Identification of Biomarkers and Mechanisms for Keloid Disorder based on Comprehensive Bioinformatics Analysis and Machine Learning Algorithms

Identification of Biomarkers and Mechanisms for Keloid Disorder based on Comprehensive Bioinformatics Analysis and Machine Learning Algorithms | 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 Research Article Identification of Biomarkers and Mechanisms for Keloid Disorder based on Comprehensive Bioinformatics Analysis and Machine Learning Algorithms Bowen Zheng, Jianxiong Qiao, Xiaoping Yu, Hanghang Zhou, Anqi Wang, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5118256/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 01 Jul, 2025 Read the published version in BMC Medical Genomics → Version 1 posted 23 You are reading this latest preprint version Abstract Background Keloid disorder (KD) is a group of fibroproliferative skin disorders characterized by hypervascularity and excessive accumulation of the extracellular matrix (ECM) and affects individuals of all age groups. The etiology of KD is complex and still poorly understood. This study aimed to investigate biomarkers and therapeutic targets in KD on the basis of comprehensive bioinformatics analysis and machine learning of RNA autosequencing data. Methods Thirteen skin tissues from KD patients (KD samples) and 14 normal control skin tissues (control samples) were collected for RNA sequencing. Initially, differentially expressed key module genes were acquired through expression analysis with weighted gene coexpression network analysis, followed by enrichment analysis. The 10 candidate genes obtained via the CytoHubba plugin were subsequently incorporated into the least absolute shrinkage and selection operator (LASSO) and support vector machine recursive feature elimination (SVM-RFE) to recognize feature genes associated with KD. Furthermore, biomarkers were determined via expression level analysis, followed by enrichment analysis and immunoinfiltration analysis to elucidate the pathogenesis of KD. Results A total of 420 differentially expressed key module genes were identified, and these 420 genes were enriched in collagen- and bone-associated biological functions, including “collagen fibril organization” and “bone development”. With respect to the 10 candidate genes, five feature genes were subsequently obtained through LASSO and SVM-RFE, and among them, NID2, MFAP2, COL8A1, and P4HA3 had significant expression differences between the KD and control samples as well as consistent expression patterns in both datasets; these genes were considered biomarkers. These four biomarkers had excellent abilities to diagnose KD patients, and there were significant positive correlations between these four biomarkers. Functional enrichment analysis suggested that the main enriched KEGG pathways for biomarkers were “steroid hormone biosynthesis”, “cytokine–cytokine receptor interaction”, etc. Furthermore, immune analysis suggested that four biomarkers were negatively linked to type 17 T helper cells and positively linked to 15 immune cells (activated B cells, central memory CD4 T cells, etc.). Conclusion NID2, MFAP2, COL8A1, and P4HA3 were identified as biomarkers for KD, providing more targeted and effective diagnostic and therapeutic strategies for KD. Keloid disorder Biomarker Molecular mechanisms Bioinformatics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 1. Introduction Keloid disorder (KD) is a specific type of abnormal scar formation characterized by raised, firm, and reddish growth that extends beyond the original wound site. The current understanding of the pathogenesis of KD is excessive deposition of collagen during the healing process, which results in a fibroproliferative inflammatory reaction with neovascularization at the advancing edge. KD can be itchy, painful, and cause skin tightness, impacting a person's appearance and quality of life[ 1 ]. During wound healing, fibroblasts and myofibroblasts play key roles in remodeling and forming the collagen matrix. They create a stiffer extracellular matrix[ 2 ]. The KD matrix is composed of various types of glycoproteins, collagens, and glycosaminoglycans[ 3 ]. Initially, the KD matrix is characterized by an overproduction of type III collagen, which over time is gradually replaced by a significant proportion of type I collagen. As a result, the fibers within KD are typically larger in size than those found in normal skin[ 4 ]. Owing to their distinctive morphology and clinically aggressive nature, KD resulting from dermal trauma is often classified as a type of nonspecific etiologic or benign fibroproliferative reticular skin tumor. The morphology and clinical features of KD are similar to those of neoplastic skin tumors, which exhibit aggressive growth, tissue invasion and excessive vascularization[ 5 ]. KD is a common problem worldwide, but certain populations are more predisposed to developing KD. Epidemiological studies have shown that KD is more prevalent in individuals with a darker skin phototype, including those of African, Hispanic, and Asian descent. Additionally, KD tends to affect younger individuals and is more common in females than in males. The exact prevalence of KD varies across different regions and populations, with some studies estimating that up to 10% of people may develop KD at some point in their lives. Genetic factors, environmental influences, and variations in wound healing processes all play a role in the development of KD. Current treatment methods for KD include corticosteroid injections, silicone gel sheets, cryotherapy, laser therapy, surgical excision, and radiation therapy[ 6 ]. However, these treatments may have limitations, such as potential recurrence, discomfort during treatment, and variable effectiveness depending on the individual's response to therapy[ 7 ]. Several general pathologic etiologies have been described for KD development, although no specific factor has been identified as the definitive cause. The immune microenvironment heavily influences KD formation, which represents another key factor in this process. This condition arises from an excessive inflammatory reaction following tissue injury. The presence of an increased number of inflammatory cells, especially T cells and macrophages, within KD tissues is a key feature of this pathological state[ 8 ]. These immune cells play a significant role by releasing a wide range of cytokines and chemokines, such as interleukins (ILs), transforming growth factor β, and fibronectin extra domain A (Fn-EDA), which are essential mediators involved in collagen deposition during KD formation[ 9 ]. However, the exact pathogenesis and molecular mechanisms underlying KD development are still not fully understood. Further research is needed to elucidate the precise mechanisms driving the occurrence and progression of KD, which will be essential in the development of more effective and targeted treatment strategies tailored to individuals affected by this condition. In this study, KD and normal control skin samples were first collected for RNA sequencing to identify differentially expressed genes (DEGs). These genes subsequently intersected with the most relevant key modular genes for KD. In addition, biomarkers were obtained via machine learning combined with expression validation, and receiver operating characteristic (ROC) curves were generated to assess the predictive performance of biomarkers for KD. For immune infiltration analysis, the infiltration levels of 28 immune cells and their correlations with biomarkers were analyzed. We constructed ceRNA regulatory networks to explore the regulatory mechanisms of biomarkers in KD. Finally, subcellular localization of biomarkers was performed, and possible drugs for the treatment of KD were investigated. 2. Materials and methods 2.1 Sample collection and data sources A total of 13 skin tissues from KD patients (KD samples) and 14 normal control skin tissues (control samples) were collected for RNA sequencing. The samples used were obtained from Lanzhou University Second Hospital Hospital, and the Animal Welfare and Ethics Committee of Lanzhou University Second Hospital (D2021–256) approved the study (2023A–324). Furthermore, a KD-related dataset (GSE158395) (GPL24676) was obtained from the Gene Expression Omnibus (GEO) database ( http://www.ncbi.nlm.nih.gov/geo/ ); four KD samples (Keloid_LS) and six control samples were selected. 2.2 RNA sequencing (RNA-seq) Following the manufacturer’s instructions, TRIzol (Invitrogen, CA, USA) was used to isolate and purify RNA from total samples. The quantity and quality of total RNA were measured with a NanoDrop ND-1000 (NanoDrop, Wilmington, DE, USA). Next, a Bioanalyzer 2100 (Agilent, CA, USA) was used to evaluate RNA integrity, followed by confirmation via gel electrophoresis. RIN values > 7.0, OD260/280 > 1.8, concentrations > 50 ng/L, and total RNA > 1 g satisfied the requirements for downstream investigations. The mRNA library was then constructed in accordance with the instructions. After the mRNA containing polyA (polyadenylate) was precisely isolated via oligo (dT) magnetic beads and broken up at high temperatures, reverse transcriptase produced the first strand of cDNA. The production of second-strand cDNA was subsequently mediated by RNase H digestion. After the cDNA ends were repaired, the A-tail was introduced, and libraries with fragment sizes of 300 bp ± 50 bp were generated through screening and purification via magnetic beads. Additionally, we implemented 2×150 bp paired-end sequencing (PE150) via an Illumina NovaSeqTM 6000 in accordance with the vendor’s recommended procedure. Afterwards, FastQC (v 0.11.9) was deployed to assess data quality, and Trimmomatic (v 0.39) was implemented to filter low-quality data[ 10 ]. Reads were mapped to the Homo sapiens GRCh38 genome via Bowtie2 and TopHat2. Immediately thereafter, the clean data were compared to the reference genome (hg19) via hisat2 (v 2.2.1)[ 11 ], with all the parameters set to default values. Finally, the gene expression matrix was acquired for further examination. 2.3 Weighted gene coexpression network analysis (WGCNA) WGCNA was completed via the ‘WGCNA’ package (v 1.69)[ 12 ] utilizing an expression matrix from RNA sequencing data, with the aim of identifying KD-related modules. Initially, unqualified and abnormal samples were identified and excluded through the ‘goodSamplesGenes’ and ‘hclust’ functions to ensure the accuracy of the subsequent analyses. Second, selecting the soft threshold was key to constructing a network topology analysis and was based on the nearly scale-free topology criterion. After the optimal soft threshold was determined, a systematic clustering tree was generated by calculating the neighborhood and similarity between genes, followed by building the expression network in accordance with the criteria of the hybrid dynamic tree cutting algorithm (minModuleSize = 50 and mergeCutHeight = 0.25). Moreover, correlation analysis was performed to assess the relationships between modules and traits (KD and control), and immediately after, we focused on the module most relevant to KD and the module genes it contained ( P 0.3). These genes were defined as key module genes highly correlated with KD. 2.4 Differential expression analysis and enrichment analysis For the RNA sequencing data, the differentially expressed genes (DEGs) between the KD and control samples were mined and analyzed via the ‘DESeq2’ package (v 1.36.0)[ 13 ], with the filtering conditions of |log 2 FoldChange(FC)|>1.5 and adj. P < 0.05. Moreover, volcano maps and heatmaps of the DEGs were generated via the ‘ggplot2’ package (v3.4.1)[ 14 ] and ‘ComplexHeatmap’ package (v1.1.9)[ 15 ], respectively. A Venn diagram of key module genes and DEGs was subsequently generated, and the intersecting genes were considered differentially expressed key modular genes for subsequent analysis. Furthermore, to mine the potential biological functions and signaling pathways linked to key differentially expressed genes, Gene Ontology (GO) annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were performed via the ‘clusterProfiler’ package (v 4.7.1)[ 16 ], and the cutoff value was P < 0.05. The GO annotations included biological process (BP), cell component (CC) and molecular function (MF) terms. 2.5 Protein‒protein interaction (PPI) analysis and machine learning algorithms To explore whether there were interactions among the key differentially expressed genes, these genes were introduced into the online Search Tool for the Retrieval of Interacting Genes (STRING, https://STRING-db.org/ ) to construct a protein–protein interaction (PPI) network (confidence = 0.4). The network was subsequently analyzed via the CytoHubba plugin, followed by the selection of the top 10 genes as candidate genes on the basis of the degree-modified node centrality (DMNC) value. A larger DMNC value indicates that the gene is more important and vice versa. Two machine learning methods, least absolute shrinkage and selection operator (LASSO) and support vector machine recursive feature elimination (SVM-RFE), were subsequently utilized to acquire feature genes in KDs on the basis of the candidate genes obtained above. For the RNA sequencing data, LASSO logistic regression analysis was implemented via the ‘glmnet’ package (v 4.0–2)[ 17 ], and the lambda.min value was applied to select the feature genes. The ‘caret’ package (v 6.0–93) was utilized to achieve SVM-RFE, which ranks various features according to their predictive potential. In the end, the feature genes identified by the above two machine learning algorithms overlapped, and the intersecting feature genes obtained were applied for follow-up analysis. 2.6 Expression level analysis and diagnostic performance evaluation To further clarify the expression levels of the intersecting feature genes in the KD and control samples, the expression data of these genes in the RNA sequencing data and the GSE158395 dataset were compared via the Wilcoxon test, with a significance threshold set at P < 0.05. Importantly, genes whose expression was significantly different between the KD and control samples, as well as genes whose expression trend was consistent between the two datasets, were selected as biomarkers for the diagnosis of KD. To further validate the expression levels of these biomarkers, Western blot analysis was also performed. Briefly, total proteins were extracted from samples collected from the KD and control groups, and the protein concentrations were determined via the BCA method. Equal amounts of protein samples were separated by SDS‒PAGE and transferred to PVDF membranes, followed by incubation with specific primary antibodies. Chemiluminescence detection was used to measure the relative expression levels of the target proteins, and semiquantitative analysis was conducted via ImageJ software. The Western blot results further confirmed the differential expression of the biomarkers between the KD and control groups. Moreover, the ‘pROC’ package (v 1.18.0)[ 18 ] was employed to create receiver operating characteristic (ROC) curves in the RNA sequencing data with the aim of assessing the biomarkers’ predictive accuracy for KD. Depending on the area under the curve (AUC), the prediction precision was classified as low (0.5–0.7), medium (0.7–0.9), or high (> 0.9). Finally, correlations between biomarkers were revealed via a Spearman test via the ‘corrplot’ package (v 0.92)[ 19 ]. 2.7 Gene set enrichment analysis (GSEA) To further uncover the underlying biological mechanisms associated with the biomarkers, the samples in the RNA sequencing data were categorized into high- and low-expression groups according to the median value of expression for each biomarker, after which the DEGs between the two expression groups were determined via the ‘DEseq2’ package (v 1.36.0)[ 20 ] and sorted according to logFC. On the basis of KEGG gene sets from the Molecular Signatures Database (MSigDB, https://www.gsea-msigdb.org/gsea/msigdb/ ), GSEA of the sorted genes was subsequently executed via the ‘clusterProfiler’ package (v 4.7.1)[ 21 ]. The threshold values were set at |NES|≥1 and adj. P < 0.05. 2.8 Immune infiltration analysis Immune cells are known to be involved in the pathogenesis of KD[ 22 ]. Therefore, this work applied the single-sample gene set enrichment analysis (ssGSEA) method to compute enrichment scores of 28 immune infiltrating cells from samples of RNA sequencing data via the ‘GSVA’ package (v 1.40.1)[ 23 ], followed by the Wilcoxon test to compare differential immune cells between KD and control samples ( P < 0.05). To further illuminate the role of immune infiltrating cells and biomarkers in the immune microenvironment, correlation analysis between immune infiltrating cells as well as between differential immune cells and biomarkers was implemented via the Spearman test. 2.9 Regulatory mechanism analysis Transcription factor (TF)-mRNA and lncRNA‒miRNA-mRNA regulatory networks were constructed to investigate the molecular regulatory mechanisms of biomarkers in KD. First, potential TFs regulating the biomarkers were retrieved by accessing the NetworkAnalyst database ( https://www.networkanalyst.ca/ ), and the TF‒mRNA network was generated via Cytoscape software (v 3.9.1)[ 24 ]. The target miRNAs of the biomarkers were subsequently predicted via the miRanda ( http://www.microrna.org/ ) and miRDB ( https://mirdb.org/ ) databases, and the intersection of the two databases was used, after which the StarBase database was accessed to obtain the upstream lncRNAs of the miRNAs (clipExpNum > 4). Ultimately, the lncRNA‒miRNA‒mRNA network was visualized via the ‘ggalluvial’ package (v 0.12.5) ( http://corybrunson.github.io/ggalluvial/ ). 2.10 Subcellular localization analysis and drug prediction Biomarkers were imported into the GeneCards database ( http://www.genecards.org/ ) to extract subcellular distribution information, and the subcellular-biomarker network was visualized via Cytoscape software (v 3.9.1)[ 25 ] (confidence ≥ 2). Finally, the comparative toxicogenomics database (CTD, http://ctdbase.org/ ) was adopted to predict drugs associated with biomarkers to explore potential drugs for the treatment of KD. 2.11 Statistical analysis All bioinformatic projects were conducted in R software. Differences were considered significant at P < 0.05. 3. Results 3.1 In total, 809 key module genes related to KD were obtained through WGCNA With respect to WGCNA, the sample clustering results revealed a lack of outliers in the samples from the RNA sequencing data (Fig. 1 A). When the soft threshold was equal to 24, R 2 was greater than 0.85 (red line), and the mean connectivity was close to 0, which was consistent with a scale-free network distribution (Fig. 1 B). Furthermore, the dynamic tree cutting algorithm yielded 11 coexpression modules (Fig. 1 C). In terms of the correlations between modules and traits (KD and control), the MEgreen module had the highest correlation with KD (cor = 0.86 and P < 0.01), and 809 genes in the MEgreen module were identified as key module genes highly associated with KD ( Fig. 1 D ) . 3.2 Differentially expressed key module genes associated with collagen- and bone-associated biological functions In total, 1,898 DEGs between the KD and control samples were mined from the RNA sequencing data, with 900 upregulated and 998 downregulated genes ( Fig. 2 A- 2 B ) . The Venn diagram revealed that 420 differentially expressed key module genes were acquired by extracting the intersections of 1,898 DEGs and 809 key module genes (Fig. 2 C). Furthermore, GO annotation and KEGG pathway analyses were performed to elucidate the underlying biological mechanisms of the 420 genes, yielding a total of 192 GO entries (146 BPs, 17 CCs, and 29 MFs) and nine KEGG signaling pathways ( P < 0.05). In the GO-BP category, these 420 genes were enriched primarily in “extracellular matrix organization”, “collagen fibril organization”, “bone development”, “collagen biosynthetic process”, and “cartilage development” (Fig. 2 D and Table S1) . The GO-CC analysis results showed that 420 genes were significantly enriched in “collagen-containing extracellular matrix”, “complex of collagen trimers”, and “fibrillar collagen trimers” (Fig. 2 D and Table S1) . GO-MF terms were enriched in “collagen binding” and “integrin binding” (Fig. 2 D and Table S1) . Moreover, KEGG analysis revealed that 420 genes were enriched in pathways related to “ECM-receptor interaction”, “PI3K-Akt signaling pathway”, and “cell adhesion molecules” (Fig. 2 E and Table S2) . Overall, collagen- and bone-associated biological functions and cell-associated signaling pathways appear to be key factors in the regulation of KD. 3.3 Identification of feature genes through LASSO and SVM-RFE The PPI network of 420 differentially expressed key module genes, including 420 points and 2,111 edges, is shown in Fig. 3 A, followed by the identification of 10 candidate genes (LOXL1, COL16A1, CD248, NID2, P4HA3, LAMB1, P3H3, COL8A1, MFAP2, and LOXL2) on the basis of DMNC values in the CytoHubba plugin (Fig. 3 B). These 10 candidate genes were subsequently incorporated into machine learning algorithms to further identify feature genes in the RNA sequencing data. Through the LASSO algorithm, five genes were identified as feature genes of KD: NID2, MFAP2, COL8A1, P4HA3, and CD248 (Fig. 3 C ) . Moreover, eight feature genes (COL8A1, MFAP2, P4HA3, LOXL2, LOXL1, P3H3, NID2, and CD248) were identified via the SVM-RFE algorithm (Fig. 3 D). Moreover, the feature genes acquired via the above two methods were intersected to acquire five intersecting feature genes (NID2, MFAP2, COL8A1, P4HA3, and CD248) for subsequent analysis ( Fig. 3 E ) . 3.4 NID2, MFAP2, COL8A1, and P4HA3 are biomarkers of KD Transcription and protein expression level analyses were carried out to identify biomarkers. Notably, NID2, MFAP2, COL8A1, and P4HA3 were significantly differentially expressed between the KD and control samples ( P < 0.05) and presented consistent expression patterns (Fig. 4 A- 4 C ) . Therefore, these four genes were utilized as biomarkers for KD in this study, and the expression of all four genes was markedly greater in the KD samples than in the control samples. Moreover, the ROC curves of the biomarkers in the RNA sequencing data suggested that the AUC values (NID2 = 0.967, MFAP2 = 1, COL8A1 = 1, and P4HA3 = 1) were all greater than 0.9 (Fig. 4 D- 4 G ) , implying that they had excellent ability to diagnose KD patients. Correlation analysis revealed significant positive correlations between these four biomarkers (cor > 0.75 and P < 0.05), with the strongest correlation between MFAP2 and P4HA3 (cor = 0.90) (Fig. 4 H ) . 3.5 Biological pathways of biomarkers explored in KD To explore the underlying biological mechanisms of biomarkers, GSEA was performed on RNA sequencing data. The KEGG gene set revealed four biomarkers with enriched pathways, including “citrate cycle TCA cycle”, “metabolism of xenobiotics by cytochrome P450”, “steroid hormone biosynthesis”, “PPAR signaling pathway”, “cytokine‒cytokine receptor interaction”, “chemokine signaling pathway”, “TGF-beta signaling pathway”, “ECM-receptor-interaction”, etc. (Fig. 5 A- 5 D ) , revealing their important roles in the pathogenesis of KD. 3.6 Biomarkers associated with immune infiltrating cells The mechanisms by which the four biomarkers influence KD progression were examined via immune analysis of RNA sequencing data. The infiltration levels of 28 immune cells in the KD and control samples are depicted in Fig. 6A. Among these 28 immune cells, the strongest positive correlation (cor=0.96) was detected between type 1 T helper cells and regulatory T cells, and the strongest negative correlation (cor=-0.78) was detected between effector memory CD4 + T cells and CD56 dim natural killer cells ( Fig. 6B) . Moreover, 16 immune cell types (activated B cells, central memory CD4+ T cells, central memory CD8+ T cells, macrophages, memory B cells, etc.) differed between the KD and control samples ( P <0.05), and all of them were more abundant in the KD samples than in the control samples ( Fig. 6C) , indicating that disease progression might be influenced by these differential immune cells. Furthermore, the correlation of biomarkers with differential immune cells indicated that all four biomarkers were negatively linked to type 17 T helper cells and positively correlated with the remaining 15 immune cells ( Fig. 6D) . 3.7 Investigation of the potential regulatory mechanism of biomarkers To elucidate the regulatory mechanisms of biomarkers, four biomarkers were entered into the NetworkAnalyst database, yielding a total of 20 upstream TFs of biomarkers, including seven, nine, seven, and three TFs targeting NID2, MFAP2, COL8A1, and P4HA3, respectively. In this TF-mRNA network, FOXC1 was a regulator of three biomarkers (NID2, MFAP2, and COL8A1), and TFs (FOS and JUN) could coregulate two biomarkers (NID2 and MFAP2) (Fig. 7 A ) . Furthermore, on the basis of the starBase, miRanda, and miRDB databases, a lncRNA‒miRNA‒mRNA network consisting of two biomarkers (NID2 and MFAP2), five miRNAs and 39 lncRNAs (Fig. 7 B ) was constructed. Multiple lncRNA‒miRNA‒mRNA relationship pairs, such as lncRNAs (SNHG1, NEAT1, MALAT1, etc.) that can regulate MFAP2 via miRNAs (hsa-miR-3163), as well as hsa-miR‒199 family members (hsa-miR‒199b‒3p and hsa-miR‒199a‒3p), were identified as regulators of NID2, were found in this network. 3.8 Subcellular localization and drug prediction The biomarkers were entered into the GeneCards database to analyze their subcellular localization, which indicated that all four genes were expressed in the extracellular space and that COL8A1, MFAP2, and NID2 were expressed in the plasma membrane (Fig. 8 A). By analyzing the CTD database, 86 biomarker-related drugs were obtained, including 40 drugs targeting NID2, 43 drugs targeting MFAP2, 40 drugs targeting COL8A1, and 22 drugs targeting P4HA3 (Fig. 8 B). Notably, seven drugs (bisphenol A, bisphenol F, bis(4-hydroxyphenyl) sulfone, benzo(a)pyrene, acetaminophen, decamethrin, and calcitriol) were coacting drugs with four biomarkers. 4. Discussion KD is characterized by the development of excess scar tissue that expands beyond the initial site of skin injury, often affecting individuals with a genetic predisposition to this condition[ 26 ]. The exact pathophysiological mechanisms underlying KD development are largely unclear[ 27 ] but likely involve skin tension[ 28 ], tissue hypoxia[ 29 ], chronic inflammation[ 30 ], autoimmunity[ 31 ], and genetic factors[ 32 ]. KD is further characterized by increased blood vessel proliferation, epidermal thickening, increased mesenchymal cell density, thick and dense hyalinized collagen fibers, and excessive fibronectin accumulation[ 33 , 34 ]. In contrast, normal skin features collagen bundles aligned parallel to the epidermis. Currently, different approaches are used to improve these scars, including intralesional corticosteroids, surgery and, more recently, laser therapy. They can result in itching and pain and adversely affect the physical and psychological well-being of patients. Therefore, finding KD-related biomarkers and exploring prospective targets for anti-KD therapies are highly important. By enrichment analysis of candidate genes, we observed significant enrichment in collagen binding, extracellular matrix and collagen fibril organization, bone development, the collagen biosynthetic process, the collagen-containing extracellular matrix, the complex of collagen trimers, and the fibrillar collagen trimer. Furthermore, the KEGG analysis identified enriched pathways, including ECM-receptor interaction, the PI3K-Akt signaling pathway, and cell adhesion molecules. These findings indicate that the reconstruction and deposition of collagen, reorganization of the extracellular matrix, and regulation of signaling pathways are crucial aspects of scar formation and collectively contribute to the development of scar tissue[ 35 ]. Further investigation into the interaction and regulatory mechanisms of these factors is essential for uncovering the biological processes underlying scar formation. In this study, four biomarkers (NID2, MFAP2, COL8A1, and P4HA3) were identified for KD by innovatively combining RNA sequencing, WGCNA, SVM-RFE, and LASSO logistic regression methods. Subsequently, the diagnostic accuracy of the ROC curve was validated. Finally, the consistent expression of the biomarkers in the dataset and clinical samples reinforced the validity of the biomarkers. Nidogen 2 (NID2) is a gene involved in the regulation of cell‒matrix interactions and is known to play a role in maintaining tissue structure and integrity. It can be a new biomarker in some cancer[ 36 , 37 ]. In KD, dysregulation of NID2 expression may lead to abnormal interactions between cells and the extracellular matrix, contributing to the excessive deposition of collagen and extracellular matrix components characteristic of KD. Additionally, NID2 may impact signaling pathways involved in fibroblast proliferation and migration, further promoting the pathological processes observed in KD[ 38 ]. Microfibril-associated protein 2(MFAP2) is a gene encoding a protein that is associated with microfibrils in the extracellular matrix. In the context of KD, MFAP2 may influence the organization and stability of the extracellular matrix, potentially contributing to abnormal scarring and fibrotic tissue formation in KD. Dysregulation of MFAP2 expression may disrupt the structural integrity of the extracellular matrix, leading to the aberrant tissue remodeling and excessive collagen deposition observed in KD lesions[ 39 ]. The collagen type VIII alpha 1 chain (COL8A1) gene encodes one of the two alpha chains of type VIII collagen. The gene product is short-chain collagen and a macromolecular component of the subendothelium. The major component of Descemet's membrane (basement membrane) of corneal endothelial cells[ 40 ]. Additionally, it is a component of the endothelia of blood vessels. This protein is necessary for the migration and proliferation of vascular smooth muscle cells and thus has a potential role in the maintenance of vessel wall integrity and structure, particularly in atherogenesis[ 41 ]. In KD, dysfunction of COL8A1 may lead to aberrant collagen synthesis and deposition, contributing to the excessive fibrosis and scar formation observed in KD lesions. This dysregulation of collagen production and organization may disrupt normal tissue repair processes, leading to the pathological scarring characteristic of KD[ 42 ]. Prolyl 4-hydroxylase subunit alpha 3 (P4HA3) is a gene encoding a subunit of prolyl 4-hydroxylase, an enzyme involved in collagen synthesis and maturation [ 43 , 44 ]. Altered expression of P4HA3 in KD may lead to abnormal collagen cross-linking and deposition, resulting in the formation of dense, fibrotic scar tissue distinctive of KD [ 45 , 46 ]. Consequently, P4HA3 likely plays a pivotal role in modulating collagen stability and organization within the extracellular matrix. Abnormalities in P4HA3 could influence collagen remodeling processes, thus contributing to the pathogenesis of KD. In this study, the four biomarkers also showed excellent performance in diagnosing KD, suggesting that biomarkers can be used to diagnose KD more accurately and can be used to achieve targeted therapy for KD by modulating the corresponding pathways. In addition, it is important to highlight that inflammatory chemicals are key mediators of immune cell activities. Immune cells invade and regulate the KD microenvironment through the secretion of cytokines[ 47 ]. These cytokines, including IL-4, IL-6, IL-10, IL-12, IL-13, IL-17, TNF-α, and TGF-β, are significantly elevated in the local microenvironment of KD[ 48 ]. They play crucial roles in both the inflammatory phase and the prolonged proliferative phase of KD[ 49 ]. In contrast, the serum concentrations of IFN-α, IFN-γ, and TNF-β, which are known to suppress collagen synthesis and fibroblast proliferation, are reduced, leading to a lack of inhibition and promoting uncontrolled collagen production[ 50 ]. Immune infiltration plays a critical role in KD development. Previous studies have highlighted the significant involvement of various inflammatory cells in the wound-healing process[ 51 ]. For example, the balance between the M1 and M2 phenotypes is crucial for regulating inflammation and tissue healing, both of which are controlled by macrophages. Macrophages in scar tissue are highly activated and polarized toward the M2 subtype [ 52 , 53 ]. M2 macrophages can produce cytokines that stimulate fibroblast proliferation, such as vascular endothelial growth factor, transforming growth factor beta , and insulin-like growth factor (IGF)-1. T cells also play a critical role. The proportions of CD8 + memory T cells and effector memory CD8 + T cells are greater in KD tissue than in normal skin tissue, as commonly observed in inflammatory skin diseases[ 54 ]. Naïve CD4 + T cells differentiate into distinct Th lineages, including Th1, Th2, Th17, and Tregs[ 55 ]. In a study comparing lesional versus normal skin, significant upregulation of markers of T-cell activation/migration (ICOS, CCR7) was observed. Additionally, elevated expression of markers associated with the Th2 pathway (IL4R, CCL11, TNFSF4/OX40L), the Th1 pathway (CXCL9/CXCL10/CXCL11), and the Th17/Th22 pathways (CCL20, S100As) was identified, along with increased JAK/STAT signaling (JAK3). Tregs act by inhibiting CD4 + and CD8 + T-cell activation and promoting M2 macrophage polarization[ 56 ]. Treg cells express TGF-β, which enhances COL1A1 and COL3A1 expression, leading to collagen overexpression[ 57 ]. Recent research has shed light on the role of dendritic cells (DCs) in influencing fibroblasts and vasculature function[ 58 ]. Wu et al. reported an elevated presence of DCs (CD11c+) in KD, along with the dendritic cell infiltrates commonly observed in atopic dermatitis[ 59 ]. Rath et al. reported that inflammatory DCs are the predominant type of DC infiltrating KD tissue, with a notably stronger correlation between inflammatory DCs and fibroblasts in KD tissue than in surrounding or healthy tissue [ 60 ]. These findings suggest that during wound healing, a KD can recruit and activate inflammatory DC precursors, leading to inflammation[ 60 ]. Furthermore, the interaction between inflammatory DCs and fibroblasts may stimulate the expression of metalloproteinases in fibroblasts, thereby promoting fibroblast activation and fibrosis[ 61 ]. In this study, we examined the infiltration level of 28 immune cells in the samples, as immune infiltration significantly affects scar formation. These findings reveal substantial variations in the levels of activated B cells, central memory CD4 + T cells, central memory CD8 + T cells, macrophages, memory B cells, effector memory CD4 + T cells, effector memory CD8 + T cells, immature B cells, natural killer cells, natural killer T cells, plasmacytoid dendritic cells, regulatory T cells, type 1 T helper cells, monocytes, MDSCs, and type 17 T helper cells between KD and normal tissues. On the basis of our findings and those of the aforementioned studies, macrophages, plasmacytoid dendritic cells, memory T cells, and T helper cells perform critical functions in KD tissue. This study also demonstrated the relationships between potential biomarkers and immune cells. Specifically, analysis revealed that the biomarkers NID2, MFAP2, COL8A1, and P4HA3 were negatively associated with type 17 T helper cells but positively correlated with the other 15 types of immune cells. In the discussion above, how different immune cells contribute to the development of KD is outlined, and the findings suggest that immunotherapy could address the challenges in treating KD. This study also retrieved small molecule drugs from the CDT database and revealed that seven top-ranked drugs, bisphenol-A, bisphenol-F, bis(4-hydroxyphenyl) sulfone, benzo(a)pyrene, acetaminophen, decamethrin, and calcitriol, were identified as coacting drugs with four specific biomarkers. These drugs may have therapeutic potential for KD. These pharmacological treatments for KD reported in the literature are still rare. Further experimental studies are needed to elucidate the therapeutic potential of these drugs in the context of KD. Certain limitations of this study warrant recognition, as it was a single-center design. Thus, our results should be validated prospectively in larger cohorts from multiple centers. Furthermore, a further subdivided immune cell subset is needed to study the mechanism behind KD development in those subdivided immune cell subsets. Finally, additional experimental validation, such as cellular and animal experiments, is needed to substantiate these findings and enhance the reliability of the results. The results of our study indicate that NID2, MFAP2, COL8A1, and P4HA3 could influence the KD microenvironment, potentially promoting growth. Future experiments could involve the development of animal models of KD to evaluate changes in scar size and histology after manipulating the expression of potential biomarkers, thus validating their roles in KD growth. Moreover, primary cells such as fibroblasts and keratinocytes can be cultured in cellular experiments to study their impact on activities such as cell proliferation, migration, and apoptosis. Gene therapy is a revolutionary new technique; however, its clinical application in KD remains relatively rare. Confirming the significance and effectiveness of these 4 promising biomarkers in KD through multiple methods can facilitate their clinical application. This approach provides a viable method to treat KD. In summary, our comprehensive analysis identified four key differentially expressed genes (NID2, MFAP2, COL8A1, and P4HA3) as potential biomarkers for keloid disease (KD). These biomarkers were found to be significantly associated with the dysregulation of collagen synthesis, extracellular matrix remodeling, and signaling pathways critical for the pathogenesis of KD. Furthermore, we observed distinct patterns of immune cell infiltration, with notable differences in the levels of activated B cells, T-cell subsets, macrophages, and dendritic cells between KD and normal skin samples. The identified biomarkers exhibited strong correlations with these immune cell populations, suggesting their potential involvement in the complex interplay between the immune system and the aberrant wound healing processes underlying KD development. These findings provide valuable insights into the molecular and cellular mechanisms driving KD development, and the identified biomarkers hold promise as diagnostic and therapeutic targets. Future research should focus on the functional validation of these biomarkers, both in vitro and in vivo, to further elucidate their specific roles in KD pathogenesis. Additionally, exploring the therapeutic potential of small-molecule drugs that may modulate the identified biomarkers and associated pathways could pave the way for the development of more effective treatment strategies for this debilitating condition. Collectively, our results lay the groundwork for advancing the understanding and management of keloid disease, with the ultimate goal of improving patient outcomes. Abbreviations KD Keloid Disorder EC Extracellular Matrix LASSO Least Absolute Shrinkage and Selection Operator SVM-RFE Support Vector Machine Recursive Feature Elimination Fn-EDA Fibronectin Extra Domain A DEGs Differentially Expressed Genes ROC Receiver Operating Characteristic WGCNA Weighted Gene Coexpression Network Analysis GO Gene Ontology KEGG Kyoto Encyclopedia of Genes and Genomes BP Biological process CC Cell Component MF Molecular Function Terms PPI Protein Protein Interaction DMNC Degree-Modified Node Centrality Declarations Acknowledgements We thank the patient’s and family for permission to publish these datas. Thanks to the reviewers and editors for their sincere comments. Authors’ contributions Bowen Zheng wrote the first draft and coordinated the manuscript. Anqi Wang, Hanghang Zhou, Xiaoping Yu and Jianxiong Qiaoprepared genetic data collection and analysis. Xuanfen Zhang contributed to data visualisation, manuscript editing, finalisation. All authors reviewed the manuscript. Funding This study was supported by the Natural Science Foundation of Gansu Province under Grant [number 24JRRA339]. Data availability The datasets and materials generated during and/or analysed during the current study are available as follow: https://pan.baidu.com/s/1YirFLmRAZIrkEXC2Nd08GA code: 6666. Competing interests The authors declare that they have no competing interests. Ethics approval and consent to participate The studies involving human participants were reviewed and approved by the Ethic Committee of Lanzhou University Second Hospital. The patients/participants provided their written informed consent to participate in this study. Consent for publication Written informed consents for publication of samples or other personal or clinical details were obtained from all participants and patients. References Feng F, Liu M, Pan L, Wu J, Wang C, Yang L, et al. Biomechanical Regulatory Factors and Therapeutic Targets in Keloid Fibrosis. Front Pharmacol. 2022;13:906212. Ashcroft KJ, Syed F, Bayat A. Site-specific keloid fibroblasts alter the behaviour of normal skin and normal scar fibroblasts through paracrine signalling. PLoS ONE. 2013;8:e75600. Satish L, Lyons-Weiler J, Hebda PA, Wells A. Gene expression patterns in isolated keloid fibroblasts. Wound Repair Regen. 2006;14:463–70. Verhaegen PDHM, van Zuijlen PPM, Pennings NM, van Marle J, Niessen FB, van der Horst CMAM, et al. Differences in collagen architecture between keloid, hypertrophic scar, normotrophic scar, and normal skin: An objective histopathological analysis. Wound Repair Regen. 2009;17:649–56. Tan S, Khumalo N, Bayat A. Understanding Keloid Pathobiology From a Quasi-Neoplastic Perspective: Less of a Scar and More of a Chronic Inflammatory Disease With Cancer-Like Tendencies. Front Immunol. 2019;10:1810. Gauglitz GG. Management of keloids and hypertrophic scars: current and emerging options. Clin Cosmet Investig Dermatol. 2013;6:103–14. Gauglitz GG, Pötschke J, Clementoni MT. [Therapy of scars with lasers]. Hautarzt. 2018;69:17–26. Shaker SA, Ayuob NN, Hajrah NH. Cell talk: a phenomenon observed in the keloid scar by immunohistochemical study. Appl Immunohistochem Mol Morphol. 2011;19:153–9. Nangole FW, Agak GW. Keloid pathophysiology: fibroblast or inflammatory disorders? JPRAS Open. 2019;22:44–54. de Sena Brandine G, Smith AD. Falco: high-speed FastQC emulation for quality control of sequencing data. F1000Res. 2019;8:1874. Kim D, Paggi JM, Park C, Bennett C, Salzberg SL. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat Biotechnol. 2019;37:907–15. Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008;9:559. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15:550. Gustavsson EK, Zhang D, Reynolds RH, Garcia-Ruiz S, Ryten M. ggtranscript: an R package for the visualization and interpretation of transcript isoforms using ggplot2. Bioinformatics. 2022;38:3844–6. Gu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics. 2016;32:2847–9. Yu G, Wang L-G, Han Y, He Q-Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS. 2012;16:284–7. Friedman J, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models via Coordinate Descent. J Stat Softw. 2010;33:1–22. Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez J-C, et al. pROC: an open-source package for R and S + to analyze and compare ROC curves. BMC Bioinformatics. 2011;12:77. Zhang X, Chao P, Zhang L, Xu L, Cui X, Wang S, et al. Single-cell RNA and transcriptome sequencing profiles identify immune-associated key genes in the development of diabetic kidney disease. Front Immunol. 2023;14:1030198. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15:550. Yu G, Wang L-G, Han Y, He Q-Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS. 2012;16:284–7. Shan M, Liu H, Song K, Liu S, Hao Y, Wang Y. Immune-related gene expression in skin, inflamed and keloid tissue from patients with keloids. Oncol Lett. 2022;23:72. Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics. 2013;14:7. Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics. 2013;14:7. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13:2498–504. Troiano M, Simeone A, Scaramuzzi G, Parisi S, Guglielmi G. Giant keloid of left buttock treated with post-excisional radiotherapy. J Radiol Case Rep. 2011;5:8–15. Troiano M, Simeone A, Scaramuzzi G, Parisi S, Guglielmi G. Giant keloid of left buttock treated with post-excisional radiotherapy. J Radiol Case Rep. 2011;5:8–15. Song H, Liu T, Wang W, Pang H, Zhou Z, Lv Y, et al. Tension enhances cell proliferation and collagen synthesis by upregulating expressions of integrin αvβ3 in human keloid-derived mesenchymal stem cells. Life Sci. 2019;219:272–82. Wolfram D, Tzankov A, Pülzl P, Piza-Katzer H. Hypertrophic scars and keloids–a review of their pathophysiology, risk factors, and therapeutic management. Dermatol Surg. 2009;35:171–81. Ogawa R. Keloid and Hypertrophic Scars Are the Result of Chronic Inflammation in the Reticular Dermis. Int J Mol Sci. 2017;18. Kazeem AA. The immunological aspects of keloid tumor formation. J Surg Oncol. 1988;38:16–8. Bayat A, Arscott G, Ollier WER, McGrouther DA, Ferguson MWJ. Keloid disease: clinical relevance of single versus multiple site scars. Br J Plast Surg. 2005;58:28–37. Lee JY-Y, Yang C-C, Chao S-C, Wong T-W. Histopathological differential diagnosis of keloid and hypertrophic scar. Am J Dermatopathol. 2004;26:379–84. Babu M, Diegelmann R, Oliver N. Fibronectin is overproduced by keloid fibroblasts during abnormal wound healing. Mol Cell Biol. 1989;9:1642–50. Mathew-Steiner SS, Roy S, Sen CK. Collagen in Wound Healing. Bioeng (Basel). 2021;8. Kiziltan R, Cikman O, Algul S, Aydin MA, Kemik O. Nidogen-2: A new biomarker in colon cancer patients. Ann Ital Chir. 2022;92:88–92. Mao C, Ma Z, Jia Y, Li W, Xie N, Zhao G, et al. Nidogen-2 Maintains the Contractile Phenotype of Vascular Smooth Muscle Cells and Prevents Neointima Formation via Bridging Jagged1-Notch3 Signaling. Circulation. 2021;144:1244–61. He Y, Tsou P-S, Khanna D, Sawalha AH. Methyl-CpG-binding protein 2 mediates antifibrotic effects in scleroderma fibroblasts. Ann Rheum Dis. 2018;77:1208–18. Yao L-W, Wu L-L, Zhang L-H, Zhou W, Wu L, He K, et al. MFAP2 is overexpressed in gastric cancer and promotes motility via the MFAP2/integrin α5β1/FAK/ERK pathway. Oncogenesis. 2020;9:17. Aldave AJ, Rayner SA, Salem AK, Yoo GL, Kim BT, Saeedian M, et al. No pathogenic mutations identified in the COL8A1 and COL8A2 genes in familial Fuchs corneal dystrophy. Invest Ophthalmol Vis Sci. 2006;47:3787–90. Xu R, Yao ZY, Xin L, Zhang Q, Li TP, Gan RB. NC1 domain of human type VIII collagen (alpha 1) inhibits bovine aortic endothelial cell proliferation and causes cell apoptosis. Biochem Biophys Res Commun. 2001;289:264–8. Hopfer U, Fukai N, Hopfer H, Wolf G, Joyce N, Li E, et al. Targeted disruption of Col8a1 and Col8a2 genes in mice leads to anterior segment abnormalities in the eye. FASEB J. 2005;19:1232–44. Kukkola L, Hieta R, Kivirikko KI, Myllyharju J. Identification and characterization of a third human, rat, and mouse collagen prolyl 4-hydroxylase isoenzyme. J Biol Chem. 2003;278:47685–93. Van Den Diepstraten C, Papay K, Bolender Z, Brown A, Pickering JG. Cloning of a novel prolyl 4-hydroxylase subunit expressed in the fibrous cap of human atherosclerotic plaque. Circulation. 2003;108:508–11. Shih B, Garside E, McGrouther DA, Bayat A. Molecular dissection of abnormal wound healing processes resulting in keloid disease. Wound Repair Regen. 2010;18:139–53. Kihara M, Amano H, Misu Y, Kubo T. Release of [3H]- and endogenous GABA from slices of the rat medulla oblongata: modification by 3-mercaptopropionic acid, nipecotic acid and diaminobutyric acid. Arch Int Pharmacodyn Ther. 1989;298:50–60. Shan M, Wang Y. Viewing keloids within the immune microenvironment. Am J Transl Res. 2022;14:718–27. Dong X, Mao S, Wen H. Upregulation of proinflammatory genes in skin lesions may be the cause of keloid formation (Review). Biomed Rep. 2013;1:833–6. Shih B, Garside E, McGrouther DA, Bayat A. Molecular dissection of abnormal wound healing processes resulting in keloid disease. Wound Repair Regen. 2010;18:139–53. Shaffer JJ, Taylor SC, Cook-Bolden F. Keloidal scars: a review with a critical look at therapeutic options. J Am Acad Dermatol. 2002;46(2 Suppl Understanding):S63–97. Jiao H, Fan J, Cai J, Pan B, Yan L, Dong P, et al. Analysis of Characteristics Similar to Autoimmune Disease in Keloid Patients. Aesthetic Plast Surg. 2015;39:818–25. Jin Q, Gui L, Niu F, Yu B, Lauda N, Liu J, et al. Macrophages in keloid are potent at promoting the differentiation and function of regulatory T cells. Exp Cell Res. 2018;362:472–6. Tardito S, Martinelli G, Soldano S, Paolino S, Pacini G, Patane M, et al. Macrophage M1/M2 polarization and rheumatoid arthritis: A systematic review. Autoimmun Rev. 2019;18:102397. Chen Z, Zhou L, Won T, Gao Z, Wu X, Lu L. Characterization of CD45RO(+) memory T lymphocytes in keloid disease. Br J Dermatol. 2018;178:940–50. Zhu J, Yamane H, Paul WE. Differentiation of effector CD4 T cell populations (*). Annu Rev Immunol. 2010;28:445–89. Schmidt A, Oberle N, Krammer PH. Molecular mechanisms of treg-mediated T cell suppression. Front Immunol. 2012;3:51. Chen Y, Jin Q, Fu X, Qiao J, Niu F. Connection between T regulatory cell enrichment and collagen deposition in keloid. Exp Cell Res. 2019;383:111549. Lu TT. Dendritic cells: novel players in fibrosis and scleroderma. Curr Rheumatol Rep. 2012;14:30–8. Wu J, Del Duca E, Espino M, Gontzes A, Cueto I, Zhang N, et al. RNA Sequencing Keloid Transcriptome Associates Keloids With Th2, Th1, Th17/Th22, and JAK3-Skewing. Front Immunol. 2020;11:597741. Rath M, Pitiot A, Kirr M, Fröhlich W, Plosnita B, Schliep S et al. Multi-Antigen Imaging Reveals Inflammatory DC, ADAM17 and Neprilysin as Effectors in Keloid Formation. Int J Mol Sci. 2021;22. Lagares D, Ghassemi-Kakroodi P, Tremblay C, Santos A, Probst CK, Franklin A, et al. ADAM10-mediated ephrin-B2 shedding promotes myofibroblast activation and organ fibrosis. Nat Med. 2017;23:1405–15. Additional Declarations No competing interests reported. Supplementary Files TableS1.csv TableS2.csv Cite Share Download PDF Status: Published Journal Publication published 01 Jul, 2025 Read the published version in BMC Medical Genomics → Version 1 posted Editorial decision: Revision requested 04 Nov, 2024 Reviews received at journal 24 Oct, 2024 Reviews received at journal 12 Oct, 2024 Reviews received at journal 11 Oct, 2024 Reviews received at journal 07 Oct, 2024 Reviews received at journal 06 Oct, 2024 Reviews received at journal 02 Oct, 2024 Reviewers agreed at journal 02 Oct, 2024 Reviewers agreed at journal 29 Sep, 2024 Reviewers agreed at journal 28 Sep, 2024 Reviews received at journal 27 Sep, 2024 Reviewers agreed at journal 27 Sep, 2024 Reviewers agreed at journal 27 Sep, 2024 Reviewers agreed at journal 27 Sep, 2024 Reviewers agreed at journal 27 Sep, 2024 Reviewers agreed at journal 27 Sep, 2024 Reviewers agreed at journal 27 Sep, 2024 Reviewers agreed at journal 25 Sep, 2024 Reviewers invited by journal 25 Sep, 2024 Editor invited by journal 20 Sep, 2024 Editor assigned by journal 20 Sep, 2024 Submission checks completed at journal 20 Sep, 2024 First submitted to journal 19 Sep, 2024 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-5118256","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":373860675,"identity":"f16fb22e-637b-4d8a-9ff5-36d57b426a3c","order_by":0,"name":"Bowen Zheng","email":"","orcid":"","institution":"The Department of Plastic Surgery, Lanzhou University Second Hospital, Lanzhou, 730030, People's Republic of China","correspondingAuthor":false,"prefix":"","firstName":"Bowen","middleName":"","lastName":"Zheng","suffix":""},{"id":373860676,"identity":"e732803c-383e-4237-a116-070b5c2373ea","order_by":1,"name":"Jianxiong Qiao","email":"","orcid":"","institution":"The Department of Plastic Surgery, Lanzhou University Second Hospital, Lanzhou, 730030, People's Republic of China","correspondingAuthor":false,"prefix":"","firstName":"Jianxiong","middleName":"","lastName":"Qiao","suffix":""},{"id":373860677,"identity":"0e28ac5d-13cd-4358-8ef9-d32d3c391c64","order_by":2,"name":"Xiaoping Yu","email":"","orcid":"","institution":"The Department of burn, Gansu Provincial Hospital,Lanzhou,Gansu, 730000, China","correspondingAuthor":false,"prefix":"","firstName":"Xiaoping","middleName":"","lastName":"Yu","suffix":""},{"id":373860678,"identity":"2c136705-496d-4be3-9955-8865c4d4346b","order_by":3,"name":"Hanghang Zhou","email":"","orcid":"","institution":"The Department of Plastic Surgery, Lanzhou University Second Hospital, Lanzhou, 730030, People's Republic of China","correspondingAuthor":false,"prefix":"","firstName":"Hanghang","middleName":"","lastName":"Zhou","suffix":""},{"id":373860679,"identity":"e1315d95-4ad0-4b06-8e4e-9152568beda0","order_by":4,"name":"Anqi Wang","email":"","orcid":"","institution":"NHC Key Laboratory of Diagnosis and Therapy of Gastrointestinal Tumor, Gansu Provincial Hospital, Lanzhou, 730000, China, China","correspondingAuthor":false,"prefix":"","firstName":"Anqi","middleName":"","lastName":"Wang","suffix":""},{"id":373860680,"identity":"af3fdba1-6465-43e9-9aac-6a9e11f6d72c","order_by":5,"name":"Xuanfen Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyElEQVRIiWNgGAWjYDACCQaGDx9+SDDzMzMffECsFsaZM3ts2CXb2ZINiNYym4ctjd/gPI+ZAFE65Gf3GDbz8ByWNj7MYMbAUGMTTVAL45xjiY1zLA4bmx1mSHvAcCwtt4GQFmaJ5OMP3vAcTgZqOW7A2HCYsBY2icTGBh62w/WbmxnbJIjSwiORfLAR6H1mA2ZmNuK0SEikJTYCA5lZ4jAbs0ECMX6Rn5Fj2ACOyv7zHx98qLEhrAUVJJCmfBSMglEwCkYBLgAA2Lg8scXBEqoAAAAASUVORK5CYII=","orcid":"","institution":"The Department of Plastic Surgery, Lanzhou University Second Hospital, Lanzhou, 730030, People's Republic of China","correspondingAuthor":true,"prefix":"","firstName":"Xuanfen","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2024-09-19 15:53:23","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5118256/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5118256/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12920-025-02174-9","type":"published","date":"2025-07-01T15:58:27+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":71475727,"identity":"3cfe5261-4924-43f2-892e-657df85b7e5e","added_by":"auto","created_at":"2024-12-16 05:03:19","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":35602,"visible":true,"origin":"","legend":"\u003cp\u003eIdentification of key module genes highly associated with KD via WGCNA. \u003cstrong\u003eA\u003c/strong\u003e Sample clustering of a total of 27 samples to detect outliers. \u003cstrong\u003eB \u003c/strong\u003eGraphs of scale independence, mean connectivity and scale-free topology; the appropriate soft power was 24. \u003cstrong\u003eC \u003c/strong\u003eCluster dendrogram of the coexpression network modules (1-TOM); branches are distinguished by different colors. \u003cstrong\u003eD\u003c/strong\u003eModule‒trait heatmap of correlations. Red indicates a positive correlation, whereas blue denotes a negative correlation, with rows representing modules and columns representing trait attributes.\u003c/p\u003e","description":"","filename":"OnlineFig.1.png","url":"https://assets-eu.researchsquare.com/files/rs-5118256/v1/6cbbd650b5a68d1498d014ee.png"},{"id":71474576,"identity":"e10357a1-2821-45b3-bd0d-8ab8b2e70d3d","added_by":"auto","created_at":"2024-12-16 04:47:18","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":88656,"visible":true,"origin":"","legend":"\u003cp\u003eIdentification of differentially expressed key module genes and their enrichment analyses. \u003cstrong\u003eA \u003c/strong\u003eHeatmap showing the DEGs identified via color gradients, where red represents high expression and green represents low expression. Darker colors indicate larger differences. \u003cstrong\u003eB \u003c/strong\u003eVolcano plot of the DEG distribution. The green and red dots represent DEGs with low and high expression, respectively. \u003cstrong\u003eC \u003c/strong\u003eVenn diagram showing the intersection genes of the DEGs and the key module genes. \u003cstrong\u003eD \u003c/strong\u003eCircos plots of the GO enrichment. \u003cstrong\u003eE\u003c/strong\u003e KEGG analysis of the key differentially expressed module genes.\u003c/p\u003e","description":"","filename":"OnlineFig.2.png","url":"https://assets-eu.researchsquare.com/files/rs-5118256/v1/86448eba21e34d746b110cfe.png"},{"id":71475412,"identity":"85d3a7ef-b973-46ba-a243-68fabb2db725","added_by":"auto","created_at":"2024-12-16 04:55:18","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":47446,"visible":true,"origin":"","legend":"\u003cp\u003eScreening feature genes. \u003cstrong\u003eA \u003c/strong\u003ePPI network of differentially expressed key module genes. \u003cstrong\u003eB\u003c/strong\u003e Screening of 10 candidate genes on the basis of DMNC values in the CytoHubba plugin. \u003cstrong\u003eC \u003c/strong\u003eCross-validation was employed to select tuning parameters for the LASSO regression model. Optimal levels were identified for five examined genes by creating dotted vertical lines. \u003cstrong\u003eD\u003c/strong\u003e SVM-RFE analysis; importance of genes (left) and predicted true value change curves (right). \u003cstrong\u003eE \u003c/strong\u003eVenn diagram illustrating the identification of feature genes through LASSO and SVM analyses.\u003c/p\u003e","description":"","filename":"OnlineFig.3.png","url":"https://assets-eu.researchsquare.com/files/rs-5118256/v1/020b791d00d9d26f7648a9ec.png"},{"id":71474585,"identity":"f82b4781-e7b1-4566-9f4b-52a4711184a5","added_by":"auto","created_at":"2024-12-16 04:47:19","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":47501,"visible":true,"origin":"","legend":"\u003cp\u003eScreening and verification of biomarkers. \u003cstrong\u003e4A, 4B \u003c/strong\u003eTranscription level validation of two datasets (A, RNA sequencing data; B, external verification set).\u003cstrong\u003e 4C \u003c/strong\u003eWestern blotting was used to verify the differences in the protein levels of the 4 biomarkers. \u003cstrong\u003e4D, 4E, 4F,\u003c/strong\u003e and \u003cstrong\u003e4G\u003c/strong\u003e Biomarker-related ROC analysis. \u003cstrong\u003e4H \u003c/strong\u003ecorrelation analysis revealed strong positive correlations (cor\u0026gt;0.75, P\u0026lt;0.05) among the four biomarkers.\u003c/p\u003e","description":"","filename":"OnlineFig.4.png","url":"https://assets-eu.researchsquare.com/files/rs-5118256/v1/db37700e7eecb4d46f92e75c.png"},{"id":71474581,"identity":"e87669b5-d1df-4b39-8dc6-6d6f72f60c37","added_by":"auto","created_at":"2024-12-16 04:47:19","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":39275,"visible":true,"origin":"","legend":"\u003cp\u003eGSEA enrichment analysis \u003cstrong\u003e5A-5D\u003c/strong\u003e GSEA–KEGG analysis of four biomarkers (A NID2, B MFAP2, C COL8A1, D P4HA3).\u003c/p\u003e","description":"","filename":"OnlineFig.5.png","url":"https://assets-eu.researchsquare.com/files/rs-5118256/v1/93cd68100a860639aa9ebace.png"},{"id":71474571,"identity":"1cdfdc97-468e-4180-9e63-8dabf2b68af2","added_by":"auto","created_at":"2024-12-16 04:47:18","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":91711,"visible":true,"origin":"","legend":"\u003cp\u003eImmune infiltration analysis \u003cstrong\u003eA \u003c/strong\u003eHeatmap of\u003cstrong\u003e \u003c/strong\u003eimmune cell infiltration. \u003cstrong\u003eB \u003c/strong\u003eCorrelation heatmap showing immune cell relationships: red indicates positive correlations, blue indicates negative correlations, and darker colors indicate stronger correlations. \u003cstrong\u003eC\u003c/strong\u003e Differences in immune cell types between KD and control samples. \u003cstrong\u003eD \u003c/strong\u003eCorrelations of the four biomarkers with various differential immune cells.\u003c/p\u003e","description":"","filename":"OnlineFig.6.png","url":"https://assets-eu.researchsquare.com/files/rs-5118256/v1/177aa427ac3444c53ef422c3.png"},{"id":71474607,"identity":"96483cc0-94d0-49d9-a08a-ec3b7637d9eb","added_by":"auto","created_at":"2024-12-16 04:47:21","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":49232,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRegulatory network analysis of biomarkers. 7A\u003c/strong\u003e TF‒mRNA network (red indicates mRNAs, and cyan blue diamonds indicate TFs). \u003cstrong\u003e7B\u003c/strong\u003e LncRNA-miRNA‒mRNA network.\u003c/p\u003e","description":"","filename":"OnlineFig.7.png","url":"https://assets-eu.researchsquare.com/files/rs-5118256/v1/d1e731202cbc74e045869818.png"},{"id":71474610,"identity":"8a837579-b024-4298-ab18-ba4ffa0ecfc3","added_by":"auto","created_at":"2024-12-16 04:47:21","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":30182,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSubcellular localization of biomarkers and drug prediction. 8A \u003c/strong\u003eSubcellular localization networks for biomarkers (pink represents biomarkers, yellow diamonds represent subcellular biomarkers). \u003cstrong\u003e8B\u003c/strong\u003eBiomarkers for the drug modulation network (red indicates a biomarker, purple indicates a drug that interacts with all four biomarkers, blue indicates a drug that interacts with three biomarkers, yellow indicates a drug that interacts with two biomarkers, and greenish blue indicates a drug that interacts with only one biomarker).\u003c/p\u003e","description":"","filename":"OnlineFig.8.png","url":"https://assets-eu.researchsquare.com/files/rs-5118256/v1/4f030c6930d7560d51e34792.png"},{"id":86179229,"identity":"69a175e4-fc79-4021-8d0e-299b80a44ca4","added_by":"auto","created_at":"2025-07-07 16:17:25","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1770180,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5118256/v1/c0d0fcce-e6d7-4db4-a09e-3360bac0435b.pdf"},{"id":71475415,"identity":"a5a91604-da3a-4442-85ba-19ae86a416af","added_by":"auto","created_at":"2024-12-16 04:55:19","extension":"csv","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":44970,"visible":true,"origin":"","legend":"","description":"","filename":"TableS1.csv","url":"https://assets-eu.researchsquare.com/files/rs-5118256/v1/0156f019ea40c8cb37bd5e31.csv"},{"id":71474578,"identity":"08184329-4539-4ead-8109-740a9a54d955","added_by":"auto","created_at":"2024-12-16 04:47:18","extension":"csv","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":31779,"visible":true,"origin":"","legend":"","description":"","filename":"TableS2.csv","url":"https://assets-eu.researchsquare.com/files/rs-5118256/v1/49dd0b6047e39a2e1df3faf2.csv"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eIdentification of Biomarkers and Mechanisms for Keloid Disorder based on Comprehensive Bioinformatics Analysis and Machine Learning Algorithms\u003c/p\u003e","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eKeloid disorder (KD) is a specific type of abnormal scar formation characterized by raised, firm, and reddish growth that extends beyond the original wound site. The current understanding of the pathogenesis of KD is excessive deposition of collagen during the healing process, which results in a fibroproliferative inflammatory reaction with neovascularization at the advancing edge. KD can be itchy, painful, and cause skin tightness, impacting a person's appearance and quality of life[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eDuring wound healing, fibroblasts and myofibroblasts play key roles in remodeling and forming the collagen matrix. They create a stiffer extracellular matrix[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. The KD matrix is composed of various types of glycoproteins, collagens, and glycosaminoglycans[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Initially, the KD matrix is characterized by an overproduction of type III collagen, which over time is gradually replaced by a significant proportion of type I collagen. As a result, the fibers within KD are typically larger in size than those found in normal skin[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Owing to their distinctive morphology and clinically aggressive nature, KD resulting from dermal trauma is often classified as a type of nonspecific etiologic or benign fibroproliferative reticular skin tumor. The morphology and clinical features of KD are similar to those of neoplastic skin tumors, which exhibit aggressive growth, tissue invasion and excessive vascularization[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eKD is a common problem worldwide, but certain populations are more predisposed to developing KD. Epidemiological studies have shown that KD is more prevalent in individuals with a darker skin phototype, including those of African, Hispanic, and Asian descent. Additionally, KD tends to affect younger individuals and is more common in females than in males. The exact prevalence of KD varies across different regions and populations, with some studies estimating that up to 10% of people may develop KD at some point in their lives. Genetic factors, environmental influences, and variations in wound healing processes all play a role in the development of KD. Current treatment methods for KD include corticosteroid injections, silicone gel sheets, cryotherapy, laser therapy, surgical excision, and radiation therapy[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. However, these treatments may have limitations, such as potential recurrence, discomfort during treatment, and variable effectiveness depending on the individual's response to therapy[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eSeveral general pathologic etiologies have been described for KD development, although no specific factor has been identified as the definitive cause. The immune microenvironment heavily influences KD formation, which represents another key factor in this process. This condition arises from an excessive inflammatory reaction following tissue injury. The presence of an increased number of inflammatory cells, especially T cells and macrophages, within KD tissues is a key feature of this pathological state[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. These immune cells play a significant role by releasing a wide range of cytokines and chemokines, such as interleukins (ILs), transforming growth factor β, and fibronectin extra domain A (Fn-EDA), which are essential mediators involved in collagen deposition during KD formation[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. However, the exact pathogenesis and molecular mechanisms underlying KD development are still not fully understood. Further research is needed to elucidate the precise mechanisms driving the occurrence and progression of KD, which will be essential in the development of more effective and targeted treatment strategies tailored to individuals affected by this condition.\u003c/p\u003e \u003cp\u003eIn this study, KD and normal control skin samples were first collected for RNA sequencing to identify differentially expressed genes (DEGs). These genes subsequently intersected with the most relevant key modular genes for KD. In addition, biomarkers were obtained via machine learning combined with expression validation, and receiver operating characteristic (ROC) curves were generated to assess the predictive performance of biomarkers for KD. For immune infiltration analysis, the infiltration levels of 28 immune cells and their correlations with biomarkers were analyzed. We constructed ceRNA regulatory networks to explore the regulatory mechanisms of biomarkers in KD. Finally, subcellular localization of biomarkers was performed, and possible drugs for the treatment of KD were investigated.\u003c/p\u003e"},{"header":"2. Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Sample collection and data sources\u003c/h2\u003e \u003cp\u003eA total of 13 skin tissues from KD patients (KD samples) and 14 normal control skin tissues (control samples) were collected for RNA sequencing. The samples used were obtained from Lanzhou University Second Hospital Hospital, and the Animal Welfare and Ethics Committee of Lanzhou University Second Hospital (D2021\u0026ndash;256) approved the study (2023A\u0026ndash;324). Furthermore, a KD-related dataset (GSE158395) (GPL24676) was obtained from the Gene Expression Omnibus (GEO) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"http://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e); four KD samples (Keloid_LS) and six control samples were selected.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 RNA sequencing (RNA-seq)\u003c/h2\u003e \u003cp\u003eFollowing the manufacturer\u0026rsquo;s instructions, TRIzol (Invitrogen, CA, USA) was used to isolate and purify RNA from total samples. The quantity and quality of total RNA were measured with a NanoDrop ND-1000 (NanoDrop, Wilmington, DE, USA). Next, a Bioanalyzer 2100 (Agilent, CA, USA) was used to evaluate RNA integrity, followed by confirmation via gel electrophoresis. RIN values\u0026thinsp;\u0026gt;\u0026thinsp;7.0, OD260/280\u0026thinsp;\u0026gt;\u0026thinsp;1.8, concentrations\u0026thinsp;\u0026gt;\u0026thinsp;50 ng/L, and total RNA\u0026thinsp;\u0026gt;\u0026thinsp;1 g satisfied the requirements for downstream investigations.\u003c/p\u003e \u003cp\u003eThe mRNA library was then constructed in accordance with the instructions. After the mRNA containing polyA (polyadenylate) was precisely isolated via oligo (dT) magnetic beads and broken up at high temperatures, reverse transcriptase produced the first strand of cDNA. The production of second-strand cDNA was subsequently mediated by RNase H digestion. After the cDNA ends were repaired, the A-tail was introduced, and libraries with fragment sizes of 300 bp\u0026thinsp;\u0026plusmn;\u0026thinsp;50 bp were generated through screening and purification via magnetic beads. Additionally, we implemented 2\u0026times;150 bp paired-end sequencing (PE150) via an Illumina NovaSeqTM 6000 in accordance with the vendor\u0026rsquo;s recommended procedure.\u003c/p\u003e \u003cp\u003eAfterwards, FastQC (v 0.11.9) was deployed to assess data quality, and Trimmomatic (v 0.39) was implemented to filter low-quality data[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Reads were mapped to the \u003cem\u003eHomo sapiens\u003c/em\u003e GRCh38 genome via Bowtie2 and TopHat2. Immediately thereafter, the clean data were compared to the reference genome (hg19) via hisat2 (v 2.2.1)[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e], with all the parameters set to default values. Finally, the gene expression matrix was acquired for further examination.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Weighted gene coexpression network analysis (WGCNA)\u003c/h2\u003e \u003cp\u003eWGCNA was completed via the \u0026lsquo;WGCNA\u0026rsquo; package (v 1.69)[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] utilizing an expression matrix from RNA sequencing data, with the aim of identifying KD-related modules. Initially, unqualified and abnormal samples were identified and excluded through the \u0026lsquo;goodSamplesGenes\u0026rsquo; and \u0026lsquo;hclust\u0026rsquo; functions to ensure the accuracy of the subsequent analyses. Second, selecting the soft threshold was key to constructing a network topology analysis and was based on the nearly scale-free topology criterion. After the optimal soft threshold was determined, a systematic clustering tree was generated by calculating the neighborhood and similarity between genes, followed by building the expression network in accordance with the criteria of the hybrid dynamic tree cutting algorithm (minModuleSize\u0026thinsp;=\u0026thinsp;50 and mergeCutHeight\u0026thinsp;=\u0026thinsp;0.25). Moreover, correlation analysis was performed to assess the relationships between modules and traits (KD and control), and immediately after, we focused on the module most relevant to KD and the module genes it contained (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and |cor|\u0026gt;0.3). These genes were defined as key module genes highly correlated with KD.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Differential expression analysis and enrichment analysis\u003c/h2\u003e \u003cp\u003eFor the RNA sequencing data, the differentially expressed genes (DEGs) between the KD and control samples were mined and analyzed via the \u0026lsquo;DESeq2\u0026rsquo; package (v 1.36.0)[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e], with the filtering conditions of |log\u003csub\u003e2\u003c/sub\u003eFoldChange(FC)|\u0026gt;1.5 and adj.\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Moreover, volcano maps and heatmaps of the DEGs were generated via the \u0026lsquo;ggplot2\u0026rsquo; package (v3.4.1)[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e] and \u0026lsquo;ComplexHeatmap\u0026rsquo; package (v1.1.9)[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], respectively. A Venn diagram of key module genes and DEGs was subsequently generated, and the intersecting genes were considered differentially expressed key modular genes for subsequent analysis. Furthermore, to mine the potential biological functions and signaling pathways linked to key differentially expressed genes, Gene Ontology (GO) annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were performed via the \u0026lsquo;clusterProfiler\u0026rsquo; package (v 4.7.1)[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e], and the cutoff value was \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05. The GO annotations included biological process (BP), cell component (CC) and molecular function (MF) terms.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Protein‒protein interaction (PPI) analysis and machine learning algorithms\u003c/h2\u003e \u003cp\u003eTo explore whether there were interactions among the key differentially expressed genes, these genes were introduced into the online Search Tool for the Retrieval of Interacting Genes (STRING, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://STRING-db.org/\u003c/span\u003e\u003cspan address=\"https://STRING-db.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) to construct a protein\u0026ndash;protein interaction (PPI) network (confidence\u0026thinsp;=\u0026thinsp;0.4). The network was subsequently analyzed via the CytoHubba plugin, followed by the selection of the top 10 genes as candidate genes on the basis of the degree-modified node centrality (DMNC) value. A larger DMNC value indicates that the gene is more important and vice versa.\u003c/p\u003e \u003cp\u003eTwo machine learning methods, least absolute shrinkage and selection operator (LASSO) and support vector machine recursive feature elimination (SVM-RFE), were subsequently utilized to acquire feature genes in KDs on the basis of the candidate genes obtained above. For the RNA sequencing data, LASSO logistic regression analysis was implemented via the \u0026lsquo;glmnet\u0026rsquo; package (v 4.0\u0026ndash;2)[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e], and the lambda.min value was applied to select the feature genes. The \u0026lsquo;caret\u0026rsquo; package (v 6.0\u0026ndash;93) was utilized to achieve SVM-RFE, which ranks various features according to their predictive potential. In the end, the feature genes identified by the above two machine learning algorithms overlapped, and the intersecting feature genes obtained were applied for follow-up analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Expression level analysis and diagnostic performance evaluation\u003c/h2\u003e \u003cp\u003eTo further clarify the expression levels of the intersecting feature genes in the KD and control samples, the expression data of these genes in the RNA sequencing data and the GSE158395 dataset were compared via the Wilcoxon test, with a significance threshold set at \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Importantly, genes whose expression was significantly different between the KD and control samples, as well as genes whose expression trend was consistent between the two datasets, were selected as biomarkers for the diagnosis of KD. To further validate the expression levels of these biomarkers, Western blot analysis was also performed. Briefly, total proteins were extracted from samples collected from the KD and control groups, and the protein concentrations were determined via the BCA method. Equal amounts of protein samples were separated by SDS‒PAGE and transferred to PVDF membranes, followed by incubation with specific primary antibodies. Chemiluminescence detection was used to measure the relative expression levels of the target proteins, and semiquantitative analysis was conducted via ImageJ software. The Western blot results further confirmed the differential expression of the biomarkers between the KD and control groups. Moreover, the \u0026lsquo;pROC\u0026rsquo; package (v 1.18.0)[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e] was employed to create receiver operating characteristic (ROC) curves in the RNA sequencing data with the aim of assessing the biomarkers\u0026rsquo; predictive accuracy for KD. Depending on the area under the curve (AUC), the prediction precision was classified as low (0.5\u0026ndash;0.7), medium (0.7\u0026ndash;0.9), or high (\u0026gt;\u0026thinsp;0.9). Finally, correlations between biomarkers were revealed via a Spearman test via the \u0026lsquo;corrplot\u0026rsquo; package (v 0.92)[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Gene set enrichment analysis (GSEA)\u003c/h2\u003e \u003cp\u003eTo further uncover the underlying biological mechanisms associated with the biomarkers, the samples in the RNA sequencing data were categorized into high- and low-expression groups according to the median value of expression for each biomarker, after which the DEGs between the two expression groups were determined via the \u0026lsquo;DEseq2\u0026rsquo; package (v 1.36.0)[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] and sorted according to logFC. On the basis of KEGG gene sets from the Molecular Signatures Database (MSigDB, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gsea-msigdb.org/gsea/msigdb/\u003c/span\u003e\u003cspan address=\"https://www.gsea-msigdb.org/gsea/msigdb/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), GSEA of the sorted genes was subsequently executed via the \u0026lsquo;clusterProfiler\u0026rsquo; package (v 4.7.1)[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. The threshold values were set at |NES|\u0026ge;1 and adj.\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Immune infiltration analysis\u003c/h2\u003e \u003cp\u003eImmune cells are known to be involved in the pathogenesis of KD[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Therefore, this work applied the single-sample gene set enrichment analysis (ssGSEA) method to compute enrichment scores of 28 immune infiltrating cells from samples of RNA sequencing data via the \u0026lsquo;GSVA\u0026rsquo; package (v 1.40.1)[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e], followed by the Wilcoxon test to compare differential immune cells between KD and control samples (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). To further illuminate the role of immune infiltrating cells and biomarkers in the immune microenvironment, correlation analysis between immune infiltrating cells as well as between differential immune cells and biomarkers was implemented via the Spearman test.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e2.9 Regulatory mechanism analysis\u003c/h2\u003e \u003cp\u003eTranscription factor (TF)-mRNA and lncRNA‒miRNA-mRNA regulatory networks were constructed to investigate the molecular regulatory mechanisms of biomarkers in KD. First, potential TFs regulating the biomarkers were retrieved by accessing the NetworkAnalyst database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.networkanalyst.ca/\u003c/span\u003e\u003cspan address=\"https://www.networkanalyst.ca/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), and the TF‒mRNA network was generated via Cytoscape software (v 3.9.1)[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. The target miRNAs of the biomarkers were subsequently predicted via the miRanda (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.microrna.org/\u003c/span\u003e\u003cspan address=\"http://www.microrna.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and miRDB (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://mirdb.org/\u003c/span\u003e\u003cspan address=\"https://mirdb.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) databases, and the intersection of the two databases was used, after which the StarBase database was accessed to obtain the upstream lncRNAs of the miRNAs (clipExpNum\u0026thinsp;\u0026gt;\u0026thinsp;4). Ultimately, the lncRNA‒miRNA‒mRNA network was visualized via the \u0026lsquo;ggalluvial\u0026rsquo; package (v 0.12.5) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://corybrunson.github.io/ggalluvial/\u003c/span\u003e\u003cspan address=\"http://corybrunson.github.io/ggalluvial/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e2.10 Subcellular localization analysis and drug prediction\u003c/h2\u003e \u003cp\u003eBiomarkers were imported into the GeneCards database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.genecards.org/\u003c/span\u003e\u003cspan address=\"http://www.genecards.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) to extract subcellular distribution information, and the subcellular-biomarker network was visualized via Cytoscape software (v 3.9.1)[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e] (confidence\u0026thinsp;\u0026ge;\u0026thinsp;2). Finally, the comparative toxicogenomics database (CTD, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://ctdbase.org/\u003c/span\u003e\u003cspan address=\"http://ctdbase.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was adopted to predict drugs associated with biomarkers to explore potential drugs for the treatment of KD.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e2.11 Statistical analysis\u003c/h2\u003e \u003cp\u003eAll bioinformatic projects were conducted in R software. Differences were considered significant at P\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.1 In total, 809 key module genes related to KD were obtained through WGCNA\u003c/h2\u003e \u003cp\u003eWith respect to WGCNA, the sample clustering results revealed a lack of outliers in the samples from the RNA sequencing data (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). When the soft threshold was equal to 24, R\u003csup\u003e2\u003c/sup\u003e was greater than 0.85 (red line), and the mean connectivity was close to 0, which was consistent with a scale-free network distribution (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). Furthermore, the dynamic tree cutting algorithm yielded 11 coexpression modules (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC). In terms of the correlations between modules and traits (KD and control), the MEgreen module had the highest correlation with KD (cor\u0026thinsp;=\u0026thinsp;0.86 and \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01), and 809 genes in the MEgreen module were identified as key module genes highly associated with KD \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Differentially expressed key module genes associated with collagen- and bone-associated biological functions\u003c/h2\u003e \u003cp\u003eIn total, 1,898 DEGs between the KD and control samples were mined from the RNA sequencing data, with 900 upregulated and 998 downregulated genes \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA-\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB\u003cb\u003e)\u003c/b\u003e. The Venn diagram revealed that 420 differentially expressed key module genes were acquired by extracting the intersections of 1,898 DEGs and 809 key module genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC). Furthermore, GO annotation and KEGG pathway analyses were performed to elucidate the underlying biological mechanisms of the 420 genes, yielding a total of 192 GO entries (146 BPs, 17 CCs, and 29 MFs) and nine KEGG signaling pathways (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). In the GO-BP category, these 420 genes were enriched primarily in \u0026ldquo;extracellular matrix organization\u0026rdquo;, \u0026ldquo;collagen fibril organization\u0026rdquo;, \u0026ldquo;bone development\u0026rdquo;, \u0026ldquo;collagen biosynthetic process\u0026rdquo;, and \u0026ldquo;cartilage development\u0026rdquo; (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD \u003cb\u003eand Table S1)\u003c/b\u003e. The GO-CC analysis results showed that 420 genes were significantly enriched in \u0026ldquo;collagen-containing extracellular matrix\u0026rdquo;, \u0026ldquo;complex of collagen trimers\u0026rdquo;, and \u0026ldquo;fibrillar collagen trimers\u0026rdquo; (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD \u003cb\u003eand Table S1)\u003c/b\u003e. GO-MF terms were enriched in \u0026ldquo;collagen binding\u0026rdquo; and \u0026ldquo;integrin binding\u0026rdquo; (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD \u003cb\u003eand Table S1)\u003c/b\u003e. Moreover, KEGG analysis revealed that 420 genes were enriched in pathways related to \u0026ldquo;ECM-receptor interaction\u0026rdquo;, \u0026ldquo;PI3K-Akt signaling pathway\u0026rdquo;, and \u0026ldquo;cell adhesion molecules\u0026rdquo; (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE \u003cb\u003eand Table S2)\u003c/b\u003e. Overall, collagen- and bone-associated biological functions and cell-associated signaling pathways appear to be key factors in the regulation of KD.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Identification of feature genes through LASSO and SVM-RFE\u003c/h2\u003e \u003cp\u003eThe PPI network of 420 differentially expressed key module genes, including 420 points and 2,111 edges, is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA, followed by the identification of 10 candidate genes (LOXL1, COL16A1, CD248, NID2, P4HA3, LAMB1, P3H3, COL8A1, MFAP2, and LOXL2) on the basis of DMNC values in the CytoHubba plugin (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB). These 10 candidate genes were subsequently incorporated into machine learning algorithms to further identify feature genes in the RNA sequencing data. Through the LASSO algorithm, five genes were identified as feature genes of KD: NID2, MFAP2, COL8A1, P4HA3, and CD248 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC\u003cb\u003e)\u003c/b\u003e. Moreover, eight feature genes (COL8A1, MFAP2, P4HA3, LOXL2, LOXL1, P3H3, NID2, and CD248) were identified via the SVM-RFE algorithm (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD). Moreover, the feature genes acquired via the above two methods were intersected to acquire five intersecting feature genes (NID2, MFAP2, COL8A1, P4HA3, and CD248) for subsequent analysis \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e3.4 NID2, MFAP2, COL8A1, and P4HA3 are biomarkers of KD\u003c/h2\u003e \u003cp\u003eTranscription and protein expression level analyses were carried out to identify biomarkers. Notably, NID2, MFAP2, COL8A1, and P4HA3 were significantly differentially expressed between the KD and control samples (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) and presented consistent expression patterns (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA-\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC\u003cb\u003e)\u003c/b\u003e. Therefore, these four genes were utilized as biomarkers for KD in this study, and the expression of all four genes was markedly greater in the KD samples than in the control samples. Moreover, the ROC curves of the biomarkers in the RNA sequencing data suggested that the AUC values (NID2\u0026thinsp;=\u0026thinsp;0.967, MFAP2\u0026thinsp;=\u0026thinsp;1, COL8A1\u0026thinsp;=\u0026thinsp;1, and P4HA3\u0026thinsp;=\u0026thinsp;1) were all greater than 0.9 (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD-\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eG\u003cb\u003e)\u003c/b\u003e, implying that they had excellent ability to diagnose KD patients. Correlation analysis revealed significant positive correlations between these four biomarkers (cor\u0026thinsp;\u0026gt;\u0026thinsp;0.75 and \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), with the strongest correlation between MFAP2 and P4HA3 (cor\u0026thinsp;=\u0026thinsp;0.90) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eH\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Biological pathways of biomarkers explored in KD\u003c/h2\u003e \u003cp\u003eTo explore the underlying biological mechanisms of biomarkers, GSEA was performed on RNA sequencing data. The KEGG gene set revealed four biomarkers with enriched pathways, including \u0026ldquo;citrate cycle TCA cycle\u0026rdquo;, \u0026ldquo;metabolism of xenobiotics by cytochrome P450\u0026rdquo;, \u0026ldquo;steroid hormone biosynthesis\u0026rdquo;, \u0026ldquo;PPAR signaling pathway\u0026rdquo;, \u0026ldquo;cytokine‒cytokine receptor interaction\u0026rdquo;, \u0026ldquo;chemokine signaling pathway\u0026rdquo;, \u0026ldquo;TGF-beta signaling pathway\u0026rdquo;, \u0026ldquo;ECM-receptor-interaction\u0026rdquo;, etc. (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD\u003cb\u003e)\u003c/b\u003e, revealing their important roles in the pathogenesis of KD.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e3.6 Biomarkers associated with immune infiltrating cells\u003c/h2\u003e \u003cp\u003eThe mechanisms by which the four biomarkers influence KD progression were examined via immune analysis of RNA sequencing data. The infiltration levels of 28 immune cells in the KD and control samples are depicted in\u0026nbsp;\u003cstrong\u003eFig. 6A.\u003c/strong\u003e Among these 28 immune cells, the strongest positive correlation (cor=0.96) was detected between type 1 T helper cells and regulatory T cells, and the strongest negative correlation (cor=-0.78) was detected between effector memory CD4\u003csup\u003e+\u003c/sup\u003e T cells and CD56\u003csup\u003edim\u003c/sup\u003e natural killer cells (\u003cstrong\u003eFig. 6B)\u003c/strong\u003e. Moreover, 16 immune cell types (activated B cells, central memory CD4+ T cells, central memory CD8+ T cells, macrophages, memory B cells, etc.) differed between the KD and control samples (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.05), and all of them were more abundant in the KD samples than in the control samples (\u003cstrong\u003eFig. 6C)\u003c/strong\u003e,\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eindicating that disease progression might be influenced by these differential immune cells. Furthermore, the correlation of biomarkers with differential immune cells indicated that all four biomarkers were negatively linked to type 17 T helper cells and positively correlated with the remaining 15 immune cells (\u003cstrong\u003eFig. 6D)\u003c/strong\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e3.7 Investigation of the potential regulatory mechanism of biomarkers\u003c/h2\u003e \u003cp\u003eTo elucidate the regulatory mechanisms of biomarkers, four biomarkers were entered into the NetworkAnalyst database, yielding a total of 20 upstream TFs of biomarkers, including seven, nine, seven, and three TFs targeting NID2, MFAP2, COL8A1, and P4HA3, respectively. In this TF-mRNA network, FOXC1 was a regulator of three biomarkers (NID2, MFAP2, and COL8A1), and TFs (FOS and JUN) could coregulate two biomarkers (NID2 and MFAP2) (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA\u003cb\u003e)\u003c/b\u003e. Furthermore, on the basis of the starBase, miRanda, and miRDB databases, a lncRNA‒miRNA‒mRNA network consisting of two biomarkers (NID2 and MFAP2), five miRNAs and 39 lncRNAs (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB\u003cb\u003e)\u003c/b\u003e was constructed. Multiple lncRNA‒miRNA‒mRNA relationship pairs, such as lncRNAs (SNHG1, NEAT1, MALAT1, etc.) that can regulate MFAP2 via miRNAs (hsa-miR-3163), as well as hsa-miR‒199 family members (hsa-miR‒199b‒3p and hsa-miR‒199a‒3p), were identified as regulators of NID2, were found in this network.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e3.8 Subcellular localization and drug prediction\u003c/h2\u003e \u003cp\u003eThe biomarkers were entered into the GeneCards database to analyze their subcellular localization, which indicated that all four genes were expressed in the extracellular space and that COL8A1, MFAP2, and NID2 were expressed in the plasma membrane (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA). By analyzing the CTD database, 86 biomarker-related drugs were obtained, including 40 drugs targeting NID2, 43 drugs targeting MFAP2, 40 drugs targeting COL8A1, and 22 drugs targeting P4HA3 (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eB). Notably, seven drugs (bisphenol A, bisphenol F, bis(4-hydroxyphenyl) sulfone, benzo(a)pyrene, acetaminophen, decamethrin, and calcitriol) were coacting drugs with four biomarkers.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eKD is characterized by the development of excess scar tissue that expands beyond the initial site of skin injury, often affecting individuals with a genetic predisposition to this condition[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. The exact pathophysiological mechanisms underlying KD development are largely unclear[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e] but likely involve skin tension[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e], tissue hypoxia[\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e], chronic inflammation[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e], autoimmunity[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e], and genetic factors[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. KD is further characterized by increased blood vessel proliferation, epidermal thickening, increased mesenchymal cell density, thick and dense hyalinized collagen fibers, and excessive fibronectin accumulation[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. In contrast, normal skin features collagen bundles aligned parallel to the epidermis. Currently, different approaches are used to improve these scars, including intralesional corticosteroids, surgery and, more recently, laser therapy. They can result in itching and pain and adversely affect the physical and psychological well-being of patients. Therefore, finding KD-related biomarkers and exploring prospective targets for anti-KD therapies are highly important.\u003c/p\u003e \u003cp\u003eBy enrichment analysis of candidate genes, we observed significant enrichment in collagen binding, extracellular matrix and collagen fibril organization, bone development, the collagen biosynthetic process, the collagen-containing extracellular matrix, the complex of collagen trimers, and the fibrillar collagen trimer. Furthermore, the KEGG analysis identified enriched pathways, including ECM-receptor interaction, the PI3K-Akt signaling pathway, and cell adhesion molecules. These findings indicate that the reconstruction and deposition of collagen, reorganization of the extracellular matrix, and regulation of signaling pathways are crucial aspects of scar formation and collectively contribute to the development of scar tissue[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. Further investigation into the interaction and regulatory mechanisms of these factors is essential for uncovering the biological processes underlying scar formation.\u003c/p\u003e \u003cp\u003eIn this study, four biomarkers (NID2, MFAP2, COL8A1, and P4HA3) were identified for KD by innovatively combining RNA sequencing, WGCNA, SVM-RFE, and LASSO logistic regression methods. Subsequently, the diagnostic accuracy of the ROC curve was validated. Finally, the consistent expression of the biomarkers in the dataset and clinical samples reinforced the validity of the biomarkers.\u003c/p\u003e \u003cp\u003eNidogen 2 (NID2) is a gene involved in the regulation of cell‒matrix interactions and is known to play a role in maintaining tissue structure and integrity. It can be a new biomarker in some cancer[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. In KD, dysregulation of NID2 expression may lead to abnormal interactions between cells and the extracellular matrix, contributing to the excessive deposition of collagen and extracellular matrix components characteristic of KD. Additionally, NID2 may impact signaling pathways involved in fibroblast proliferation and migration, further promoting the pathological processes observed in KD[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eMicrofibril-associated protein 2(MFAP2) is a gene encoding a protein that is associated with microfibrils in the extracellular matrix. In the context of KD, MFAP2 may influence the organization and stability of the extracellular matrix, potentially contributing to abnormal scarring and fibrotic tissue formation in KD. Dysregulation of MFAP2 expression may disrupt the structural integrity of the extracellular matrix, leading to the aberrant tissue remodeling and excessive collagen deposition observed in KD lesions[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe collagen type VIII alpha 1 chain (COL8A1) gene encodes one of the two alpha chains of type VIII collagen. The gene product is short-chain collagen and a macromolecular component of the subendothelium. The major component of Descemet's membrane (basement membrane) of corneal endothelial cells[\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. Additionally, it is a component of the endothelia of blood vessels. This protein is necessary for the migration and proliferation of vascular smooth muscle cells and thus has a potential role in the maintenance of vessel wall integrity and structure, particularly in atherogenesis[\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. In KD, dysfunction of COL8A1 may lead to aberrant collagen synthesis and deposition, contributing to the excessive fibrosis and scar formation observed in KD lesions. This dysregulation of collagen production and organization may disrupt normal tissue repair processes, leading to the pathological scarring characteristic of KD[\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eProlyl 4-hydroxylase subunit alpha 3 (P4HA3) is a gene encoding a subunit of prolyl 4-hydroxylase, an enzyme involved in collagen synthesis and maturation [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e]. Altered expression of P4HA3 in KD may lead to abnormal collagen cross-linking and deposition, resulting in the formation of dense, fibrotic scar tissue distinctive of KD [\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]. Consequently, P4HA3 likely plays a pivotal role in modulating collagen stability and organization within the extracellular matrix. Abnormalities in P4HA3 could influence collagen remodeling processes, thus contributing to the pathogenesis of KD. In this study, the four biomarkers also showed excellent performance in diagnosing KD, suggesting that biomarkers can be used to diagnose KD more accurately and can be used to achieve targeted therapy for KD by modulating the corresponding pathways.\u003c/p\u003e \u003cp\u003eIn addition, it is important to highlight that inflammatory chemicals are key mediators of immune cell activities. Immune cells invade and regulate the KD microenvironment through the secretion of cytokines[\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. These cytokines, including IL-4, IL-6, IL-10, IL-12, IL-13, IL-17, TNF-α, and TGF-β, are significantly elevated in the local microenvironment of KD[\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e]. They play crucial roles in both the inflammatory phase and the prolonged proliferative phase of KD[\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e]. In contrast, the serum concentrations of IFN-α, IFN-γ, and TNF-β, which are known to suppress collagen synthesis and fibroblast proliferation, are reduced, leading to a lack of inhibition and promoting uncontrolled collagen production[\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eImmune infiltration plays a critical role in KD development. Previous studies have highlighted the significant involvement of various inflammatory cells in the wound-healing process[\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]. For example, the balance between the M1 and M2 phenotypes is crucial for regulating inflammation and tissue healing, both of which are controlled by macrophages. Macrophages in scar tissue are highly activated and polarized toward the M2 subtype [\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e]. M2 macrophages can produce cytokines that stimulate fibroblast proliferation, such as vascular endothelial growth factor, transforming \u003cb\u003egrowth factor beta\u003c/b\u003e, and insulin-like growth factor (IGF)-1. T cells also play a critical role. The proportions of CD8\u0026thinsp;+\u0026thinsp;memory T cells and effector memory CD8\u0026thinsp;+\u0026thinsp;T cells are greater in KD tissue than in normal skin tissue, as commonly observed in inflammatory skin diseases[\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e]. Na\u0026iuml;ve CD4\u0026thinsp;+\u0026thinsp;T cells differentiate into distinct Th lineages, including Th1, Th2, Th17, and Tregs[\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e]. In a study comparing lesional versus normal skin, significant upregulation of markers of T-cell activation/migration (ICOS, CCR7) was observed. Additionally, elevated expression of markers associated with the Th2 pathway (IL4R, CCL11, TNFSF4/OX40L), the Th1 pathway (CXCL9/CXCL10/CXCL11), and the Th17/Th22 pathways (CCL20, S100As) was identified, along with increased JAK/STAT signaling (JAK3). Tregs act by inhibiting CD4\u0026thinsp;+\u0026thinsp;and CD8\u0026thinsp;+\u0026thinsp;T-cell activation and promoting M2 macrophage polarization[\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e]. Treg cells express TGF-β, which enhances COL1A1 and COL3A1 expression, leading to collagen overexpression[\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eRecent research has shed light on the role of dendritic cells (DCs) in influencing fibroblasts and vasculature function[\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e]. Wu et al. reported an elevated presence of DCs (CD11c+) in KD, along with the dendritic cell infiltrates commonly observed in atopic dermatitis[\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e]. Rath et al. reported that inflammatory DCs are the predominant type of DC infiltrating KD tissue, with a notably stronger correlation between inflammatory DCs and fibroblasts in KD tissue than in surrounding or healthy tissue [\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e]. These findings suggest that during wound healing, a KD can recruit and activate inflammatory DC precursors, leading to inflammation[\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e]. Furthermore, the interaction between inflammatory DCs and fibroblasts may stimulate the expression of metalloproteinases in fibroblasts, thereby promoting fibroblast activation and fibrosis[\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn this study, we examined the infiltration level of 28 immune cells in the samples, as immune infiltration significantly affects scar formation. These findings reveal substantial variations in the levels of activated B cells, central memory CD4\u0026thinsp;+\u0026thinsp;T cells, central memory CD8\u0026thinsp;+\u0026thinsp;T cells, macrophages, memory B cells, effector memory CD4\u0026thinsp;+\u0026thinsp;T cells, effector memory CD8\u0026thinsp;+\u0026thinsp;T cells, immature B cells, natural killer cells, natural killer T cells, plasmacytoid dendritic cells, regulatory T cells, type 1 T helper cells, monocytes, MDSCs, and type 17 T helper cells between KD and normal tissues. On the basis of our findings and those of the aforementioned studies, macrophages, plasmacytoid dendritic cells, memory T cells, and T helper cells perform critical functions in KD tissue. This study also demonstrated the relationships between potential biomarkers and immune cells. Specifically, analysis revealed that the biomarkers NID2, MFAP2, COL8A1, and P4HA3 were negatively associated with type 17 T helper cells but positively correlated with the other 15 types of immune cells. In the discussion above, how different immune cells contribute to the development of KD is outlined, and the findings suggest that immunotherapy could address the challenges in treating KD. This study also retrieved small molecule drugs from the CDT database and revealed that seven top-ranked drugs, bisphenol-A, bisphenol-F, bis(4-hydroxyphenyl) sulfone, benzo(a)pyrene, acetaminophen, decamethrin, and calcitriol, were identified as coacting drugs with four specific biomarkers. These drugs may have therapeutic potential for KD. These pharmacological treatments for KD reported in the literature are still rare. Further experimental studies are needed to elucidate the therapeutic potential of these drugs in the context of KD.\u003c/p\u003e \u003cp\u003eCertain limitations of this study warrant recognition, as it was a single-center design. Thus, our results should be validated prospectively in larger cohorts from multiple centers. Furthermore, a further subdivided immune cell subset is needed to study the mechanism behind KD development in those subdivided immune cell subsets. Finally, additional experimental validation, such as cellular and animal experiments, is needed to substantiate these findings and enhance the reliability of the results. The results of our study indicate that NID2, MFAP2, COL8A1, and P4HA3 could influence the KD microenvironment, potentially promoting growth. Future experiments could involve the development of animal models of KD to evaluate changes in scar size and histology after manipulating the expression of potential biomarkers, thus validating their roles in KD growth. Moreover, primary cells such as fibroblasts and keratinocytes can be cultured in cellular experiments to study their impact on activities such as cell proliferation, migration, and apoptosis. Gene therapy is a revolutionary new technique; however, its clinical application in KD remains relatively rare. Confirming the significance and effectiveness of these 4 promising biomarkers in KD through multiple methods can facilitate their clinical application. This approach provides a viable method to treat KD.\u003c/p\u003e \u003cp\u003eIn summary, our comprehensive analysis identified four key differentially expressed genes (NID2, MFAP2, COL8A1, and P4HA3) as potential biomarkers for keloid disease (KD). These biomarkers were found to be significantly associated with the dysregulation of collagen synthesis, extracellular matrix remodeling, and signaling pathways critical for the pathogenesis of KD. Furthermore, we observed distinct patterns of immune cell infiltration, with notable differences in the levels of activated B cells, T-cell subsets, macrophages, and dendritic cells between KD and normal skin samples. The identified biomarkers exhibited strong correlations with these immune cell populations, suggesting their potential involvement in the complex interplay between the immune system and the aberrant wound healing processes underlying KD development.\u003c/p\u003e \u003cp\u003eThese findings provide valuable insights into the molecular and cellular mechanisms driving KD development, and the identified biomarkers hold promise as diagnostic and therapeutic targets. Future research should focus on the functional validation of these biomarkers, both in vitro and in vivo, to further elucidate their specific roles in KD pathogenesis. Additionally, exploring the therapeutic potential of small-molecule drugs that may modulate the identified biomarkers and associated pathways could pave the way for the development of more effective treatment strategies for this debilitating condition. Collectively, our results lay the groundwork for advancing the understanding and management of keloid disease, with the ultimate goal of improving patient outcomes.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eKD Keloid Disorder\u003c/p\u003e \u003cp\u003eEC Extracellular Matrix\u003c/p\u003e \u003cp\u003eLASSO Least Absolute Shrinkage and Selection Operator\u003c/p\u003e \u003cp\u003eSVM-RFE Support Vector Machine Recursive Feature Elimination\u003c/p\u003e \u003cp\u003eFn-EDA Fibronectin Extra Domain A\u003c/p\u003e \u003cp\u003eDEGs Differentially Expressed Genes\u003c/p\u003e \u003cp\u003eROC Receiver Operating Characteristic\u003c/p\u003e \u003cp\u003eWGCNA Weighted Gene Coexpression Network Analysis\u003c/p\u003e \u003cp\u003eGO Gene Ontology\u003c/p\u003e \u003cp\u003eKEGG Kyoto Encyclopedia of Genes and Genomes\u003c/p\u003e \u003cp\u003eBP Biological process\u003c/p\u003e \u003cp\u003eCC Cell Component\u003c/p\u003e \u003cp\u003eMF Molecular Function Terms\u003c/p\u003e \u003cp\u003ePPI Protein Protein Interaction\u003c/p\u003e \u003cp\u003eDMNC Degree-Modified Node Centrality\u003c/p\u003e "},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank the patient\u0026rsquo;s and family for permission to publish these datas.\u0026nbsp;Thanks to the reviewers and editors for their sincere comments.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBowen Zheng\u0026nbsp;wrote the first draft and coordinated the manuscript.\u0026nbsp;Anqi Wang, Hanghang Zhou, Xiaoping Yu and Jianxiong Qiaoprepared genetic data collection and analysis.\u0026nbsp;Xuanfen Zhang\u0026nbsp;contributed to data visualisation, manuscript editing, finalisation. All authors reviewed the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by the Natural Science Foundation of Gansu Province under Grant [number 24JRRA339].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets and materials generated during and/or analysed during the current study are available as follow: https://pan.baidu.com/s/1YirFLmRAZIrkEXC2Nd08GA code: 6666.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe studies involving human participants were reviewed and approved by the Ethic Committee of Lanzhou University Second Hospital. The patients/participants provided their written informed consent to participate in this study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWritten informed consents for publication of samples or other personal or clinical details were obtained from all participants and patients.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eFeng F, Liu M, Pan L, Wu J, Wang C, Yang L, et al. Biomechanical Regulatory Factors and Therapeutic Targets in Keloid Fibrosis. Front Pharmacol. 2022;13:906212.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAshcroft KJ, Syed F, Bayat A. Site-specific keloid fibroblasts alter the behaviour of normal skin and normal scar fibroblasts through paracrine signalling. PLoS ONE. 2013;8:e75600.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSatish L, Lyons-Weiler J, Hebda PA, Wells A. Gene expression patterns in isolated keloid fibroblasts. Wound Repair Regen. 2006;14:463\u0026ndash;70.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVerhaegen PDHM, van Zuijlen PPM, Pennings NM, van Marle J, Niessen FB, van der Horst CMAM, et al. Differences in collagen architecture between keloid, hypertrophic scar, normotrophic scar, and normal skin: An objective histopathological analysis. Wound Repair Regen. 2009;17:649\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTan S, Khumalo N, Bayat A. Understanding Keloid Pathobiology From a Quasi-Neoplastic Perspective: Less of a Scar and More of a Chronic Inflammatory Disease With Cancer-Like Tendencies. Front Immunol. 2019;10:1810.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGauglitz GG. Management of keloids and hypertrophic scars: current and emerging options. Clin Cosmet Investig Dermatol. 2013;6:103\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGauglitz GG, P\u0026ouml;tschke J, Clementoni MT. [Therapy of scars with lasers]. Hautarzt. 2018;69:17\u0026ndash;26.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShaker SA, Ayuob NN, Hajrah NH. Cell talk: a phenomenon observed in the keloid scar by immunohistochemical study. Appl Immunohistochem Mol Morphol. 2011;19:153\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNangole FW, Agak GW. Keloid pathophysiology: fibroblast or inflammatory disorders? JPRAS Open. 2019;22:44\u0026ndash;54.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ede Sena Brandine G, Smith AD. Falco: high-speed FastQC emulation for quality control of sequencing data. F1000Res. 2019;8:1874.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim D, Paggi JM, Park C, Bennett C, Salzberg SL. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat Biotechnol. 2019;37:907\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLangfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008;9:559.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLove MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15:550.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGustavsson EK, Zhang D, Reynolds RH, Garcia-Ruiz S, Ryten M. ggtranscript: an R package for the visualization and interpretation of transcript isoforms using ggplot2. Bioinformatics. 2022;38:3844\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics. 2016;32:2847\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu G, Wang L-G, Han Y, He Q-Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS. 2012;16:284\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFriedman J, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models via Coordinate Descent. J Stat Softw. 2010;33:1\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRobin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez J-C, et al. pROC: an open-source package for R and S\u0026thinsp;+\u0026thinsp;to analyze and compare ROC curves. BMC Bioinformatics. 2011;12:77.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang X, Chao P, Zhang L, Xu L, Cui X, Wang S, et al. Single-cell RNA and transcriptome sequencing profiles identify immune-associated key genes in the development of diabetic kidney disease. Front Immunol. 2023;14:1030198.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLove MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15:550.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu G, Wang L-G, Han Y, He Q-Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS. 2012;16:284\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShan M, Liu H, Song K, Liu S, Hao Y, Wang Y. Immune-related gene expression in skin, inflamed and keloid tissue from patients with keloids. Oncol Lett. 2022;23:72.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH\u0026auml;nzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics. 2013;14:7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH\u0026auml;nzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics. 2013;14:7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13:2498\u0026ndash;504.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTroiano M, Simeone A, Scaramuzzi G, Parisi S, Guglielmi G. Giant keloid of left buttock treated with post-excisional radiotherapy. J Radiol Case Rep. 2011;5:8\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTroiano M, Simeone A, Scaramuzzi G, Parisi S, Guglielmi G. Giant keloid of left buttock treated with post-excisional radiotherapy. J Radiol Case Rep. 2011;5:8\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSong H, Liu T, Wang W, Pang H, Zhou Z, Lv Y, et al. Tension enhances cell proliferation and collagen synthesis by upregulating expressions of integrin αvβ3 in human keloid-derived mesenchymal stem cells. Life Sci. 2019;219:272\u0026ndash;82.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWolfram D, Tzankov A, P\u0026uuml;lzl P, Piza-Katzer H. Hypertrophic scars and keloids\u0026ndash;a review of their pathophysiology, risk factors, and therapeutic management. Dermatol Surg. 2009;35:171\u0026ndash;81.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOgawa R. Keloid and Hypertrophic Scars Are the Result of Chronic Inflammation in the Reticular Dermis. Int J Mol Sci. 2017;18.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKazeem AA. The immunological aspects of keloid tumor formation. J Surg Oncol. 1988;38:16\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBayat A, Arscott G, Ollier WER, McGrouther DA, Ferguson MWJ. Keloid disease: clinical relevance of single versus multiple site scars. Br J Plast Surg. 2005;58:28\u0026ndash;37.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee JY-Y, Yang C-C, Chao S-C, Wong T-W. Histopathological differential diagnosis of keloid and hypertrophic scar. Am J Dermatopathol. 2004;26:379\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBabu M, Diegelmann R, Oliver N. Fibronectin is overproduced by keloid fibroblasts during abnormal wound healing. Mol Cell Biol. 1989;9:1642\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMathew-Steiner SS, Roy S, Sen CK. Collagen in Wound Healing. Bioeng (Basel). 2021;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKiziltan R, Cikman O, Algul S, Aydin MA, Kemik O. Nidogen-2: A new biomarker in colon cancer patients. Ann Ital Chir. 2022;92:88\u0026ndash;92.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMao C, Ma Z, Jia Y, Li W, Xie N, Zhao G, et al. Nidogen-2 Maintains the Contractile Phenotype of Vascular Smooth Muscle Cells and Prevents Neointima Formation via Bridging Jagged1-Notch3 Signaling. Circulation. 2021;144:1244\u0026ndash;61.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe Y, Tsou P-S, Khanna D, Sawalha AH. Methyl-CpG-binding protein 2 mediates antifibrotic effects in scleroderma fibroblasts. Ann Rheum Dis. 2018;77:1208\u0026ndash;18.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYao L-W, Wu L-L, Zhang L-H, Zhou W, Wu L, He K, et al. MFAP2 is overexpressed in gastric cancer and promotes motility via the MFAP2/integrin α5β1/FAK/ERK pathway. Oncogenesis. 2020;9:17.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAldave AJ, Rayner SA, Salem AK, Yoo GL, Kim BT, Saeedian M, et al. No pathogenic mutations identified in the COL8A1 and COL8A2 genes in familial Fuchs corneal dystrophy. Invest Ophthalmol Vis Sci. 2006;47:3787\u0026ndash;90.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu R, Yao ZY, Xin L, Zhang Q, Li TP, Gan RB. NC1 domain of human type VIII collagen (alpha 1) inhibits bovine aortic endothelial cell proliferation and causes cell apoptosis. Biochem Biophys Res Commun. 2001;289:264\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHopfer U, Fukai N, Hopfer H, Wolf G, Joyce N, Li E, et al. Targeted disruption of Col8a1 and Col8a2 genes in mice leads to anterior segment abnormalities in the eye. FASEB J. 2005;19:1232\u0026ndash;44.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKukkola L, Hieta R, Kivirikko KI, Myllyharju J. Identification and characterization of a third human, rat, and mouse collagen prolyl 4-hydroxylase isoenzyme. J Biol Chem. 2003;278:47685\u0026ndash;93.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVan Den Diepstraten C, Papay K, Bolender Z, Brown A, Pickering JG. Cloning of a novel prolyl 4-hydroxylase subunit expressed in the fibrous cap of human atherosclerotic plaque. Circulation. 2003;108:508\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShih B, Garside E, McGrouther DA, Bayat A. Molecular dissection of abnormal wound healing processes resulting in keloid disease. Wound Repair Regen. 2010;18:139\u0026ndash;53.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKihara M, Amano H, Misu Y, Kubo T. Release of [3H]- and endogenous GABA from slices of the rat medulla oblongata: modification by 3-mercaptopropionic acid, nipecotic acid and diaminobutyric acid. Arch Int Pharmacodyn Ther. 1989;298:50\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShan M, Wang Y. Viewing keloids within the immune microenvironment. Am J Transl Res. 2022;14:718\u0026ndash;27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDong X, Mao S, Wen H. Upregulation of proinflammatory genes in skin lesions may be the cause of keloid formation (Review). Biomed Rep. 2013;1:833\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShih B, Garside E, McGrouther DA, Bayat A. Molecular dissection of abnormal wound healing processes resulting in keloid disease. Wound Repair Regen. 2010;18:139\u0026ndash;53.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShaffer JJ, Taylor SC, Cook-Bolden F. Keloidal scars: a review with a critical look at therapeutic options. J Am Acad Dermatol. 2002;46(2 Suppl Understanding):S63\u0026ndash;97.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiao H, Fan J, Cai J, Pan B, Yan L, Dong P, et al. Analysis of Characteristics Similar to Autoimmune Disease in Keloid Patients. Aesthetic Plast Surg. 2015;39:818\u0026ndash;25.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJin Q, Gui L, Niu F, Yu B, Lauda N, Liu J, et al. Macrophages in keloid are potent at promoting the differentiation and function of regulatory T cells. Exp Cell Res. 2018;362:472\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTardito S, Martinelli G, Soldano S, Paolino S, Pacini G, Patane M, et al. Macrophage M1/M2 polarization and rheumatoid arthritis: A systematic review. Autoimmun Rev. 2019;18:102397.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen Z, Zhou L, Won T, Gao Z, Wu X, Lu L. Characterization of CD45RO(+) memory T lymphocytes in keloid disease. Br J Dermatol. 2018;178:940\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu J, Yamane H, Paul WE. Differentiation of effector CD4 T cell populations (*). Annu Rev Immunol. 2010;28:445\u0026ndash;89.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmidt A, Oberle N, Krammer PH. Molecular mechanisms of treg-mediated T cell suppression. Front Immunol. 2012;3:51.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen Y, Jin Q, Fu X, Qiao J, Niu F. Connection between T regulatory cell enrichment and collagen deposition in keloid. Exp Cell Res. 2019;383:111549.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu TT. Dendritic cells: novel players in fibrosis and scleroderma. Curr Rheumatol Rep. 2012;14:30\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu J, Del Duca E, Espino M, Gontzes A, Cueto I, Zhang N, et al. RNA Sequencing Keloid Transcriptome Associates Keloids With Th2, Th1, Th17/Th22, and JAK3-Skewing. Front Immunol. 2020;11:597741.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRath M, Pitiot A, Kirr M, Fr\u0026ouml;hlich W, Plosnita B, Schliep S et al. Multi-Antigen Imaging Reveals Inflammatory DC, ADAM17 and Neprilysin as Effectors in Keloid Formation. Int J Mol Sci. 2021;22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLagares D, Ghassemi-Kakroodi P, Tremblay C, Santos A, Probst CK, Franklin A, et al. ADAM10-mediated ephrin-B2 shedding promotes myofibroblast activation and organ fibrosis. Nat Med. 2017;23:1405\u0026ndash;15.\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":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-genomics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"mgnm","sideBox":"Learn more about [BMC Medical Genomics](http://bmcmedgenomics.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/mgnm/default.aspx","title":"BMC Medical Genomics","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Keloid disorder, Biomarker, Molecular mechanisms, Bioinformatics","lastPublishedDoi":"10.21203/rs.3.rs-5118256/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5118256/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eKeloid disorder (KD) is a group of fibroproliferative skin disorders characterized by hypervascularity and excessive accumulation of the extracellular matrix (ECM) and affects individuals of all age groups. The etiology of KD is complex and still poorly understood. This study aimed to investigate biomarkers and therapeutic targets in KD on the basis of comprehensive bioinformatics analysis and machine learning of RNA autosequencing data.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eThirteen skin tissues from KD patients (KD samples) and 14 normal control skin tissues (control samples) were collected for RNA sequencing. Initially, differentially expressed key module genes were acquired through expression analysis with weighted gene coexpression network analysis, followed by enrichment analysis. The 10 candidate genes obtained via the CytoHubba plugin were subsequently incorporated into the least absolute shrinkage and selection operator (LASSO) and support vector machine recursive feature elimination (SVM-RFE) to recognize feature genes associated with KD. Furthermore, biomarkers were determined via expression level analysis, followed by enrichment analysis and immunoinfiltration analysis to elucidate the pathogenesis of KD.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eA total of 420 differentially expressed key module genes were identified, and these 420 genes were enriched in collagen- and bone-associated biological functions, including \u0026ldquo;collagen fibril organization\u0026rdquo; and \u0026ldquo;bone development\u0026rdquo;. With respect to the 10 candidate genes, five feature genes were subsequently obtained through LASSO and SVM-RFE, and among them, NID2, MFAP2, COL8A1, and P4HA3 had significant expression differences between the KD and control samples as well as consistent expression patterns in both datasets; these genes were considered biomarkers. These four biomarkers had excellent abilities to diagnose KD patients, and there were significant positive correlations between these four biomarkers. Functional enrichment analysis suggested that the main enriched KEGG pathways for biomarkers were \u0026ldquo;steroid hormone biosynthesis\u0026rdquo;, \u0026ldquo;cytokine\u0026ndash;cytokine receptor interaction\u0026rdquo;, etc. Furthermore, immune analysis suggested that four biomarkers were negatively linked to type 17 T helper cells and positively linked to 15 immune cells (activated B cells, central memory CD4 T cells, etc.).\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eNID2, MFAP2, COL8A1, and P4HA3 were identified as biomarkers for KD, providing more targeted and effective diagnostic and therapeutic strategies for KD.\u003c/p\u003e","manuscriptTitle":"Identification of Biomarkers and Mechanisms for Keloid Disorder based on Comprehensive Bioinformatics Analysis and Machine Learning Algorithms","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-12-16 04:47:13","doi":"10.21203/rs.3.rs-5118256/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-11-04T12:50:39+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-24T20:03:59+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-12T07:48:07+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-11T17:08:30+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-07T09:25:45+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-06T05:41:57+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-02T19:20:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"236624162087080875120152941692486784330","date":"2024-10-02T06:09:52+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"192023764928082749251137821644240631433","date":"2024-09-29T17:56:37+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"208573363350907456587720976665837073014","date":"2024-09-28T04:08:56+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-09-27T23:50:35+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"261318485565956223979697730755311591664","date":"2024-09-27T18:15:26+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"312679685558893803786732528627971062381","date":"2024-09-27T16:53:15+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"18437802430102258302670474035532767322","date":"2024-09-27T09:35:39+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"38678789085398724205534109048425655082","date":"2024-09-27T06:31:24+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"85920622556523001149162398852072364839","date":"2024-09-27T05:53:52+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"176526435567999232820671995264253174456","date":"2024-09-27T05:22:34+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"253466424409666898087135777852860345723","date":"2024-09-25T10:43:06+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-09-25T09:25:08+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2024-09-20T16:14:42+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-09-20T13:31:42+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-09-20T13:27:25+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Genomics","date":"2024-09-19T15:48:30+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-genomics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"mgnm","sideBox":"Learn more about [BMC Medical Genomics](http://bmcmedgenomics.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/mgnm/default.aspx","title":"BMC Medical Genomics","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9f1b6333-d9ed-4e22-89d4-9e1829cbda11","owner":[],"postedDate":"December 16th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-07-07T16:07:42+00:00","versionOfRecord":{"articleIdentity":"rs-5118256","link":"https://doi.org/10.1186/s12920-025-02174-9","journal":{"identity":"bmc-medical-genomics","isVorOnly":false,"title":"BMC Medical Genomics"},"publishedOn":"2025-07-01 15:58:27","publishedOnDateReadable":"July 1st, 2025"},"versionCreatedAt":"2024-12-16 04:47:13","video":"","vorDoi":"10.1186/s12920-025-02174-9","vorDoiUrl":"https://doi.org/10.1186/s12920-025-02174-9","workflowStages":[]},"version":"v1","identity":"rs-5118256","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5118256","identity":"rs-5118256","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}