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) for BPs, CCs and MFs included \"cellular response to tumor necrosis factor\", \"late endosome membrane\", and \"DNA-binding transcription activator activity\" (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ed). After excluding four candidate genes encoding discrete proteins, a PPI network was constructed with the remaining 13 proteins (e.g., ETV6) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ee).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 ETV6, LAPTM5, and MCOLN3 were identified as key genes\u003c/h2\u003e \u003cp\u003eAmong the 17 candidate genes, LASSO analysis yielded the lowest prediction error when lambda was set to lambda. min\u0026thinsp;=\u0026thinsp;0.03331 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea). At this specific lambda value, seven genes exhibited non-zero regression coefficients, and these genes (LAPTM5, SMIM3, GJA1, LRMDA, ETV6, SLC16A3, and MCOLN3) were subsequently identified as LASSO characteristic genes. Moreover, the SVM-RFE model achieved peak predictive accuracy when 16 variables were selected (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eb). The following 16 genes were consequently identified as SVM-RFE characteristic genes: MCOLN3, NR4A1, SLC16A3, FABP4, ETV6, LAPTM5, RGS19, CDKN1C, GJA1, PRDX6, SMIM3, LRMDA, KLF9, CEBPD, CXCL3, and RHOB. Furthermore, the Boruta algorithm pinpointed 12 genes that exhibited importance scores exceeding those of ShadowMax, thereby designating them as Boruta characteristic genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ec). These genes included FABP4, ASAH1, PRDX6, RHOB, LAPTM5, SMIM3, GJA1, KLF9, LRMDA, ETV6, MCOLN3, and RGS19. Subsequently, the overlap among the characteristic genes identified by the three aforementioned machine learning algorithms was determined (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ed), yielding six common characteristic genes: LAPTM5, SMIM3, GJA1, LRMDA, ETV6, and MCOLN3. In datasets GSE156906 and GSE25401, ETV6, LAPTM5, and MCOLN3 exhibited AUC values exceeding 0.7 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ee-f). Moreover, within GSE156906 and GSE25401, ETV6 and LAPTM5 demonstrated markedly elevated expression levels in the obese group (P\u0026thinsp;\u0026lt;\u0026thinsp;0.0001), while MCOLN3 levels were significantly reduced (P\u0026thinsp;\u0026lt;\u0026thinsp;0.0001) (\u003cb\u003eFigure. 6g-h\u003c/b\u003e). The results indicated that ETV6, LAPTM5, and MCOLN3 exhibited good stability and high diagnostic potential, rendering them promising candidates for further exploration as key genes associated with obesity.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Nomogram demonstrated favorable performance in assessing the diagnosis of obesity\u003c/h2\u003e \u003cp\u003eA nomogram was constructed to evaluate the diagnostic effectiveness of ETV6, LAPTM5, and MCOLN3 for identifying obesity. According to the nomogram, a higher total score derived from these three genes was associated with a greater likelihood of obesity onset (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ea). The calibration curve demonstrated satisfactory performance by passing the HL goodness-of-fit test, yielding a P value of 0.252 (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eb). This result suggested no significant discrepancy between predicted and observed outcomes, thus validating the nomogram's high accuracy. The DCA findings revealed that the nomogram model demonstrated a notably higher net benefit than utilizing individual key genes alone (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ec), suggesting that it could help clinicians to make more accurate early diagnoses of obesity.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Subcellular localization, GGI network establishment, and drug/compound interactions associated with ETV6, LAPTM5, and MCOLN3\u003c/h2\u003e \u003cp\u003eSubsequent analysis was performed on the structures of ETV6, LAPTM5, and MCOLN3. In particular, ETV6 and MCOLN3 were mainly localized within the nucleus, whereas LAPTM5 exhibited predominant cytoplasmic localization (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ea). The GGI network comprised 23 genes, encompassing ETV6, LAPTM5, and MCOLN3, in addition to 20 other related genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eb). MCOLN3, along with MCOLN1 and MCOLN2, collectively participated in multiple biological pathways, such as \"calcium ion transmembrane transporter activity\" and \"divalent inorganic cation transmembrane transporter activity.\" As predicted by DsigDB, ETV6, LAPTM5, and MCOLN3 targeted 51 drugs/compounds (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ec). Of these, cyclosporin A, retinoic acid and aflatoxin B1 were targeted by both LAPTM5 and MCOLN3, arsenious acid and formaldehyde were targeted by both ETV6 and LAPTM5, and trichostatin A was targeted by both ETV6 and MCOLN3.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e2.9 RT-qPCR analysis of ETV6, LAPTM5, and MCOLN3 expression\u003c/h2\u003e \u003cp\u003eConsistent with our bioinformatic predictions, RT-qPCR validation confirmed that mRNA levels of ETV6 and LAPTM5 were significantly upregulated in the obese group (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001; Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ea-b), while MCOLN3 expression was notably downregulated (P\u0026thinsp;\u0026lt;\u0026thinsp;0.01; Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ec). The RT-qPCR results for ETV6, LAPTM5, and MCOLN3 were in agreement with their predicted expression patterns from the GSE156906 and GSE25401 datasets, collectively confirming the accuracy of our bioinformatic predictions.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"3. Discussion","content":"\u003cp\u003eObesity is a complex metabolic disorder closely connected to chronic inflammation, immune imbalance, and multiple molecular signaling pathways. Research indicates that TRPM4, a member of the transient receptor potential channel family, plays a crucial regulatory role in various immunological and metabolic diseases, suggesting that it may also contribute to the development and progression of obesity[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. To investigate the role of TRPM4 in obesity, this study combined bulk and single-cell transcriptomics data, revealing significant differences in TRPM4 expression between endothelial cells in obese and control groups. Endothelial cells were identified as a key cell type. hdWGCNA was employed to identify crucial module genes associated with endothelial cells, followed by screening using machine learning, ROC analysis, and validation. This process identified three essential genes: ETV6, LAPTM5, and MCOLN3. These genes could enhance understanding of TRPM4 regulatory networks in obesity and serve as potential targets for diagnosis and treatment. This study, for the first time, indicates an association between TRPM4 and the occurrence and development of obesity, offering a new perspective for research on TRPM4-related factors in obesity. This discovery deepens our understanding of the molecular mechanisms behind obesity and offers valuable theoretical insights and practical hints for creating new diagnostic markers and intervention strategies focused on these key genes.\u003c/p\u003e \u003cp\u003escRNA-seq analysis revealed significantly altered TRPM4 expression in endothelial cells of obese subjects compared to controls. Endothelial cells, identified as key regulators, provide insights into their role in obesity and TRPM4-related functions. Endothelial cell dysfunction is a crucial factor in the pathophysiology of obesity. In obesity, inflammatory factors (TNF-α, IL-6) and adipokines secreted by adipose tissue suppress the expression of insulin receptor substrate (Irs) in endothelial cells, thereby impairing insulin-induced vasodilation and glucose uptake, which leads to systemic insulin resistance[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Meanwhile, endothelial cell dysfunction is characterized by reduced vasodilatory capacity (decreased eNOS activity, lower NO production), decreased vascular density, and abnormal remodeling. These changes actively contribute to the development and persistence of metabolic disorders by impairing nutrient transport and oxygen supply to tissues[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. Studies have shown that TRPM4 regulates calcium homeostasis, vascular barrier integrity, vascular tension, and inflammatory signaling pathways in endothelial cells[\u003cspan additionalcitationids=\"CR28\" citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Additionally, TRPM4 also affects vascular permeability and blood perfusion, indirectly regulating the metabolic supply of adjacent tissues[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Therefore, the results of this study suggest that RPM4 may participate in the process of obesity-related endothelial dysfunction and metabolic disorders by regulating the calcium signal transduction and barrier function of endothelial cells. This provides an important theoretical basis for understanding the mechanism of sodium overload-induced necrosis (NECSO) mediated by TRPM4 in the pathophysiology of obesity.\u003c/p\u003e \u003cp\u003eThis study thoroughly explores the potential mechanistic roles of three key genes\u0026mdash;ETV6, LAPTM5, and MCOLN3. First, it clarifies the biological functions of these three genes under normal physiological conditions. Then, by integrating previous related research and analyzing the expression differences between obese individuals and the control group in this study, the expression patterns and possible roles of these genes in obesity and related metabolic diseases are thoroughly explored. Notably, some of these genes have not been previously reported in obesity. Additionally, the results of the RT-qPCR experiment can be combined for discussion: ETV6 and LAPTM5 were significantly upregulated in the obese group, while MCOLN3 was downregulated considerably (P\u0026thinsp;\u0026lt;\u0026thinsp;0.01). This trend in expressions was consistent with the database's prediction results. The expression patterns shown in the experimental data match those in the database analysis, indicating these genes have stable expression changes under specific conditions. This finding not only supports the reliability of bioinformatics predictions at the experimental level but also further strengthens the credibility and scientific basis of this study\u0026rsquo;s conclusions[\u003cspan additionalcitationids=\"CR33\" citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eETV6 (ETS variant transcription factor 6) is a type of transcription repressor within the ETS transcription factor family. It plays a crucial regulatory role in the hematopoietic system[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e].Additionally, ETV6 plays a vital role in the maintenance and differentiation of hematopoietic stem cells. Its mutation or abnormal expression has been widely linked to various hematologic malignancies and myelodysplastic syndromes (MDS)[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. In addition to its role in the blood system, recent studies have indicated that ETV6 may also serve a significant regulatory function in metabolic disorders. Notably, a study investigating the epigenetic landscape of human subcutaneous and visceral adipose tissue observed an inverse correlation between ETV6 promoter methylation and gene expression. This phenomenon was consistently validated across two independent population cohorts[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. These findings suggest that ETV6 may be involved in adipogenesis, lipid metabolism, and the growth and function of adipose tissue through epigenetic mechanisms, potentially contributing to the development of obesity. In summary, ETV6 plays a crucial role in blood development and also regulates fat tissue metabolism. This suggests that it could serve as a potential link between metabolic disorders and hematopoietic system diseases, warranting further research.\u003c/p\u003e \u003cp\u003eLAPTM5 (Lysosomal-Associated Protein Transmembrane 5) is a protein associated with lysosomes, featuring multiple transmembrane domains. It is primarily expressed in hematopoietic cells, especially in B cells and macrophages. LAPTM5 is situated on the lysosomal membrane and is involved in lysosome-mediated protein degradation, signal transduction, and immune regulation. It plays a vital role in maintaining cellular homeostasis and immune responses[\u003cspan additionalcitationids=\"CR41 CR42\" citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]. Recent studies have demonstrated that the role of LAPTM5 has expanded from traditional immune regulation to the area of metabolic diseases. Gene expression analysis revealed that in conditions of obesity and insulin resistance, LAPTM5 was upregulated, alongside several genes related to lysosomal function, suggesting that it may play a role in regulating lipid metabolism and the abnormal lysosomal functions associated with obesity[\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e]. The bioinformatics research further identified LAPTM5 as a key hub gene in the obesity-related chronic inflammation network. Its expression level is closely connected to the infiltration of M1 macrophages, suggesting its potential involvement in regulating the immune response following bariatric surgery[\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]. Studies on childhood obesity have also shown that LAPTM5 is involved in the process of polyamine metabolism disorder, leading to an imbalance in the immune microenvironment through the lysosome pathway, further supporting its significant role in the development of obesity[\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]. Additionally, animal experiments have provided indirect evidence for the metabolic regulatory function of LAPTM5: liver cell-specific knockout of Laptm5 worsens non-alcoholic fatty liver disease (NAFLD), while overexpression of LAPTM5 can significantly reduce liver inflammation and lipid buildup by promoting the lysosomal degradation of CDC42 and inhibiting the MAPK signaling pathway, emphasizing its protective role in lipid metabolism[\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. In conclusion, LAPTM5, as a key molecule linking the immune system and metabolic disorders, not only performs the traditional lysosomal regulatory function in immune cells but also shows extensive regulatory potential in obesity and related metabolic diseases. Its value as a diagnostic or therapeutic target is worth in-depth exploration.\u003c/p\u003e \u003cp\u003eMCOLN3 (Mucolipin 3), also known as TRPML3 (Transient Receptor Potential Mucolipin 3), belongs to the mucolipin subfamily of the TRP (transient receptor potential) ion channel family. MCOLN3 is a cation channel situated in the inner membrane system of cells, widely present on the membranes of early endosomes, late endosomes, lysosomes, and other organelles. It primarily regulates Ca\u0026sup2;⁺ homeostasis within cells and participates in various biological processes, such as vesicle transport, regulation of the endocytosis pathway, and maintenance of lysosomal function[\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e, \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e]. The activity of MCOLN3 in the lysosome-endosome system is essential for the cell's response to osmotic pressure changes, regulation of vesicle acidification, and facilitation of membrane fusion and other cellular stress and transport processes[\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e, \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]. Although MCOLN3 is crucial for organelle function and calcium signaling regulation, there has been no systematic research on its specific roles in obesity or metabolism-related conditions, such as abnormal fat processing or energy balance issues. Our PCR results demonstrated significantly lower MCOLN3 expression in the obese group compared to controls. This downregulation suggests a potential role for MCOLN3 in obesity pathogenesis, possibly mediated through its known functions in regulating intracellular calcium levels and vesicle transport. The reduced expression of MCOLN3 could impair the lysosome-endosome system, which would affect lipid metabolism and cellular energy regulation. However, further experimental validation is necessary to confirm this finding mechanism. This finding opens up new avenues for investigating the functions of MCOLN3 within the energy metabolism network. It may broaden the scope of research into the roles of TRPML family channels in metabolic diseases. It also suggests potential new interaction mechanisms between calcium channels and lipid metabolism in the endosomal system, providing valuable theoretical insights and promising research opportunities.\u003c/p\u003e \u003cp\u003eBased on the prediction results of drugs and compounds, explore potential intervention targets and mechanisms of key genes: Both LAPTM5 and MCOLN3 target cyclosporin A and retinoic acid. Additionally, ETV6 and MCOLN3 target trichostatin A. These findings suggest that these compounds may participate in regulating obesity-related immune and metabolic pathways through a multi-target synergistic mechanism and could have notable intervention potential.\u003c/p\u003e \u003cp\u003eCyclosporin A (Cyclosporin A) is a well-known calcineurin inhibitor widely used in immunosuppressive therapy after organ transplantation. Recent research has shown that it also plays a significant role in non-immune systems[\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e]. Cyclosporine A influences the inflammatory response and metabolic balance of adipose tissue by disrupting mitochondrial function[\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e]. In this study, the drug was predicted to be a targeted molecule that acts on both LAPTM5 and MCOLN3, suggesting that cyclosporine A may participate in regulating adipose tissue inflammation and metabolic homeostasis through pathways involving LAPTM5 and MCOLN3.\u003c/p\u003e \u003cp\u003eRetinoic acid (RA) is a metabolite of vitamin A that can regulate gene expression by activating nuclear receptors RAR/RXR. Retinoic acid has been extensively studied in the processes of adipocyte fate determination, brown fat activation, and adipose tissue remodeling[\u003cspan additionalcitationids=\"CR56\" citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e]. We found that LAPTM5 and MCOLN3 are also potential targets for retinoic acid, indicating that these two genes may influence retinoic acid-driven lipid metabolism processes.\u003c/p\u003e \u003cp\u003eTrichostatin A (TSA) is a widely used histone deacetylase inhibitor (HDACi) that significantly influences chromatin structure and transcriptional activity by inhibiting the functions of Class I and Class II HDACs[\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e]. TSA may also influence the expression of genes involved in adipogenesis and metabolism[\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e, \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e]. In this study, TSA was predicted to target both ETV6 and MCOLN3 simultaneously, suggesting that these two may be involved in regulating the influence of epigenetic mechanisms on obesity.\u003c/p\u003e \u003cp\u003eLAPTM5 is the gene among the three key genes that has the most drug interactions. It is also associated with metabolic regulatory functions, such as cyclosporine A and retinoic acid. This suggests that it may act as a central hub in the obesity-related immune and metabolic regulation network. This finding highlights the importance of LAPTM5 as a potential target and further confirms its role in regulating multi-pathway and multi-drug response networks. It is worth noting that these drug-gene interaction predictions are primarily based on analyzing gene expression correlations and may involve various mechanisms, including direct binding, transcriptional regulation, and regulation of signaling pathways. The specific molecular interaction patterns still need further experimental verification to be confirmed. Moving forward, a drug-gene-phenotype functional verification system could be developed around LAPTM5 to evaluate its translational potential in the treatment of obesity thoroughly.\u003c/p\u003e \u003cp\u003eThis study investigates the potential molecular mechanisms behind obesity. It starts with single-cell transcriptomic analysis to identify endothelial cells as key regulatory types in adipose tissue, guiding future efforts to target specific cell populations. Further research, combining various machine learning algorithms with bioinformatics methods, carefully screens and identifies three critical genes\u0026mdash;ETV6, LAPTM5, and MCOLN3. These genes support and extend the functional regulatory network centered on TRPM4, offering new insights into the complex interactions of obesity-related signaling pathways. Although this research provides innovative bioinformatic evidence for the development of obesity, some limitations remain. For example, some analyses rely on data from public databases, where sample size and data quality may limit the applicability of the findings. Additionally, these key genes need validation through larger clinical samples, animal models, or in vitro functional tests to better understand their functions at the tissue level and within the broader metabolic framework. Moving forward, we will continue to explore the expression patterns and regulatory mechanisms of ETV6, LAPTM5, and MCOLN3 across various obesity types. Our goal is to thoroughly understand their roles in TRPM4-related signaling pathways and the immunometabolism disruptions associated with obesity. This research aims to strengthen the scientific foundation and offer new directions for early diagnosis, biomarker identification, and targeted therapies for obesity.\u003c/p\u003e"},{"header":"4. Methods","content":"\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Data acquisition\u003c/h2\u003e \u003cp\u003eThe obese datasets employed in the study included GSE156906 (platform: GPL24676), GSE25401 (platform: GPL6244), and GSE155960 (platform: GPL24676), all of which were retrieved from the Gene Expression Omnibus (GEO) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The analysis utilized subcutaneous adipose tissue samples from two cohorts: the training set (GSE156906) included 25 obese patients and 14 healthy controls, and the validation set (GSE25401) comprised 30 obese patients and 26 healthy controls. The single-cell RNA sequencing (scRNA-seq) dataset (GSE155960) included matrix files of CD45-positive and CD45-negative cells derived from subcutaneous adipose tissue, with six samples each from the obese and control groups.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e4.2 Evaluation of TRPM4 expression and gene set enrichment analysis (GSEA)\u003c/h2\u003e \u003cp\u003eTRPM4 expression was assessed between obese and control samples in the GSE156906 dataset, employing the Wilcoxon test (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). GSEA was performed to investigate TRPM4-associated signaling pathways in obesity, utilizing the \"c2.cp.v2024.1.Hs.symbols.gmt\" collection from the Molecular Signatures Database (MSigDB) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gsea-msigdb.org/\u003c/span\u003e\u003cspan address=\"https://www.gsea-msigdb.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) as the gene set background. Using the cor function from the R stats package (v 4.3.3; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.R-project.org/\u003c/span\u003e\u003cspan address=\"https://www.R-project.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), we computed Spearman correlations between TRPM4 and all other genes in GSE156906, then ranked the genes in descending order to create a TRPM4-associated gene list. Thereafter, we performed GSEA for TRPM4 using the clusterProfiler package (v 4.8.3)[\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e], employing significance criteria of |NES| \u0026gt; 1 and adjusted P-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e4.3 construction of molecular regulatory networks\u003c/h2\u003e \u003cp\u003eThe identification of microRNAs (miRNAs) and transcription factors (TFs) regulating TRPM4 enhanced the understanding of the gene's regulatory logic and facilitated the discovery of potential key regulators. Specifically, the StarBase v3.0 database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://starbase.sysu.edu.cn/\u003c/span\u003e\u003cspan address=\"https://starbase.sysu.edu.cn/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was searched for miRNAs associated with TRPM4, and these were filtered based on the criterion of \"pancancerNum\u0026thinsp;\u0026gt;\u0026thinsp;6\". Additionally, TFs regulating TRPM4 were retrieved from the Cistrome database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://cistrome.org/\u003c/span\u003e\u003cspan address=\"http://cistrome.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) based on the criterion of a \"regulatory potential score\" exceeding 0.7. The miRNA-TRPM4 and TF-TRPM4 regulatory networks were constructed and visualized using Cytoscape software (v 3.8.0)[\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e4.4 Processing of scRNA-seq data and identification of crucial cell types\u003c/h2\u003e \u003cp\u003eQuality control (QC) procedures for dataset GSE155960 were carried out utilizing the \"PercentageFeatureSet function\" from the Seurat package (v 5.0.1)[\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e]. Following this filtering, all downstream analyses were performed exclusively with the functions provided in this package. Cells with an nFeature_RNA (number of detected genes) below 200 or above 4,000 were excluded. Cells with an nCount_RNA (total number of RNA counts per cell) below 500 or above 20,000 were also filtered out. Subsequently, the \"NormalizeData function\" was employed to standardize the data. Following this, the top 2,000 highly variable genes (HVGs) were identified utilizing the variance-stabilizing transformation (vst), which was implemented in the \"FindVariableFeatures function\". The top 10 genes were labeled utilizing the \"LabelPoints function\" for visualization. After that, the data were normalized via the \"Scale Data function\". Subsequently, principal component analysis (PCA) was applied to reduce data dimensionality, and the Harmony method was applied for batch effect correction. PCA facilitated the embedding of cells into a low-dimensional space, whereas Harmony efficiently removed batch-related discrepancies through an iterative optimization algorithm conducted within the PCA-transformed space. The \"JackStraw function\" was used to identify the most significant principal components (PCs) by selecting those with a higher number of genes with low P values (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). The \"ElbowPlot function\" was employed to generate a scree plot, where inflection points served as indicators to identify the optimal number of PCs for subsequent analysis. Thereafter, unsupervised clustering was conducted with a resolution parameter of 0.5 to identify the distinct cell populations. The Uniform Manifold Approximation and Projection (UMAP) method allowed for unsupervised clustering and provided an unbiased visualization of cell clusters in two dimensions. Additionally, to identify cluster-specific marker genes, we employed the \"FindAllMarkers\" function, applying thresholds of |log2 fold change (FC)| \u0026gt; 1 and an adjusted P-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05. The SingleR package (v 2.2.0)[\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e] and the CellMarker database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://xteam.xbio.top/CellMarker/\u003c/span\u003e\u003cspan address=\"http://xteam.xbio.top/CellMarker/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) were used as the primary references for cell type annotation, while the HumanPrimaryCellAtlasData, BlueprintEncodeData, and ImmuneCellExpressionData were used as supplementary references. Based on the clustering results and concerning relevant literature[\u003cspan additionalcitationids=\"CR66\" citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e], cell subclusters were annotated with corresponding cell types. The distribution of annotated cell types in the GSE155960 dataset was visualized using stacked plots generated with ggplot2 (v 3.5.1)[\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e]. Subsequently, crucial cell types were identified as those exhibiting significant differences (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) in TRPM4 expression between obese and control groups, as determined by the Wilcoxon test.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e4.5 Cellular communication and pseudo-temporal trajectory analyses\u003c/h2\u003e \u003cp\u003eThe CellChat package (v 1.6.1)[\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e] was employed to analyse patterns of cell-to-cell communication, facilitating the quantification and characterization of interactions between crucial cell types and other annotated cell types. The reference human ligand-receptor database, CellChatDB (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.cellchat.org/\u003c/span\u003e\u003cspan address=\"http://www.cellchat.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), was employed in the analysis to evaluate human intercellular communication. Additionally, the same package was employed to analyse ligand-receptor interaction pairs across these annotated cell types. Furthermore, dimensionality reduction and clustering were performed on the crucial cell types again, utilizing the same methodologies as those previously employed for processing the scRNA-seq data. Moreover, the differentiation trajectories of the crucial cell types were reconstructed using the Monocle package (v 2.28.0)[\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e], and the trajectories for each crucial cell type were depicted utilizing the \"DDRTree function\" within the same package. To visualize the dynamic expression patterns of significant genes in crucial cell types, we generated an expression heatmap across pseudotime using the plot_pseudo_time_heatmap function. Adipocyte progenitor cells (APCs) play a pivotal role in obesity. Consequently, dimensionality reduction and clustering, along with pseudo-temporal trajectory analyses, were conducted on APCs. The dynamic expression pattern of TRPM4 during APC differentiation was delineated utilizing the monocle package (v 2.28.0).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e4.6 Gene set variation analysis (GSVA) and prediction of TF activity\u003c/h2\u003e \u003cp\u003eA comprehensive assessment of the signaling pathways implicated in both obese and control groups was carried out by employing GSVA on all samples from GSE155960.\u003c/p\u003e \u003cp\u003eThe GSVA package (v 1.53.28)[\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e] was utilized to analyze all samples in GSE155960, calculating enrichment scores for each sample across gene sets, which subsequently facilitated Gene Ontology (GO) enrichment analysis. Subsequently, we performed differential analysis between the obese and control groups with the limma package (v 3.56.2)[\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e] (|t| \u0026gt; 2, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Additionally, the enrichment of these pathways within each annotated cell type was analyzed. To predict the transcription factors activated under obese conditions, the activities of gene regulatory networks (GRNs) and TFs were identified utilizing the SCENIC Python workflow (v 0.9.1)[\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e] with default parameters (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/aertslab/pySCENIC\u003c/span\u003e\u003cspan address=\"https://github.com/aertslab/pySCENIC\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The list of human TF genes was compiled from the same source. The binary matrix was used to identify activated TFs, and the specificity of key TFs in different cell types was analysed alongside the regulon specificity score (RSS) values.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e4.7 High-dimensional weighted gene co-expression network analysis (hdWGCNA)\u003c/h2\u003e \u003cp\u003eThe hdWGCNA was conducted to identify significant genes linked to critical cell types in the samples derived from GSE155960. The \"MetacellsByGroups function\" of the hdWGCNA package (v 0.4.0)[\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e] was employed to construct metacells for critical cell types within each sample from GSE155960. During this process, the k-nearest neighbour parameter was set to 25, and the maximum number of cells that could be shared between any two metacells was limited to 10. Subsequently, following the standard hdWGCNA analysis pipeline, the \"TestSoftPowers function\" was employed to evaluate and select an appropriate soft power threshold. The minimum soft-thresholding power yielding a scale-free topology fit index\u0026thinsp;≧\u0026thinsp;0.8 was selected. Following this determination, a weighted co-expression network was built using the \"ConstructNetwork\" function. This function was then applied to perform topological overlap matrix (TOM)-based hierarchical clustering on the clustered samples, enabling a clustering dendrogram to be established for module identification. The \"ModuleCorrelogram function\" was employed to calculate correlations between different modules (|correlation coefficient (cor)| \u0026gt; 0.3 and P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Furthermore, the \"FindDMEs function\" was employed to conduct the Wilcoxon test on genes within each module in both the obese and control groups (adj.P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). We defined key modules as those exhibiting significantly distinct expression profiles, and termed their constituent genes as key module genes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e4.8 Analysis of differential gene expression\u003c/h2\u003e \u003cp\u003eDifferentially expressed genes (DEGs) between obese and control groups in GSE156906 were identified utilizing the DESeq2 package (v 1.40.2)[\u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e], with the following criteria applied: |log₂ FC| \u0026gt; 0.5 and P\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Furthermore, A volcano plot visualizing the differentially expressed genes (DEGs) was generated using the ggplot2 package (v 3.5.1), and a heatmap illustrating DEGs was generated with the ComplexHeatmap package (v 2.16.0)[\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e4.9 Identification and related functional analysis of candidate genes\u003c/h2\u003e \u003cp\u003eThe ggvenn package (v 0.1.10)[\u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e] was used to visualize the intersection between DEGs and key module genes, aiming to identify candidate genes potentially linked to TRPM4 in crucial obesity-related cell types. Following this, the clusterProfiler package (v 4.8.3) was used to perform Gene Ontology (GO) enrichment analysis on the candidate genes (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05), thereby clarifying their molecular functions (MFs), cellular components (CCs), and biological processes (BPs). Subsequently, the genes enriched in each BP, CC, and MF category were ranked in descending order according to their counts. We displayed the top 10 most significantly enriched pathways for BP, CC, and MF, respectively. A protein-protein interaction (PPI) network was constructed using the Retrieval of Interacting Genes (STRING) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.string-db.org\u003c/span\u003e\u003cspan address=\"https://www.string-db.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e; score\u0026thinsp;\u0026gt;\u0026thinsp;0.15) to explore interactions among candidate genes and was then visualized in Cytoscape (v 3.8.0).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e4.10 Discernment of key genes through machine learning, receiver operating characteristic (ROC) analysis, and expression validation\u003c/h2\u003e \u003cp\u003eAll samples within GSE156906 were subjected to analysis utilizing three different machine learning approaches, which leveraged candidate genes to pinpoint characteristic genes linked to obesity. Initially, Least Absolute Shrinkage and Selection Operator (LASSO) regression was employed to minimize redundant information and eliminate irrelevant variables, ultimately shrinking the coefficients of unimportant variables to zero while retaining only a small number of significant variables with non-zero coefficients. We applied LASSO regression to the candidate genes via the \"cv.glmnet\" function in the glmnet package (v 4.1.8)[\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e], with five-fold cross-validation applied. Genes with non-zero coefficients at the optimal lambda (lambda.min), which yielded the lowest cross-validation error, were selected as the LASSO feature set. Concurrently, we employed Support Vector Machine-Recursive Feature Elimination (SVM-RFE), a feature selection technique that sequentially removes features with the least importance based on the SVM's maximum margin principle. It trains the model iteratively, ranks the features based on their scores, and progressively eliminates the lowest-scoring features until the desired set is selected. The \"svmRFE function\" from the e1071 package (v 1.7.16)[\u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e79\u003c/span\u003e] was employed to construct an SVM-RFE model. This model utilized candidate genes as feature variables and implemented five-fold cross-validation. Additionally, the parameter halve.above was set to 100 to evaluate the feature importance and ranking of each candidate gene. During the model iteration process, the classification error rate and accuracy for each round of feature combinations were recorded. The feature combination corresponding to the minimum error rate was selected as the final optimal gene set, which was defined as the SVM-RFE characteristic genes. The core of the Boruta feature selection method can be distilled into two primary steps: constructing shadow features and employing a random forest-based voting mechanism. If an original feature is significantly more important than the shadow features, it is classified as important; otherwise, it is regarded as unimportant. Specifically, the Boruta algorithm was employed for feature selection using the Boruta package (v 8.0.0)[\u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e80\u003c/span\u003e], was employed for feature selection from candidate genes. Features with importance scores exceeding shadowMax were categorized as \"important\" and subsequently labeled as Boruta characteristic genes. Subsequently, the ggvenn package (v 0.1.10) was utilized to identify genes that overlapped among those selected by the LASSO, SVM-RFE, and Boruta algorithms. These overlapping genes were then designated as common characteristic genes. Furthermore, the pROC package (v 1.18.5)[\u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e81\u003c/span\u003e] was employed to conduct ROC analysis on common characteristic genes, and the area under the curve (AUC) values were calculated. We defined genes with AUC values exceeding 0.7, indicating good obese-control discriminative ability, as candidate key genes. Additionally, the expression profiles of the candidate key genes were compared between obese and control samples in the GSE156906 and GSE25401 datasets using the Wilcoxon test. The candidate key genes showing notably differential expression between obese and control samples (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) with a consistent trend in both GSE156906 and GSE25401 were identified as key genes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e4.11 Establishment and assessment of a nomogram\u003c/h2\u003e \u003cp\u003eA nomogram was constructed utilizing the rms package (v 6.8.1)[\u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e82\u003c/span\u003e] to evaluate the predictive probability of obesity onset based on key genes in GSE156906. A nomogram was developed where each key gene contributed a specific point value, and the summed score predicted obesity risk. The model's calibration was quantified by a curve generated with the regplot package (v 1.1)[\u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e83\u003c/span\u003e] and confirmed by a non-significant Hosmer-Lemeshow test statistic (P\u0026thinsp;\u0026gt;\u0026thinsp;0.05). We employed decision curve analysis (DCA) with the ggDCA package (v 1.1)[\u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e84\u003c/span\u003e] to evaluate the clinical net benefit of the nomogram.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec25\" class=\"Section2\"\u003e \u003ch2\u003e4.12 Subcellular localization and gene-gene interaction (GGI) network construction\u003c/h2\u003e \u003cp\u003eSubcellular localization analysis was carried out to acquire a deeper understanding of the roles played by the key genes. In particular, FASTA-formatted sequence files corresponding to the key genes were retrieved from the National Center for Biotechnology Information (NCBI) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/gene/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/gene/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The subcellular localization of the key genes was predicted using the mRNALocater database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://bio-bigdata.cn/mRNALocater/\u003c/span\u003e\u003cspan address=\"http://bio-bigdata.cn/mRNALocater/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). To further elucidate their functional relationships and shared biological roles, a GGI network was constructed via the GeneMANIA database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.genemania.org/\u003c/span\u003e\u003cspan address=\"http://www.genemania.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section2\"\u003e \u003ch2\u003e4.13 Drugs/compounds prediction\u003c/h2\u003e \u003cp\u003eThe identification of potential drugs and compounds capable of targeting specific key genes was facilitated by utilizing the Drug Signatures Database (DSigDB) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://dsigdb.tanlab.org/DSigDBv1.0/\u003c/span\u003e\u003cspan address=\"https://dsigdb.tanlab.org/DSigDBv1.0/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Subsequently, a drug/compound-key gene interaction network was generated using Cytoscape software (v 3.8.0).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section2\"\u003e \u003ch2\u003e4.14 Reverse transcription-quantitative PCR (RT-qPCR)\u003c/h2\u003e \u003cp\u003eTo experimentally validate the differential expression of key genes between obese and control samples, RT-qPCR analysis was performed using subcutaneous adipose tissue samples obtained from five obese patients, alongside five samples from control subjects, sourced from shenzhen qianHai taikang hospital. This research received approval from the medical ethics committee of Shenzhen Qianhai Taikang Hospital and all participating patients provided their informed consent by signing the relevant form. Total RNA was isolated from whole blood samples employing the TRIzol kit (Vazyme Biotech Co., Ltd., Catalog No. R401-01, Nanjing, China), Total RNA was reverse-transcribed into cDNA using the HP All-in-one qRT Master Mix II (Yungen Biotechnology, Cat. No. 24Y0124) according to the manufacturer's protocol. The resulting cDNA was diluted 5- to 20-fold with RNase/DNase-free ddH\u003csub\u003e2\u003c/sub\u003eO, and a 10 \u0026micro;L qPCR reaction mixture was prepared containing 3 \u0026micro;L of diluted cDNA, 5 \u0026micro;L of 2\u0026times; Universal Blue SYBR Green qPCR Master Mix, and 1 \u0026micro;L each of forward and reverse primers (10 \u0026micro;M each). Amplification was performed for 40 cycles on a CFX Connect real-time PCR system (BIO-RAD, Cat. No. XLFZ006) using a program that omitted the pre-denaturation step. The program details are outlined in \u003cb\u003eSupplementary Table\u0026nbsp;1\u003c/b\u003e. The primer sequences for key genes are listed in Table S2. GAPDH served as the reference gene, and relative expression levels were determined using the 2\u003csup\u003e\u0026minus;△△CT\u003c/sup\u003e method.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003e4.15 Statistical analysis\u003c/h2\u003e \u003cp\u003eStatistical analyses were primarily conducted in R (v 4.3.3), employing the Wilcoxon test for group comparisons, while RT-qPCR Ct values were analyzed by unpaired two-tailed t-tests in GraphPad Prism (v 5.0). A uniform significance threshold of P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was applied to all tests.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements (optional)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to express our sincere gratitude to all individuals and organizations who supported and assisted us throughout this research. Special thanks to the following Yanfen Li.In conclusion, we extend our thanks to everyone who has supported and assisted us along the way. Without your support, this research would not have been possible.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eYang Shen:Conceptualization, Data curation, Validation, Visualization, Writing\u0026ndash;original draft, Writing\u0026ndash;review \u0026amp; editing. Yanfen Li:Data curation, Validation, Visualization, Writing\u0026ndash;review \u0026amp; editing. Cijing Cai:Visualization, Writing\u0026ndash;review \u0026amp; editing. Shenghong Lei: Conceptualization, Supervision, Writing\u0026ndash;review \u0026amp; editing. This work currently described has not been published, is not being considered for publication elsewhere, and its publication was approved by all authors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets ANALYZED for this study can be found in the [Gene Expression Omnibus (GEO) database] [http://www.ncbi.nlm.nih.gov/geo/].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was conducted in accordance with the Declaration of Helsinki and received approval from the Medical Ethics Committee of Shenzhen Qianhai Taikang Hospital (Approval No.: AF-LLPJ-20251022-02; Date: October 22, 2025). All patients provided written informed consent when clinical samples were collected for RT-qPCR experiments to ensure that the research process was in accordance with ethical norms and that the patients\u0026apos; rights and wishes were fully respected.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to publish\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot Applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research received no specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eRubino, F. et al. Definition and diagnostic criteria of clinical obesity. \u003cem\u003eLancet Diabetes Endocrinol.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 221\u0026ndash;262 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNovelli, G., Cassadonte, C., Sbraccia, P. \u0026amp; Biancolella, M. Genetics: A Starting Point for the Prevention and the Treatment of Obesity. \u003cem\u003eNutrients\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJebeile, H., Kelly, A. S., O'Malley, G. \u0026amp; Baur, L. A. Obesity in children and adolescents: epidemiology, causes, assessment, and management. \u003cem\u003eLancet Diabetes Endocrinol.\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e, 351\u0026ndash;365 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKoenen, M., Hill, M. A., Cohen, P., Sowers, J. R. \u0026amp; Obesity Adipose Tissue and Vascular Dysfunction. \u003cem\u003eCirc. Res.\u003c/em\u003e \u003cb\u003e128\u003c/b\u003e, 951\u0026ndash;968 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePerdomo, C. M., Cohen, R. V., Sumithran, P., Cl\u0026eacute;ment, K. \u0026amp; Fr\u0026uuml;hbeck, G. Contemporary medical, device, and surgical therapies for obesity in adults. \u003cem\u003eLancet\u003c/em\u003e \u003cb\u003e401\u003c/b\u003e, 1116\u0026ndash;1130 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYanovski, S. Z. \u0026amp; Yanovski, J. A. Approach to Obesity Treatment in Primary Care: A Review. \u003cem\u003eJAMA Intern. Med.\u003c/em\u003e \u003cb\u003e184\u003c/b\u003e, 818\u0026ndash;829 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eElmaleh-Sachs, A. et al. Obesity Management in Adults: A Review. \u003cem\u003eJama\u003c/em\u003e \u003cb\u003e330\u003c/b\u003e, 2000\u0026ndash;2015 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM\u0026uuml;ller, T. D., Bl\u0026uuml;her, M., Tsch\u0026ouml;p, M. H. \u0026amp; DiMarchi, R. D. Anti-obesity drug discovery: advances and challenges. \u003cem\u003eNat. Rev. Drug Discov\u003c/em\u003e. \u003cb\u003e21\u003c/b\u003e, 201\u0026ndash;223 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFu, W. et al. Persistent activation of TRPM4 triggers necrotic cell death characterized by sodium overload. \u003cem\u003eNat. Chem. Biol.\u003c/em\u003e \u003cb\u003e21\u003c/b\u003e, 1238\u0026ndash;1249 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLaunay, P. et al. TRPM4 is a Ca2+-activated nonselective cation channel mediating cell membrane depolarization. \u003cem\u003eCell\u003c/em\u003e \u003cb\u003e109\u003c/b\u003e, 397\u0026ndash;407 (2002).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKappel, S., Ross-Kaschitza, D., Hauert, B., Rother, K. \u0026amp; Peinelt, C. p53 alters intracellular Ca(2+) signaling through regulation of TRPM4. \u003cem\u003eCell. Calcium\u003c/em\u003e. \u003cb\u003e104\u003c/b\u003e, 102591 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCs\u0026iacute;pő, T. et al. A Central Role for TRPM4 in Ca(2+)-Signal Amplification and Vasoconstriction. \u003cem\u003eInt J. Mol. Sci\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBishnoi, M., Khare, P., Brown, L. \u0026amp; Panchal, S. K. Transient receptor potential (TRP) channels: a metabolic TR(i)P to obesity prevention and therapy. \u003cem\u003eObes. Rev.\u003c/em\u003e \u003cb\u003e19\u003c/b\u003e, 1269\u0026ndash;1292 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZsombok, A. \u0026amp; Derbenev, A. V. TRP Channels as Therapeutic Targets in Diabetes and Obesity. \u003cem\u003ePharmaceuticals (Basel)\u003c/em\u003e 9 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarigo, V., Courville, K., Hsu, W. H., Feng, J. M. \u0026amp; Cheng, H. TRPM4 impacts on Ca2\u0026thinsp;+\u0026thinsp;signals during agonist-induced insulin secretion in pancreatic beta-cells. \u003cem\u003eMol. Cell. Endocrinol.\u003c/em\u003e \u003cb\u003e299\u003c/b\u003e, 194\u0026ndash;203 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTran, T. D. et al. Histamine-induced Ca\u0026sup2;⁺ signalling is mediated by TRPM4 channels in human adipose-derived stem cells. \u003cem\u003eBiochem. J.\u003c/em\u003e \u003cb\u003e463\u003c/b\u003e, 123\u0026ndash;134 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSlovin, S. et al. Single-Cell RNA Sequencing Analysis: A Step-by-Step Overview. \u003cem\u003eMethods Mol. Biol.\u003c/em\u003e \u003cb\u003e2284\u003c/b\u003e, 343\u0026ndash;365 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJovic, D. et al. Single-cell RNA sequencing technologies and applications: A brief overview. \u003cem\u003eClin. Transl Med.\u003c/em\u003e \u003cb\u003e12\u003c/b\u003e, e694 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSo, J. et al. Robust single-nucleus RNA sequencing reveals depot-specific cell population dynamics in adipose tissue remodeling during obesity. \u003cem\u003eElife\u003c/em\u003e 13 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAcharjee, A., Wijesinghe, S. N., Russ, D., Gkoutos, G. \u0026amp; Jones, S. W. Cross-species transcriptomics identifies obesity associated genes between human and mouse studies. \u003cem\u003eJ. Transl Med.\u003c/em\u003e \u003cb\u003e22\u003c/b\u003e, 592 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, D. et al. Employing single-cell RNA sequencing coupled with an array of bioinformatics approaches to ascertain the shared genetic characteristics between osteoporosis and obesity. \u003cem\u003eBone Joint Res.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 573\u0026ndash;587 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGhosh, S. et al. Plasma Membrane Channel TRPM4 Mediates Immunogenic Therapy-Induced Necrosis. \u003cem\u003eCancer Res.\u003c/em\u003e \u003cb\u003e83\u003c/b\u003e, 3115\u0026ndash;3130 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu, F. et al. Loss of the TRPM4 channel in humans causes immune dysregulation with defective monocyte migration. \u003cem\u003eJ. Allergy Clin. Immunol.\u003c/em\u003e \u003cb\u003e154\u003c/b\u003e, 792\u0026ndash;806 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSabaratnam, R. \u0026amp; Svenningsen, P. Adipocyte-Endothelium Crosstalk in Obesity. \u003cem\u003eFront. Endocrinol. (Lausanne)\u003c/em\u003e. \u003cb\u003e12\u003c/b\u003e, 681290 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, M., Qian, M., Kyler, K. \u0026amp; Xu, J. Adipose Tissue-Endothelial Cell Interactions in Obesity-Induced Endothelial Dysfunction. \u003cem\u003eFront. Cardiovasc. Med.\u003c/em\u003e \u003cb\u003e8\u003c/b\u003e, 681581 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGraupera, M. \u0026amp; Claret, M. Endothelial Cells: New Players in Obesity and Related Metabolic Disorders. \u003cem\u003eTrends Endocrinol. Metab.\u003c/em\u003e \u003cb\u003e29\u003c/b\u003e, 781\u0026ndash;794 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGarland, C. J. et al. TRPM4 inhibitor 9-phenanthrol activates endothelial cell intermediate conductance calcium-activated potassium channels in rat isolated mesenteric artery. \u003cem\u003eBr. J. Pharmacol.\u003c/em\u003e \u003cb\u003e172\u003c/b\u003e, 1114\u0026ndash;1123 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWei, S. et al. Comparison of Anti-oncotic Effect of TRPM4 Blocking Antibody in Neuron, Astrocyte and Vascular Endothelial Cell Under Hypoxia. \u003cem\u003eFront. Cell. Dev. Biol.\u003c/em\u003e \u003cb\u003e8\u003c/b\u003e, 562584 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBecerra, A. et al. Transient receptor potential melastatin 4 inhibition prevents lipopolysaccharide-induced endothelial cell death. \u003cem\u003eCardiovasc. Res.\u003c/em\u003e \u003cb\u003e91\u003c/b\u003e, 677\u0026ndash;684 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, C., Naruse, K. \u0026amp; Takahashi, K. Role of the TRPM4 Channel in Cardiovascular Physiology and Pathophysiology. \u003cem\u003eCells\u003c/em\u003e. 7 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbriel, H., Syam, N., Sottas, V., Amarouch, M. Y. \u0026amp; Rougier, J. S. TRPM4 channels in the cardiovascular system: physiology, pathophysiology, and pharmacology. \u003cem\u003eBiochem. Pharmacol.\u003c/em\u003e \u003cb\u003e84\u003c/b\u003e, 873\u0026ndash;881 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiao, L. et al. Circulating miR-3659 may be a potential biomarker of dyslipidemia in patients with obesity. \u003cem\u003eJ. Transl Med.\u003c/em\u003e \u003cb\u003e17\u003c/b\u003e, 25 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, Y. et al. Identification of metabolism related biomarkers in obesity based on adipose bioinformatics and machine learning. \u003cem\u003eJ. Transl Med.\u003c/em\u003e \u003cb\u003e22\u003c/b\u003e, 986 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShi, S. et al. Identification of hub genes in the crosstalk between type 2 diabetic nephropathy and obesity according to bioinformatics analysis. \u003cem\u003eAdipocyte\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 2423723 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDe, S. et al. Steric mechanism of auto-inhibitory regulation of specific and non-specific DNA binding by the ETS transcriptional repressor ETV6. \u003cem\u003eJ. Mol. Biol.\u003c/em\u003e \u003cb\u003e426\u003c/b\u003e, 1390\u0026ndash;1406 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBohlander, S. K. ETV6: a versatile player in leukemogenesis. \u003cem\u003eSemin Cancer Biol.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e, 162\u0026ndash;174 (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEldeeb, M. et al. Ontogeny shapes the ability of ETV6::RUNX1 to enhance hematopoietic stem cell self-renewal and disrupt early lymphopoiesis. \u003cem\u003eLeukemia\u003c/em\u003e \u003cb\u003e38\u003c/b\u003e, 455\u0026ndash;459 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHock, H. et al. Tel/Etv6 is an essential and selective regulator of adult hematopoietic stem cell survival. \u003cem\u003eGenes Dev.\u003c/em\u003e \u003cb\u003e18\u003c/b\u003e, 2336\u0026ndash;2341 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKeller, M. et al. Genome-wide DNA promoter methylation and transcriptome analysis in human adipose tissue unravels novel candidate genes for obesity. \u003cem\u003eMol. Metab.\u003c/em\u003e \u003cb\u003e6\u003c/b\u003e, 86\u0026ndash;100 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, Y. et al. LAPTM5 mediates immature B cell apoptosis and B cell tolerance by regulating the WWP2-PTEN-AKT pathway. \u003cem\u003eProc. Natl. Acad. Sci. U S A\u003c/em\u003e. \u003cb\u003e119\u003c/b\u003e, e2205629119 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGlowacka, W. K., Alberts, P., Ouchida, R., Wang, J. Y. \u0026amp; Rotin, D. LAPTM5 protein is a positive regulator of proinflammatory signaling pathways in macrophages. \u003cem\u003eJ. Biol. Chem.\u003c/em\u003e \u003cb\u003e287\u003c/b\u003e, 27691\u0026ndash;27702 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang, B. et al. Lysosomal protein transmembrane 5 promotes lung-specific metastasis by regulating BMPR1A lysosomal degradation. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 4141 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, M. M. et al. The function and mechanism of LAPTM5 in diseases. \u003cem\u003eBiomed. Pharmacother\u003c/em\u003e. \u003cb\u003e178\u003c/b\u003e, 117237 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu, X. et al. Obesity activates a program of lysosomal-dependent lipid metabolism in adipose tissue macrophages independently of classic activation. \u003cem\u003eCell. Metab.\u003c/em\u003e \u003cb\u003e18\u003c/b\u003e, 816\u0026ndash;830 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYan, K. et al. Integrative analyses of hub genes and their association with immune infiltration in adipose tissue, liver tissue and skeletal muscle of obese patients after bariatric surgery. \u003cem\u003eAdipocyte\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e, 190\u0026ndash;201 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun, M. et al. Transcriptome combined with Mendelian randomization to identify key genes related to polyamine metabolism in childhood obesity and elucidate their molecular regulatory mechanisms. \u003cem\u003eSci. Rep.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e, 17799 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang, L. et al. Lysosomal-associated protein transmembrane 5 ameliorates non-alcoholic steatohepatitis by promoting the degradation of CDC42 in mice. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cb\u003e14\u003c/b\u003e, 2654 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim, S. W. et al. Palmitoylation controls trafficking of the intracellular Ca(2+) channel MCOLN3/TRPML3 to regulate autophagy. \u003cem\u003eAutophagy\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e, 327\u0026ndash;340 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLei, Y. \u0026amp; Klionsky, D. J. MCOLN3/TRPML3 bridges the regulation of autophagosome biogenesis by PtdIns3P and the calcium channel. \u003cem\u003eAutophagy\u003c/em\u003e \u003cb\u003e19\u003c/b\u003e, 377\u0026ndash;378 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSpix, B. et al. Lung emphysema and impaired macrophage elastase clearance in mucolipin 3 deficient mice. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 318 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartina, J. A., Lelouvier, B. \u0026amp; Puertollano, R. The calcium channel mucolipin-3 is a novel regulator of trafficking along the endosomal pathway. \u003cem\u003eTraffic\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e, 1143\u0026ndash;1156 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQadir, M. et al. Cyclosporin A is a broad-spectrum multidrug resistance modulator. \u003cem\u003eClin. Cancer Res.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e, 2320\u0026ndash;2326 (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFuhrmann, A. et al. Molecular mechanisms underlying the effects of cyclosporin A and sirolimus on glucose and lipid metabolism in liver, skeletal muscle and adipose tissue in an in vivo rat model. \u003cem\u003eBiochem. Pharmacol.\u003c/em\u003e \u003cb\u003e88\u003c/b\u003e, 216\u0026ndash;228 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang, M. et al. Cyclosporin A attenuates weight gain and improves glucose tolerance in diet-induced obese mice. \u003cem\u003eMol. Cell. Endocrinol.\u003c/em\u003e \u003cb\u003e370\u003c/b\u003e, 96\u0026ndash;102 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, B. et al. Nutrigenomic regulation of adipose tissue development - role of retinoic acid: A review. \u003cem\u003eMeat Sci.\u003c/em\u003e \u003cb\u003e120\u003c/b\u003e, 100\u0026ndash;106 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu, S. et al. Adipose tissue plays a major role in retinoic acid-mediated metabolic homoeostasis. \u003cem\u003eAdipocyte\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e, 47\u0026ndash;55 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, B. et al. Retinoic acid induces white adipose tissue browning by increasing adipose vascularity and inducing beige adipogenesis of PDGFRα(+) adipose progenitors. \u003cem\u003eCell. Discov\u003c/em\u003e. \u003cb\u003e3\u003c/b\u003e, 17036 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMogal, A. \u0026amp; Abdulkadir, S. A. Effects of Histone Deacetylase Inhibitor (HDACi); Trichostatin-A (TSA) on the expression of housekeeping genes. \u003cem\u003eMol. Cell. Probes\u003c/em\u003e. \u003cb\u003e20\u003c/b\u003e, 81\u0026ndash;86 (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu, J. C. et al. Trichostatin A modulates thiazolidinedione-mediated suppression of tumor necrosis factor α-induced lipolysis in 3T3-L1 adipocytes. \u003cem\u003ePLoS One\u003c/em\u003e. \u003cb\u003e8\u003c/b\u003e, e71517 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, X. et al. Trichostatin A regulates fibro/adipogenic progenitor adipogenesis epigenetically and reduces rotator cuff muscle fatty infiltration. \u003cem\u003eJ. Orthop. Res.\u003c/em\u003e \u003cb\u003e39\u003c/b\u003e, 1452\u0026ndash;1462 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu, T. et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. \u003cem\u003eInnov. (Camb)\u003c/em\u003e. \u003cb\u003e2\u003c/b\u003e, 100141 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. \u003cem\u003eGenome Res.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 2498\u0026ndash;2504 (2003).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHao, Y. et al. Dictionary learning for integrative, multimodal and scalable single-cell analysis. \u003cem\u003eNat. Biotechnol.\u003c/em\u003e \u003cb\u003e42\u003c/b\u003e, 293\u0026ndash;304 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAran, D. et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. \u003cem\u003eNat. Immunol.\u003c/em\u003e \u003cb\u003e20\u003c/b\u003e, 163\u0026ndash;172 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, Y. et al. Development and validation of a hypoxia- and mitochondrial dysfunction- related prognostic model based on integrated single-cell and bulk RNA sequencing analyses in gastric cancer. \u003cem\u003eFront. Immunol.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e, 1419133 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin, S. et al. Single-nucleus RNA sequencing reveals heterogenous microenvironments and specific drug response between cervical squamous cell carcinoma and adenocarcinoma. \u003cem\u003eEBioMedicine\u003c/em\u003e. 97, 104846 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, H. et al. Integrated immunological analysis of single-cell and bulky tissue transcriptomes reveals the role of interactions between M0 macrophages and na\u0026iuml;ve CD4(+) T cells in the immunosuppressive microenvironment of cervical cancer. \u003cem\u003eComput. Biol. Med.\u003c/em\u003e \u003cb\u003e163\u003c/b\u003e, 107151 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGustavsson, E. K., Zhang, D., Reynolds, R. H., Garcia-Ruiz, S. \u0026amp; Ryten, M. ggtranscript: an R package for the visualization and interpretation of transcript isoforms using ggplot2. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cb\u003e38\u003c/b\u003e, 3844\u0026ndash;3846 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJin, S., Plikus, M. V. \u0026amp; Nie, Q. CellChat for systematic analysis of cell-cell communication from single-cell transcriptomics. \u003cem\u003eNat. Protoc.\u003c/em\u003e \u003cb\u003e20\u003c/b\u003e, 180\u0026ndash;219 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. \u003cem\u003eNat. Biotechnol.\u003c/em\u003e \u003cb\u003e32\u003c/b\u003e, 381\u0026ndash;386 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH\u0026auml;nzelmann, S., Castelo, R. \u0026amp; Guinney, J. GSVA: gene set variation analysis for microarray and RNA-seq data. \u003cem\u003eBMC Bioinform.\u003c/em\u003e \u003cb\u003e14\u003c/b\u003e, 7 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRitchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cb\u003e43\u003c/b\u003e, e47 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVan de Sande, B. et al. A scalable SCENIC workflow for single-cell gene regulatory network analysis. \u003cem\u003eNat. Protoc.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e, 2247\u0026ndash;2276 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorabito, S., Reese, F., Rahimzadeh, N., Miyoshi, E. \u0026amp; Swarup, V. hdWGCNA identifies co-expression networks in high-dimensional transcriptomics data. \u003cem\u003eCell. Rep. Methods\u003c/em\u003e. \u003cb\u003e3\u003c/b\u003e, 100498 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLove, M. I., Huber, W. \u0026amp; Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. \u003cem\u003eGenome Biol.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e, 550 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGu, Z. Complex heatmap visualization. \u003cem\u003eImeta\u003c/em\u003e \u003cb\u003e1\u003c/b\u003e, e43 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOrt\u0026uacute;zar, M., Riesco, R., Criado, M., Alonso, M. D. P. \u0026amp; Trujillo, M. E. Unraveling the dynamic interplay of microbial communities associated to Lupinus angustifolius in response to environmental and cultivation conditions. \u003cem\u003eSci. Total Environ.\u003c/em\u003e \u003cb\u003e946\u003c/b\u003e, 174277 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFriedman, J., Hastie, T. \u0026amp; Tibshirani, R. Regularization Paths for Generalized Linear Models via Coordinate Descent. \u003cem\u003eJ. Stat. Softw.\u003c/em\u003e \u003cb\u003e33\u003c/b\u003e, 1\u0026ndash;22 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuan, S. et al. Identifying potential targets for preventing cancer progression through the PLA2G1B recombinant protein using bioinformatics and machine learning methods. \u003cem\u003eInt. J. Biol. Macromol.\u003c/em\u003e \u003cb\u003e276\u003c/b\u003e, 133918 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMahmoudian, M., Ven\u0026auml;l\u0026auml;inen, M. S., Kl\u0026eacute;n, R. \u0026amp; Elo, L. L. \u003cem\u003eStable Iterative Variable Selection Bioinformatics\u003c/em\u003e \u003cb\u003e37\u003c/b\u003e, 4810\u0026ndash;4817 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRobin, X. et al. pROC: an open-source package for R and S\u0026thinsp;+\u0026thinsp;to analyze and compare ROC curves. \u003cem\u003eBMC Bioinform.\u003c/em\u003e \u003cb\u003e12\u003c/b\u003e, 77 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eN\u0026uacute;\u0026ntilde;ez, E., Steyerberg, E. W. \u0026amp; N\u0026uacute;\u0026ntilde;ez, J. [Regression modeling strategies]. \u003cem\u003eRev. Esp. Cardiol.\u003c/em\u003e \u003cb\u003e64\u003c/b\u003e, 501\u0026ndash;507 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCui, L. et al. Development of a nomogram for predicting 90-day mortality in patients with sepsis-associated liver injury. \u003cem\u003eSci. Rep.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 3662 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVickers, A. J. \u0026amp; Elkin, E. B. Decision curve analysis: a novel method for evaluating prediction models. \u003cem\u003eMed. Decis. Mak.\u003c/em\u003e \u003cb\u003e26\u003c/b\u003e, 565\u0026ndash;574 (2006).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Obesity, TRPM4, WGCNA, Single-cell sequencing analysis","lastPublishedDoi":"10.21203/rs.3.rs-8754472/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8754472/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWhile TRPM4 is crucial in immune and metabolic diseases, its role in obesity remains unclear. This study aimed to identify TRPM4-associated key genes in obesity. Obesity-related transcriptomic data were analyzed using single-cell RNA sequencing (scRNA-seq) to identify critical cell types, high-Dimensional Weighted Gene Co-expression Network Analysis (hdWGCNA) to determine module genes, and machine learning to screen key genes. A diagnostic nomogram was constructed, potential therapeutics were predicted, and reverse transcription quantitative PCR (RT-qPCR) validated gene expression in clinical samples. Endothelial cells were identified as critical, with ETV6, LAPTM5, and MCOLN3 highlighted as key genes. The nomogram demonstrated strong diagnostic efficacy, and compounds including cyclosporin A, retinoic acid, arsenious acid, and trichostatin A were predicted as potential modulators. RT-qPCR confirmed elevated ETV6 and LAPTM5 and reduced MCOLN3 expression in obesity (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001 and P\u0026thinsp;\u0026lt;\u0026thinsp;0.01, respectively). Integrating bulk and scRNA-seq analyses identified endothelial cells and genes ETV6, LAPTM5, and MCOLN3 as key players, offering potential targets for TRPM4-related obesity interventions.\u003c/p\u003e","manuscriptTitle":"Single-cell WGCNA combined with transcriptome sequencing to study the molecular mechanisms of TRPM4-related genes in obesity with experimental validation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-15 18:38:26","doi":"10.21203/rs.3.rs-8754472/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-03-03T20:17:36+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-24T19:11:52+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"126166116673597033627199465308679093629","date":"2026-02-11T19:02:04+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"141123212442086138703153861276819333671","date":"2026-02-11T18:55:18+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"196419403877616435164185559069953802546","date":"2026-02-09T17:34:41+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-09T16:23:14+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-09T16:18:45+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-02-06T10:47:55+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-05T13:58:02+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2026-02-05T13:31:09+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f21055f3-d8ad-471a-af70-2046a089d158","owner":[],"postedDate":"February 15th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":62609580,"name":"Health sciences/Biomarkers"},{"id":62609581,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":62609582,"name":"Health sciences/Diseases"},{"id":62609583,"name":"Biological sciences/Genetics"},{"id":62609584,"name":"Biological sciences/Molecular biology"}],"tags":[],"updatedAt":"2026-02-15T18:38:26+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-15 18:38:26","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8754472","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8754472","identity":"rs-8754472","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}