Identification of Key Genes Involved in Lactic Acid Metabolism in Periodontitis Based on Bioinformatics Analysis.

Identification of Key Genes Involved in Lactic Acid Metabolism in Periodontitis Based on Bioinformatics Analysis - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice J Cell Mol Med . 2026 Apr 17;30(8):e71141. doi: 10.1111/jcmm.71141 Search in PMC Search in PubMed View in NLM Catalog Add to search Identification of Key Genes Involved in Lactic Acid Metabolism in Periodontitis Based on Bioinformatics Analysis Bijun Zhu Bijun Zhu 1 Department of Cariology and Endodontics, Nanjing Stomatological Hospital, Affiliated Hospital of Medical School, Institute of Stomatology, Nanjing University, Nanjing, Jiangsu, China Find articles by Bijun Zhu 1 , Shi Xu Shi Xu 1 Department of Cariology and Endodontics, Nanjing Stomatological Hospital, Affiliated Hospital of Medical School, Institute of Stomatology, Nanjing University, Nanjing, Jiangsu, China Find articles by Shi Xu 1 , Mingyue Cheng Mingyue Cheng 1 Department of Cariology and Endodontics, Nanjing Stomatological Hospital, Affiliated Hospital of Medical School, Institute of Stomatology, Nanjing University, Nanjing, Jiangsu, China Find articles by Mingyue Cheng 1 , Leiying Miao Leiying Miao 2 Department of Endodontics, The Affiliated Stomatological Hospital of Nanjing Medical University, Nanjing, Jiangsu, China Find articles by Leiying Miao 2, ✉ Author information Article notes Copyright and License information 1 Department of Cariology and Endodontics, Nanjing Stomatological Hospital, Affiliated Hospital of Medical School, Institute of Stomatology, Nanjing University, Nanjing, Jiangsu, China 2 Department of Endodontics, The Affiliated Stomatological Hospital of Nanjing Medical University, Nanjing, Jiangsu, China * Correspondence: Leiying Miao ( miaoleiying80@163.com ) ✉ Corresponding author. Revised 2026 Mar 10; Received 2025 Aug 30; Accepted 2026 Mar 31; Collection date 2026 Apr. © 2026 The Author(s). Journal of Cellular and Molecular Medicine published by Foundation for Cellular and Molecular Medicine and John Wiley & Sons Ltd. This is an open access article under the terms of the http://creativecommons.org/licenses/by/4.0/ License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. PMC Copyright notice PMCID: PMC13090170  PMID: 41998822 ABSTRACT Lactic acid metabolism correlates with periodontitis progression. Using machine learning (LASSO and SVM‐RFE) on GSE16134 and GSE173078 datasets, CFI and COQ2 were identified as key genes, with good diagnostic performance (AUC > 0.75). GSEA linked these genes to MAPK signalling, endocytosis, and ubiquitin‐mediated proteolysis. Immune infiltration analysis revealed higher plasma cells and lower CD8 T cells in periodontitis. Molecular docking identified amiodarone as a potential compound with strong binding affinity to CFI and COQ2. The expression trends of CFI (upregulated) and COQ2 (downregulated) were first confirmed in TNF‐α/IL‐1β‐stimulated human periodontal ligament cells and further validated in a rat ligature‐induced periodontitis model. These findings suggest that CFI and COQ2 are potential diagnostic biomarkers and therapeutic targets for periodontitis. Keywords: immune infiltration analysis, lactic acid metabolism, machine learning, periodontitis 1. Introduction Periodontitis is a chronic inflammatory disease caused by dysbiosis of the oral microbiome and immune dysregulation. Currently recognized as the sixth most prevalent disease globally, it severely affects the health of nearly 10% of the world's population and imposes a substantial burden on healthcare expenditures [ 1 ]. The hallmark of periodontitis is the destruction of periodontal supporting tissues, characterized by clinical attachment loss, periodontal pocket formation, alveolar bone resorption, and tooth loss. Additionally, periodontitis has been linked to systemic conditions, including cardiovascular diseases [ 2 ] and diabetes [ 3 ]. Early clinical symptoms of periodontitis, such as gingival bleeding and halitosis, are often subtle, leading to delayed patient consultation. Traditional diagnostic methods, relying on periodontal probing depth and radiographic imaging, lack sensitivity and fail to accurately reflect disease activity [ 4 ]. Therefore, identifying key molecular biomarkers associated with the pathogenesis of periodontitis is critical for developing novel diagnostic tools and targeted therapies. The pathogenesis of periodontitis is primarily attributed to the colonization of specific pathogenic microorganisms in dental plaque and dysregulated host immune responses. However, recent studies suggest that metabolic disturbances, including alterations in lactate, cholesterol, and D‐glucose levels, may contribute to disease progression, as evidenced by distinct metabolic profiles between gingivitis patients and healthy controls [ 5 ]. This implies that microenvironmental metabolic shifts may play a pivotal role in disease advancement. Lactate, the end product of the glycolytic pathway, was traditionally regarded as a metabolic waste. Emerging evidence, however, highlights its multifaceted roles as an energy substrate, signalling molecule, and immunomodulator, participating in cellular metabolic reprogramming, angiogenesis, inflammatory responses, and immune evasion [ 6 ]. A novel role of lactate is protein lactylation, which has recently emerged as a post‐translational modification of proteins to regulate gene expression [ 7 ]. For instance, lactate has been implicated in tumour progression and inflammatory diseases, yet its association with periodontitis remains unclear. The role of lactate in inflammatory diseases is complex and context‐dependent. On one hand, lactate enhances HIF‐1α stability, promoting the secretion of pro‐inflammatory cytokines such as IL‐1β and TNF‐α [ 8 ]. Conversely, lactate suppresses NF‐κB signalling via GPR81 receptor binding [ 9 , 10 , 11 ], or modulates anti‐inflammatory factor secretion through lactylation modifications [ 12 ]. A study has reported decreased lactate production and lactylation levels in the periodontal tissues of periodontitis rats and LPS‐stimulated human PDLSCs [ 13 ]. These findings suggest that lactate metabolism‐related genes may influence periodontitis progression, though their precise regulatory networks require further exploration. This study leverages public datasets to identify lactate metabolism‐related differentially expressed genes (DEGs) through differential expression analysis. Key genes are validated using machine learning, expression validation, and receiver operating characteristic (ROC) curve analysis. Subsequent functional enrichment analysis, immune infiltration profiling, molecular network construction, drug prediction, and molecular docking are performed to elucidate the molecular mechanisms of these key genes. Finally, the expression of DEGs was validated in vivo. The findings aim to provide novel insights for the diagnosis and treatment of periodontitis (Figure 1 ). FIGURE 1. Open in a new tab Flowchart of this study. Training and validation datasets were screened from public databases. Multiple bioinformatic analyses—including enrichment analysis, immune cell infiltration assessment, and molecular regulatory network construction—were integrated to identify and validate key genes. The study culminated in in vivo validation of differentially expressed genes (DEGs). 2. Material and Methods 2.1. Data Collection The datasets relevant to periodontitis were retrieved from the Gene Expression Omnibus (GEO) database ( http://www.ncbi.nlm.nih.gov/geo/ ). Among them, the dataset GSE16134 (platform GPL570 ) encompassed gingival tissue samples from 241 periodontitis and 69 control, which was designated as a training set. The validation set GSE173078 (platform GPL20301 ) encompassed gingival tissue samples from 12 periodontitis and 12 control (12 of the gingivitis samples were excluded). Furthermore, a total of 273 lactic acid metabolism‐related genes (LAMRGs) were obtained from the database (Supplementary Table 1 ). 2.2. Differential Expression Analysis and Identification of Candidate Genes To identify differentially expressed genes (DEGs) between periodontitis and control samples within the training set, the ‘limma’ package (v 3.54.0) [ 14 ] was used for differential expression analysis, and DEGs were screened based on p < 0.05 and |log 2 fold change (FC)| > 0.5. Subsequently, the ‘ggplot2’ package (v 3.4.1) [ 15 ] was utilized to construct a volcano plot for visualizing the DEGs, and the ‘pheatmap’ package (v 1.0.12) [ 16 ] was used to generate a heatmap of DEGs. To identify the DEGs linked to lactic acid metabolism in periodontitis, the ‘ggvenn’ package (v 0.1.9) [ 17 ] was used to intersect the DEGs and LAMRGs, and the intersection genes were used as candidate genes. 2.3. Identification and Validation of Key Genes In the training set, in light of the candidate genes obtained above, the candidate feature genes were first screened using the least absolute shrinkage and selection operator (LASSO) regression analysis with 10‐fold cross‐validation via the ‘glmnet’ package (v 4.1.7) [ 18 ]. The optimal penalty parameter lambda was selected based on the minimum binomial deviance criterion, and genes with non‐zero coefficients at the optimal lambda were considered candidate features. Furthermore, the candidate feature genes were also screened using the support vector machine recursive feature elimination (SVM‐RFE) of the ‘e1071’ package (v 1.7.13) [ 19 ]. A linear kernel function was applied, and features were recursively eliminated with a step size of 1 to obtain the optimal feature subset based on the highest accuracy across cross‐validation folds. Then the Venn diagram was constructed using the ‘ggvenn’ package (v 0.1.9), and the genes obtained by the 2 algorithms were crossed and used as feature genes. To observe whether the expression and trend of the feature genes were concordant in both the validation set and training set, the expression levels of the feature genes in the periodontitis and control samples were compared by Wilcoxon test in the 2 datasets ( p < 0.05). Subsequently, the genes whose expression levels differed in the different samples and that had a consistent trend in the 2 datasets were selected and named as the candidate key genes. To evaluate the ability of candidate key genes to discriminate periodontitis and control samples in different datasets, the ‘pROC’ package (v 1.18.0) [ 20 ] was utilized to plot the receiver operating characteristic (ROC) curves in the 2 datasets, and the area under the curve (AUC) was computed. Moreover, genes with an AUC value greater than 0.7 in both of the 2 datasets were named key genes for subsequent analyses. 2.4. Gene Set Enrichment Analysis (GSEA) To access the functional pathways related to key genes in periodontitis, GSEA of key genes was performed. The reference gene set ‘c2.cp.kegg.v7.0.symbols.gmt’ was carefully chosen from the MSigDB. Firstly, based on all samples in the training set, Spearman correlation analyses were performed between each key gene and all the remaining genes separately using the ‘psych’ package (v 2.4.1) [ 21 ] to obtain the correlation coefficients (cor). Subsequently, genes were arranged in descending order according to these coefficients. The sorted data were then utilized to conduct GSEA (|normalized enrichment score (NES)| > 1 and p < 0.05) using the implementation of the ‘clusterProfiler’ package (v 4.2.2) [ 22 ]. 2.5. Immune Infiltration Analysis To understand the immune infiltration of different samples in the training set, the CIBERSORT algorithm (v 0.1.0) [ 23 ] was utilized to understand the distribution of 22 distinct immune cells [ 24 ] in both periodontitis and control samples. Subsequently, the Wilcoxon test was harnessed to investigate the disparities in the infiltration levels of these 22 immune cells between periodontitis and control ( p < 0.05), and the outcomes were visualized using the ‘ggplot2’ package (v 3.4.1). Furthermore, Spearman correlation was carried out to investigate the connections between the differential immune cells and key genes (|cor| > 0.3 and p < 0.05) using the ‘cor’ package (v 1.18.3) [ 25 ]. 2.6. Construction of Molecular Regulatory Network To probe the intricacy and multiplicity of the regulation process of key genes' expression by constructing molecular regulatory network, the transcription factors (TFs) of key genes were forecasted using the JASPAR ( https://jaspar.elixir.no/ ). Moreover, the miRNAs of key genes were forecasted using the ENCORI database ( https://rnasysu.com/encori/ ). Finally, the TF‐mRNA and miRNA‐mRNA regulatory network was assembled by ‘Cytoscape’ software (v 3.8.2) [ 26 ]. 2.7. Analysis of RNA‐Binding Protein (RBP) Network and Protein Post‐Translational Modifications To understand the RBPs of key genes, the ENCORI database was also used for prediction, and the ‘Cytoscape’ software (v 3.8.2) was utilized to construct the regulatory network of key gene‐RBP. To investigate the types of post‐translational modifications and related sites of key genes at the protein level, PhosphoSitePlus ( https://www.phosphosite.org/homeAction.action ) was used for analysis. 2.8. Compound Prediction and Molecular Docking To probe the potential compounds for periodontitis and study potential associations between key genes and compounds, the Comparative Toxicogenomics Database (CTD, https://ctdbase.org/ ) was used to forecast compounds potentially connected with key genes, and ‘Cytoscape’ software (v 3.8.2) was used to render a visual representation of the compound‐key gene network. To further validate the binding ability of compounds to key genes, molecular docking of each key gene to the corresponding compound was performed. The compounds shared by the key genes were selected, and after excluding the toxic and carcinogenic substances, the remaining compounds were subjected to molecular docking with the key genes. Initially, the 3D structure of compounds was extracted from the PubChem database. Subsequently, the 3D structure of the protein corresponding to the key genes was extracted from the RCSB Protein Data Bank. Thereafter, the docking simulation results were visualized using ‘MOE’ software (v 2022.02) [ 27 ]. 2.9. Cell Culture Human periodontal ligament cells (PDLCs) were purchased from ScienCell Research Laboratories (Carlsbad, CA, USA) and cultured in DMEM medium (Gibco, USA, C11995500BT) supplemented with 10% foetal bovine serum (Gibco, USA, 10099141) and 1% penicillin–streptomycin. For the assessment of osteogenesis‐related indicators, osteogenic components (β‐glycerophosphate, ascorbic acid, and dexamethasone) were additionally added to the culture medium. 2.10. Detection of Inflammatory Markers For the detection of inflammation‐related markers, PDLCs were treated with DMEM medium containing 10 ng/mL TNF‐α and 10 ng/mL IL‐1β for 48 h based on references [ 28 , 29 ], followed by RT‐qPCR experiments to measure the mRNA levels of the inflammatory factors IL‐1β, IL‐6, IL‐8, and TNF‐α. Total RNA was extracted using an RNA extraction kit (Vazyme, Nanjing, China). RNA was reverse‐transcribed into cDNA using the HiScript III RT SuperMix for qPCR (+gDNA wiper) reverse transcription kit (Vazyme, Nanjing, China). Using the primer sequences as templates, amplification was performed with a real‐time quantitative PCR instrument. The primer sequences are listed in Supplementary Table 2 . The relative expression levels of the target genes were analyzed using the 2 ^(‐ΔΔCt) relative quantification method. 2.11. Detection of Osteogenesis‐Related Indicators After 7 days of osteogenic induction in PDLCs, the culture medium in the wells was removed, and the cell samples were washed 3–5 times with PBS for 3–5 min each time. Cells were stained using an Alkaline Phosphatase (ALP) Staining Kit (Beyotime, Shanghai, China). An appropriate amount of staining working solution was added, and the samples were incubated at room temperature protected from light for 30 min before images were captured. ALP activity was measured using an ALP Activity Assay Kit (Beyotime, Shanghai, China). Cells were lysed using cell lysis buffer, and ALP activity was calculated according to the kit instructions. mRNA levels of osteogenesis‐related markers were detected on day 14 of PDLCs culture, using the specific method described in section 2.10; primer sequences are listed in Supplementary Table 2 . After 21 days of osteogenic culture, calcium nodule formation was assessed in PDLCs. Specifically, cells were fixed with 4% paraformaldehyde for 15 min and stained with Alizarin Red S staining solution at room temperature for 30 min. Subsequently, 1 mL of 10% cetylpyridinium chloride was added to each well and fully dissolved at room temperature. The absorbance at 562 nm was measured using a microplate reader as a semi‐quantitative analysis result for calcium nodules. 2.12. In Vitro Validation of Key Genes and Their Protein Expression After incubating PDLCs in medium with or without 10 ng/mL TNF‐α and 10 ng/mL IL‐1β for 48 h, the expression of CFI and COQ2 was detected. The specific method was the same as that described in section 2.10, and the primer sequences are listed in Supplementary Table 2 . Total protein was extracted, and the protein concentration was determined according to the instructions of the BCA Protein Assay Kit (Beyotime, Shanghai, China). Based on the protein concentrations measured by the BCA assay, different volumes of protein were loaded for electrophoresis. After protein separation, proteins were transferred to a PVDF membrane, which was then placed in blocking solution and blocked on a shaker at room temperature for 10 min. Subsequently, the membrane was incubated overnight at 4°C with the following primary antibodies: CFI (Abmart, Shanghai, China) and COQ2 (Bioss, Beijing, China), both diluted at a ratio of 1:1000. The membrane was washed three times with TBST on a shaker for 10 min each time. Afterwards, it was incubated with a secondary antibody at room temperature for 1 h, followed by exposure using a chemiluminescence imaging system. Band intensities were analysed using Image J software. 2.13. Construction of Rat Periodontitis Model and Validation of Key Gene Expression All animal experiments were approved by the Animal Ethics Committee of Nanjing University (Approval No. IACUCD240258) and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Seven‐week‐old male Sprague–Dawley (SD) rats were selected and randomly divided into two groups: a control group and a ligation grouAfter one week of acclimatization feeding, the rats were anaesthetised. Under sterile conditions, the bilateral maxillary second molars of rats in the ligation group were ligated using sterilized 0.2 mm diameter orthodontic ligature wire. The control group had the wire immediately removed to exclude the influence of the ligation procedure. After 4 weeks of bilateral maxillary second molar ligation in rats, the animals were euthanized via cervical dislocation under excessive sodium pentobarbital anaesthesia, and maxillary bone samples were collected. Maxillary alveolar bone morphology and quality were scanned using micro‐CT (Bruker, Germany). Maxillary tomographic images were reconstructed using CTvox and DataViewer software. CTAn software was used to measure the distance from the cementoenamel junction (CEJ) to the alveolar crest at six periodontal sites per maxillary second molar. This distance, indicative of bone loss level, was statistically analysed to confirm the successful establishment of the rat periodontitis model. Immunohistochemistry was performed to detect the expression of CFI (Abmart, China) and COQ2 (Bioss, China) in the periodontal tissues surrounding the maxillary second molars. 2.14. Statistical Analysis Bioinformatics analyses were conducted using the R programming language (v 4.2.2). The differences between the 2 groups were assessed by Wilcoxon test, with a significance level of p < 0.05 being deemed statistically significant. 3. Results 3.1. Identification of Candidate Genes Related to Lactic Acid Metabolism There were 1605 DEGs between the periodontitis and control samples in the training set, including 956 up‐regulated genes and 649 down‐regulated genes (Figure 2A,B ). To further identify DEGs associated with lactic acid metabolism, 1605 DEGs and 273 LAMRGs were intersected, and 17 intersection genes were obtained as candidate genes (Figure 2C ). Overall, the comprehensive analysis had offered a wealth of information that could guide further research efforts aimed at understanding the underlying mechanisms of periodontitis and developing innovative therapeutic strategies. FIGURE 2. Open in a new tab Identification and screening of differentially expressed genes (DEGs). (A) Volcano plot depicting DEG distribution between periodontitis and control groups. Down regulated genes are shown in blue, not changed genes in grey, and up regulated genes in red. (B) Heatmap of gene expression patterns displaying the top 50 DEGs from the GSE16134 dataset. (C) Venn diagram illustrating the overlap between DEGs and lactate metabolism‐related genes (LANRGs). 3.2. A Total of 2 Key Genes That Impact Periodontitis Were Discovered Initially, based on the candidate genes, 9 candidate feature genes were acquired by LASSO regression analysis with the lambda .min value of 0.01528 (Figure 3A,B ). In addition, through the SVM‐RFE, a total of 12 candidate feature genes were obtained (Figure 3C,D ). Finally, 8 intersection genes of the 2 algorithms ( ACAT2 , CFI , COL4A1 , COQ2 , FLI1 , MLIP , PER2 , and PFKFB2 ) were obtained, and they were used as feature genes (Figure 3E ). FIGURE 3. Open in a new tab Machine learning‐based screening of candidate key genes CFI and COQ2 . (A‐B) Least absolute shrinkage and selection operator (LASSO) algorithm identified 9 candidate feature genes, (A) Coefficient profile plot of LASSO regression; (B) Cross‐validation error curve ( lambda selection). (C‐D) Support vector machine (SVM) analysis screened 12 candidate feature genes, (C) Error rate variation under 10‐fold cross‐validation; (D) Accuracy rate variation under 10‐fold cross‐validation. (E) Venn diagram showing 8 overlapping genes from both algorithms. (F) Expression analysis of feature genes in the training set. (G) Expression analysis of feature genes in the validation set. (H) Receiver operating characteristic (ROC) curves for CFI and COQ2 in the training set. (I) ROC curves for CFI and COQ2 in the validation set. Moreover, in the training set, these 8 feature genes manifested pronounced disparities in expression between periodontitis and control samples (Figure 3F ); while in the validation set, only the CFI and COQ2 genes were significantly different between the periodontitis and control samples, and their expression trend was consistent with the training set (Figure 3G ). Among them, CFI was upregulated in periodontitis and COQ2 was downregulated in periodontitis compared with control. Therefore, CFI and COQ2 were utilized as candidate key genes. Furthermore, to assess the diagnostic utility of candidate key genes in periodontitis, the ROC curve analysis was conducted. Interestingly, it was found that the AUC values of candidate key genes ( CFI and COQ2 ) were 0.86, 0.89 respectively (Figure 3H ). In the validation set, the AUC values of candidate key genes ( CFI and COQ2 ) were 0.77, 0.75 respectively (Figure 3I ). These outcomes showed that the diagnostic efficacy of CFI and COQ2 was satisfactory, and thus they could serve as key genes for subsequent analysis. 3.3. Key Genes Participated in the Regulation of Periodontitis Through Diverse Pathways and They Were Associated With Immune Cells GSEA results for the 2 key genes showed that CFI was enriched in 106 pathways, and the top 5 pathways were MAPK signalling pathway, regulation of actin cytoskeleton, pathways in cancer, cytokine‐cytokine receptor interaction, and focal adhesion (Figure 4A , Supplementary Table 3 ); COQ2 was enriched in 96 pathways, and the top 5 pathways were endocytosis, Huntingtons disease, Wnt signalling pathway, Alzheimer's disease, and ubiquitin mediated proteolysis (Figure 4B , Supplementary Table 4 ). FIGURE 4. Open in a new tab Functional profiling of key genes. (A) Gene Set Enrichment Analysis (GSEA) for CFI showing the top five enriched pathways ranked by normalized enrichment score (NES). (B) GSEA for COQ2 displaying the top five enriched pathways based on NES. (C) Heatmap of infiltration levels for 22 immune cell types in healthy controls versus periodontitis samples. (D) Box plot comparing immune cell distributions between control (blue) and periodontitis (red) groups across 22 cell subtypes. (E) Correlation matrix of 15 differentially infiltrated immune cells, with negative correlations in blue and positive correlations in red. * means p < 0.05, ** means p < 0.01, *** means p < 0.001 relative to the Control group. The proportional distribution of 22 immune cells in periodontitis and control samples was presented in Figure 4C . There were 15 differential immune cells, including activated NK cells, CD8 T cells, follicular helper T cells, M1 macrophages, M2 macrophages, etc. The outcomes revealed that compared with the control, the periodontitis samples had a higher proportion of plasma cells and a lower proportion of CD8 T cells ( p < 0.05) (Figure 4D ). Furthermore, CFI had the most notable positive relationship with neutrophils and the most notable negative relationship with resting dendritic cells (cor = 0.33, −0.48, p < 0.001), while COQ2 exhibited the most notable positive relationship with resting dendritic cells and the most notable negative relationship with plasma cells (cor = 0.57, −0.59, p < 0.001) (Figure 4E ). The above results showed that key genes might affect the immune microenvironment of periodontitis patients, which could offer a valuable reference for the clinical decision‐making in the periodontitis treatment. 3.4. Multiple Molecules and Compounds Might Be Related to the Key Genes in Periodontitis The results of the transcription factor regulatory network showed that a total of 9 TFs regulated the expression of key genes. Among them, 8 TFs jointly regulated the expression of both CFI and COQ2 , such as TFAP2A, NFIL3, and AR (Figure 5A ). A total of 17 miRNAs targeting CFI were predicted, such as hsa‐miR‐6759‐3p and hsa‐miR‐3679‐3p; 45 miRNAs targeting COQ2 were predicted, such as hsa‐miR‐302b‐5p and hsa‐miR‐621 (Figure 5B,C ). A total of 1 RBP interacting with CFI was identified, namely PCBP2. For COQ2 , 6 interacting RBPs were found, including ALYREF, DDX3X, ELAVL1, RBMX, TARDBP, and YTHDF1. (Figure 5D ). FIGURE 5. Open in a new tab Molecular regulatory networks of key genes. (A) mRNA‐transcription factor (TF) regulatory network for CFI and COQ2 . Red nodes denote the two key genes; blue nodes represent TFs. (B) Predicted mRNA‐miRNA interaction network for CFI . (C) Predicted mRNA‐miRNA interaction network for COQ2 . Red nodes indicate key genes; blue nodes denote miRNAs. (D) mRNA‐RNA‐binding protein (RBP) interaction network. Red nodes mark key genes; purple nodes signify RBPs. The findings regarding the post‐translational modifications of the proteins encoded by the 2 key genes indicated that CFI harboured 7 phosphorylation sites, 1 acetylation site, and 1 ubiquitination site (Figure 6A ). COQ2 was found to have 1 phosphorylation site and 1 ubiquitination site (Figure 6B ). These results not only facilitated a more in‐depth comprehension of the post‐translational modification profiles of the key gene‐derived proteins but also offered valuable references for the treatment of periodontitis. FIGURE 6. Open in a new tab Post‐translational modification (PTM) analysis of key genes. (A) PTM types and associated residues of the CFI protein. (B) PTM types and associated residues of the COQ2 protein. Based on the CTD, the potential compounds of key genes were predicted. Among them, 120 compounds were predicted for CFI . In total, 55 compounds were forecasted for COQ2 (Figure 7 ). A total of 20 common compounds were predicted for the 2 key genes, including abrine, acetaminophen, and amiodarone. After excluding the toxic and carcinogenic substances, a total of 4 common compounds were obtained, including acetaminophen, amiodarone, genistein, and pirinixic acid. The molecular docking outcomes demonstrated that the binding affinities of CFI and COQ2 with amiodarone were −6.71 and −7.27 (Table 1 ), respectively. In addition, the binding energies of genistein and pirinixic acid with COQ2 were −5.92 and −6.85, respectively, which were all less than −5 kcal/mol, signifying that the key genes possessed strong affinity with compounds. The analysis of binding conformation revealed that the compound infiltrated into the binding site of each key gene, and there were abundant hydrogen bond donors and acceptors around the binding sites (Figure 8A,B , Supplementary Figure 1 ). The association of key genes with other molecules and compounds could potentially offer a valuable reference for the clinical management of periodontitis. FIGURE 7. Open in a new tab Drug‐gene targeting network. Red nodes represent key genes; purple nodes indicate predicted drugs. TABLE 1. Molecular docking results of CFI and COQ2. protein PDB ID/AFold ID intersection chem Compound CID S CFI 2XRC Acetaminophen 1983 / Amiodarone 2157 −6.713953 Genistein 5,280,961 / pirinixic acid 5694 / COQ2 AF‐ Q96H96 ‐F1 Acetaminophen 1983 −4.84126234 Amiodarone 2157 −7.27498102 Genistein 5,280,961 −5.91755962 pirinixic acid 5694 −6.85080624 Open in a new tab FIGURE 8. Open in a new tab Molecular docking of key genes with drug candidates. (A) Docking conformations between COQ2 and small‐molecule ligands: Acetaminophen, pirinixic acid, amiodarone, and genistein. (B) Binding mode of CFI with Amiodarone. 3.5. In Vivo and in Vitro Validation of Key Gene Expression Changes Given the observed changes in the expression of the CFI and COQ2 genes within the dataset, and based on existing literature, we utilized a cytokine mixture of TNF‐α and IL‐1β to co‐culture with Human Periodontal Ligament Cells (PDLCs) to establish an in vitro model of periodontitis. This approach aimed to detect alterations in both the gene and protein levels of CFI and COQ2 at the cellular level. After treating PDLCs with TNF‐α and IL‐1β, we first measured the mRNA levels of several hallmark inflammatory factors. The mRNA levels of these inflammatory factors were significantly increased (Figure 9A ). Osteogenesis‐related indicators, including Alkaline Phosphatase (ALP) activity, the quantity of calcium nodule formation, and the mRNA levels of osteogenesis‐related factors, were all significantly decreased. These changes are consistent with the characteristics of periodontitis and confirmed the successful establishment of the model (Figure 9B–D ). FIGURE 9. Open in a new tab Changes in the mRNA and protein expression of CFI and COQ2 were confirmed in the in vitro periodontal inflammation model. (A) mRNA levels of inflammation‐related cytokines (IL‐1β, IL‐6, IL‐8, TNF‐α) under inflammatory conditions. (B) Changes in Alkaline Phosphatase (ALP) activity in the periodontal inflammation model. The left panel shows representative images of ALP staining, and the right panel shows the quantitative results of ALP activity. (C) Calcium nodule formation in the periodontal inflammation model. The left panel shows representative images of Alizarin Red S (ARS) staining, and the right panel shows the semi‐quantitative analysis of ARS results. (D) mRNA expression levels of osteogenesis‐related factors ( RUNX‐2 , OCN , OPN ). (E) mRNA expression levels of the key genes CFI and COQ2 . (F) Protein expression levels of CFI and COQ2. * means p < 0.05, ** means p < 0.01, *** means p < 0.001 relative to the Control group. RT‐qPCR results demonstrated that PDLCs in the inflammatory environment exhibited increased CFI expression and decreased COQ2 expression (Figure 9E ). Detection at the protein level similarly confirmed these findings. Under inflammatory conditions, the expression level of CFI was elevated, while the expression of COQ2 was reduced (Figure 9F ). These discoveries verify that the mRNA and protein expression levels of CFI and COQ2 in the periodontal inflammatory environment are consistent with the trends observed in the dataset. To further validate the diagnostic value of key genes CFI and COQ2 , we established a rat model of experimental periodontitis using the ligature‐induced method. Orthodontic ligature wires were placed around the maxillary second molars to induce periodontal inflammation. Gene expression was verified at both tissue and molecular levels. Micro‐CT analysis was first performed. Sagittal sectional and 3D‐reconstructed images revealed alveolar bone conditions, demonstrating reduced bone height surrounding the teeth, confirming successful periodontitis induction (Figure 10A ). The cementoenamel junction‐to‐alveolar bone crest (CEJ‐ABC) distance was measured at six sites of the second molars: mesiopalatal (MP), distopalatal (DP), mesial (M), distal (D), mesiobuccal (MB), and distobuccal (DB). Significant increases in CEJ‐ABC distances in the periodontitis group validated alveolar bone loss in the rat model (Figure 10B ). Immunohistochemical staining showed elevated expression of CFI (Figure 10C ) and reduced expression of COQ2 (Figure 10D ) in periodontitis‐induced rats compared to healthy controls. FIGURE 10. Open in a new tab In vivo validation of key gene expression alterations in a rat ligature‐induced periodontitis model. (A) Micro‐CT analysis depicting sagittal sections and 3D reconstructions of maxillary second molars. (B) Cementoenamel junction‐to‐alveolar bone crest (CEJ‐ABC) distances at six periodontal sites: Mesiopalatal (MP), distopalatal (DP), mesial (M), distal (D), mesiobuccal (MB), and distobuccal (DB). (C) Immunohistochemical staining revealed increased CFI expression in periodontitis tissues. (D) Decreased COQ2 expression was observed in periodontitis tissues by immunohistochemistry. ** means p < 0.01, *** means p < 0.001 relative to the Control group. 4. Discussion Periodontitis is a chronic inflammatory disease associated with systemic disorders. Recent studies suggest potential links between metabolic dysregulation and periodontitis [ 30 ]. Lactate metabolism plays a complex, dual role in the pathogenesis of periodontitis. This study integrated multi‐level bioinformatics analyses and machine learning to identify CFI and COQ2 as key genes bridging lactate metabolism and periodontitis progression. These findings expand the understanding of periodontal pathogenesis and strengthen the connection between periodontitis and metabolic research. Compound‐binding predictions for differential genes reveal potential therapeutic agents, contributing to the advancement of precision medicine in periodontitis treatment. The data sources for this study include two datasets: the training set GSE16134 and the validation set GSE173078 , used for differential expression analysis and external validation. The training set GSE16134 contains 241 affected gingival tissue samples from periodontitis patients and 69 unaffected gingival tissue samples from the same patients (control samples from healthy gingival tissues), significantly reducing inter‐individual heterogeneity and improving the accuracy of genetic detection. Differential expression analysis was performed on the training set samples. Genes related to lactate metabolism were intersected from the obtained DEGs. Subsequently, two independent machine learning methods, LASSO and SVM‐RFE models, were used to further screen the identified feature genes. Combined with the validation set GSE173078 , external validation of candidate key genes was conducted, yielding the key genes CFI and COQ2 . CFI and COQ2 exhibited AUC > 0.7 in both datasets, indicating their high predictive ability for periodontitis. The CFI gene encodes complement factor I, a serine protease that cleaves and inactivates complement components C4b and C3b, thereby preventing the assembly of C3 and C5 convertases, which is critical for regulating the complement cascade. Recent studies have shown elevated expression of C3, C3b, and C4b in gingival tissues of periodontitis patients, with their levels positively correlated to disease severity. Inhibitors of C3b/C4b have been shown to mitigate alveolar bone loss in periodontitis [ 31 , 32 ]. This suggests that the upregulation of CFI observed in periodontitis groups in this study may reflect a host defence response, where CFI likely modulates C3b and C4b to influence disease progression. In terms of lactate metabolism, lactate activates the GPR81 receptor, promoting M2 macrophage polarization, which may indirectly affect the CFI‐associated complement system. The COQ2 gene encodes 4‐hydroxybenzoate polyprenyltransferase, an enzyme that catalyses the second critical step in coenzyme Q (CoQ) biosynthesis. CoQ is essential for mitochondrial energy metabolism and antioxidant defence [ 33 , 34 ]. The role of COQ2 in periodontitis may involve mitochondrial dysfunction, oxidative stress, and inflammatory regulation. In this study, COQ2 was found to be downregulated in periodontitis groups, leading to reduced mitochondrial oxidative phosphorylation efficiency, increased mitochondrial reactive oxygen species (ROS), and impaired oxidation of glycolytic NADH via the electron transport chain, further promoting lactate accumulation [ 35 ]. Notably, recent periodontitis studies using machine learning have identified COQ2 as a key gene and explored its mechanistic role in the disease [ 34 ]. Although CFI and COQ2 were identified as key lactate metabolism‐related genes in this study, no significant enrichment of metabolism‐related pathways was observed in the GSEA results (Figure 4 ). This discrepancy may be explained by several factors. First, GSEA evaluates pathway enrichment based on the overall expression profile of all genes, whereas CFI and COQ2 were identified through machine learning algorithms that focus on individual predictive power rather than global pathway activity [ 36 , 37 ]. Second, lactate metabolism involves complex, multi‐step enzymatic reactions and regulatory networks that may not be fully captured by conventional pathway databases such as KEGG or GO. Third, the functional roles of CFI and COQ2 in periodontitis may extend beyond metabolism, involving immune regulation and complement cascades, as supported by studies on immunometabolism [ 38 , 39 ] and the pleiotropic nature of metabolic genes [ 40 ]. Therefore, the lack of metabolic pathway enrichment in GSEA does not contradict the identification of these genes as lactate metabolism‐related, but rather reflects the multifaceted nature of their biological functions. To further explore the functional roles of CFI and COQ2 , we analysed immune cell infiltration in periodontitis. Our results revealed significant alterations in 15 types of immune cells between periodontitis and control samples, consistent with the inflammatory cascade characteristic of this disease. Notably, CFI showed positive correlations with neutrophils and plasma cells, while COQ2 exhibited strong negative correlations with plasma cells, aligning with the activation of neutrophils and plasma cells in periodontitis samples, reflecting immune responses to pathogenic bacterial infection. Additionally, the expression of reparative M2 macrophages, regulatory T cells, and resting dendritic cells decreased in periodontitis groups, suggesting impaired tissue repair. CFI demonstrated negative correlations with M2 macrophages, regulatory T cells, and resting dendritic cells, whereas COQ2 showed positive correlations with M2 macrophages and resting dendritic cells, consistent with these observations. We further constructed a molecular regulatory network based on the functional roles of these key genes. Future studies may investigate how CFI and COQ2 regulate the immune microenvironment in periodontitis and develop novel immunomodulatory therapies targeting these genes. Based on these findings, we propose that CFI and COQ2 could serve as potential targets for immunomodulatory therapy in periodontitis. Given that CFI is upregulated in periodontitis, strategies aimed at modulating its activity—rather than simply enhancing it—may help restore complement balance and limit excessive inflammation. In contrast, since COQ2 is downregulated, upregulating its expression could improve mitochondrial function and reduce oxidative stress in periodontal tissues. Future studies should explore the therapeutic potential of modulating these genes using small molecules, gene editing, or biologics in preclinical periodontitis models. Amiodarone, primarily used as an antiarrhythmic drug, has not been extensively studied or validated for its potential role in periodontitis treatment. Recent studies indicate that amiodarone inhibits the NF‐κB inflammatory signalling pathway [ 41 ] and synergizes with antibiotics to combat drug‐resistant bacteria and specific pathogens [ 42 ], suggesting its potential as an adjunct therapy for periodontitis. In this study, predictions of potential compounds targeting the key genes revealed strong binding affinity between amiodarone and CFI/COQ2, highlighting its therapeutic prospects. However, it should be noted that these findings are predictive in nature and require further experimental validation. Regarding lactate metabolism, amiodarone has been shown to elevate lactate levels in hepatocytes, likely due to its inhibition of mitochondrial electron transport system complexes I (CI) and II (CII) [ 35 ], which indirectly enhances anaerobic glycolysis and lactate accumulation. However, systemic side effects of amiodarone, such as pulmonary toxicity and cardiotoxicity [ 43 ], limit its direct application in periodontitis therapy. Future studies should focus on validating the anti‐inflammatory and osteoprotective effects of amiodarone in cellular and animal models of periodontitis, as well as developing localized delivery systems or structural derivatives to mitigate toxicity while retaining therapeutic benefits. In vitro experiments further confirmed that the expression patterns of CFI and COQ2 were consistent with our bioinformatics predictions. In the in vivo rat model of ligature‐induced periodontitis, immunohistochemical staining revealed that CFI expression was elevated and COQ2 expression was reduced in periodontal tissues compared to healthy controls. Notably, both CFI and COQ2 were widely expressed in multiple cell types within periodontal tissues. This finding does not contradict our correlation analysis, as bulk RNA‐seq reflects average expression across all cells and its association with immune cell proportions, rather than indicating cell type‐specific expression. Although these genes are broadly expressed in structural cells, their expression levels in specific immune cells may be dynamically regulated during periodontitis and contribute to the observed correlations. Future studies using single‐cell technologies are needed to further validate these findings. In conclusion, this study revealed the correlation between periodontitis and lactate metabolism through bioinformatics analysis, identifying two key genes ( CFI and COQ2 ) associated with lactate metabolism in periodontitis. Investigations into their mechanistic roles demonstrated that these genes may participate in periodontitis pathogenesis through multiple pathways and interact with immune cells, providing potential therapeutic guidance for patients. To our knowledge, this is the first bioinformatics study exploring the relationship between periodontitis and lactate‐related genes. CFI and COQ2 represent the most significant crosstalk genes linking periodontitis and lactate metabolism, serving as potential diagnostic biomarkers. Overall, we propose a hypothesis that lactate‐targeted therapy or modulation of lactate‐associated biomarkers could emerge as novel strategies for periodontitis treatment. This study has several limitations. First, while differential gene screening was performed using bioinformatics and machine learning on periodontitis versus healthy periodontal tissues, cell type‐specific gene expression variations within heterogeneous periodontal tissues were not analysed. Distinct cell populations may exhibit divergent gene expression patterns. Future single‐cell sequencing to clarify gene expression dynamics or investigations into epigenetic regulation (e.g., methylation, non‐coding RNAs) of these key genes could refine current conclusions. Second, although we validated the mRNA and protein expression of CFI and COQ2 in vitro, the in vivo validation was primarily based on semi‐quantitative immunohistochemistry in rat periodontal tissues. Therefore, these findings should be interpreted with caution and require further confirmation in human clinical samples using more quantitative methods such as RT‐qPCR or Western blot. Additionally, the datasets used may lack sufficient statistical power to generate robust evidence. Larger‐scale patient cohorts are required for external validation. Ultimately, experimental validation through cellular or animal models will be critical to confirm the biological relevance and accuracy of these findings. Author Contributions Bijun Zhu: conceptualization, formal analysis, methodology, writing – original draft, data curation, software, visualization. Shi Xu: writing – review and editing, formal analysis, visualization, validation. Mingyue Cheng: writing – review and editing, software. Leiying Miao: writing – review and editing, resources, project administration, conceptualization, funding acquisition, supervision. Funding This work was supported by the National Natural Science Foundation of China grant (22104054), Key Project supported by Medical Science and Technology Development Foundation, Nanjing Department of Health (ZKX24056), the Jiangsu Provincial Cadre Healthcare Program (BJ24041), and High‐Level Hospital Construction Project of Nanjing Stomatological Hospital, Affiliated Hospital of Medical School, Institute of Stomatology, Nanjing University (0024C006 and 0224C051). Consent The authors have nothing to report. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Figure S1: Molecular docking pose and interaction details of the candidate drug with the key genes. (A) The predicted binding pattern of the candidate drug to COQ2. (B) The predicted binding pattern of the candidate drug to CFI. JCMM-30-e71141-s001.docx (448.6KB, docx) Table S1: The Lactic acid metabolism‐related genes. JCMM-30-e71141-s003.xlsx (13KB, xlsx) Table S2: The sequences of primes. JCMM-30-e71141-s005.docx (14.6KB, docx) Table S3: Gene set enrichment analysis for CFI. JCMM-30-e71141-s004.xlsx (33.7KB, xlsx) Table S4: Gene set enrichment analysis for COQ2. JCMM-30-e71141-s002.xlsx (32.1KB, xlsx) Acknowledgements The authors sincerely thank related contributors for uploading their datasets and acknowledge GEO database for providing their platforms. Data Availability Statement The data that support the findings of this study are openly available in Gene Expression Omnibus (GEO) at https://www.ncbi.nlm.nih.gov/geo/ , reference number GSE16134 , GSE173078 . References 1. Bernabe E. and et al , “Global, Regional, and National Levels and Trends in Burden of Oral Conditions From 1990 to 2017: A Systematic Analysis for the Global Burden of Disease 2017 Study,” Journal of Dental Research 99, no. 4 (2020): 362–373. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 2. Sanz M., Marco del Castillo A., Jepsen S., et al., “Periodontitis and Cardiovascular Diseases: Consensus Report,” Journal of Clinical Periodontology 47, no. 3 (2020): 268–288. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Simpson T. C., “Treatment of Periodontitis for Glycaemic Control in People With Diabetes Mellitus,” Cochrane Database of Systematic Reviews 4, no. 4 (2022): Cd004714. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 4. Tonetti M. S., Greenwell H., and Kornman K. S., “Staging and Grading of Periodontitis: Framework and Proposal of a New Classification and Case Definition,” Journal of Periodontology 89, no. 1 (2018): S159–s172. [ DOI ] [ PubMed ] [ Google Scholar ] 5. Wang W., Wang X., Lu S., et al., “Metabolic Disturbance and Th17/Treg Imbalance Are Associated With Progression of Gingivitis,” Frontiers in Immunology 12 (2021): 670178. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 6. Liberti M. V. and Locasale J. W., “The Warburg Effect: How Does It Benefit Cancer Cells?,” Trends in Biochemical Sciences 41, no. 3 (2016): 211–218. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 7. Zhang D., Tang Z., Huang H., et al., “Metabolic Regulation of Gene Expression by Histone Lactylation,” Nature 574, no. 7779 (2019): 575–580. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. Feng T., Zhao X., Gu P., et al., “Adipocyte‐Derived Lactate Is a Signalling Metabolite That Potentiates Adipose Macrophage Inflammation via Targeting PHD2,” Nature Communications 13, no. 1 (2022): 5208. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 9. Liu C., Wu J., Zhu J., et al., “Lactate Inhibits Lipolysis in Fat Cells Through Activation of an Orphan G‐Protein‐Coupled Receptor, GPR81,” Journal of Biological Chemistry 284, no. 5 (2009): 2811–2822. [ DOI ] [ PubMed ] [ Google Scholar ] 10. Brown T. P. and Ganapathy V., “Lactate/GPR81 Signaling and Proton Motive Force in Cancer: Role in Angiogenesis, Immune Escape, Nutrition, and Warburg Phenomenon,” Pharmacology & Therapeutics 206 (2020): 107451. [ DOI ] [ PubMed ] [ Google Scholar ] 11. Hoque R., Farooq A., Ghani A., Gorelick F., and Mehal W. Z., “Lactate Reduces Liver and Pancreatic Injury in Toll‐Like Receptor‐ and Inflammasome‐Mediated Inflammation via GPR81‐Mediated Suppression of Innate Immunity,” Gastroenterology 146, no. 7 (2014): 1763–1774. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 12. Liu X., Wang J., Lao M., et al., “Study on the Effect of Protein Lysine Lactylation Modification in Macrophages on Inhibiting Periodontitis in Rats,” Journal of Periodontology 95, no. 1 (2024): 50–63. [ DOI ] [ PubMed ] [ Google Scholar ] 13. Wu Y., Wang X., Zhang Y., et al., “Proanthocyanidins Ameliorate LPS‐Inhibited Osteogenesis of PDLSCs by Restoring Lysine Lactylation,” International Journal of Molecular Sciences 25, no. 5 (2024): 2947. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 14. Ritchie M. E., Phipson B., Wu D., et al., “Limma Powers Differential Expression Analyses for RNA‐Sequencing and Microarray Studies,” Nucleic Acids Research 43, no. 7 (2015): e47. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 15. Gustavsson E. K., Zhang D., Reynolds R. H., Garcia‐Ruiz S., and Ryten M., “Ggtranscript: An R Package for the Visualization and Interpretation of Transcript Isoforms Using ggplot2,” Bioinformatics 38, no. 15 (2022): 3844–3846. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 16. Wang Y., Zhuang H., Jiang X. H., Zou R. H., Wang H. Y., and Fan Z. N., “Unveiling the Key Genes, Environmental Toxins, and Drug Exposures in Modulating the Severity of Ulcerative Colitis: A Comprehensive Analysis,” Frontiers in Immunology 14 (2023): 1162458. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 17. Chen H. and Boutros P. C., “VennDiagram: A Package for the Generation of Highly‐Customizable Venn and Euler Diagrams in R,” BMC Bioinformatics 12 (2011): 35. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 18. Li Y., Lu F., and Yin Y., “Applying Logistic LASSO Regression for the Diagnosis of Atypical Crohn's Disease,” Scientific Reports 12, no. 1 (2022): 11340. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Long G. and Yang C., “A Six‐Gene Support Vector Machine Classifier Contributes to the Diagnosis of Pediatric Septic Shock,” Molecular Medicine Reports 21, no. 3 (2020): 1561–1571. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Robin X., Turck N., Hainard A., et al., “pROC: An Open‐Source Package for R and S+ to Analyze and Compare ROC Curves,” BMC Bioinformatics 12 (2011): 77. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 21. Robles‐Jimenez L. E.. “Worldwide Traceability of Antibiotic Residues From Livestock in Wastewater and Soil: A Systematic Review,” Animals (Basel) 12, no. 1 (2021): 60. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 22. Wu T., Hu E., Xu S., et al., “clusterProfiler 4.0: A Universal Enrichment Tool for Interpreting Omics Data,” Innovation (Camb) 2, no. 3 (2021): 100141. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 23. Craven K. E., Gökmen‐Polar Y., and Badve S. S., “CIBERSORT Analysis of TCGA and METABRIC Identifies Subgroups With Better Outcomes in Triple Negative Breast Cancer,” Scientific Reports 11, no. 1 (2021): 4691. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Liu W., Han F., Wan M., and Yang X. Z., “Integrated Bioinformatics Analysis Identifies Shared Immune Changes Between Ischemic Stroke and COVID 19,” Frontiers in Immunology 14 (2023): 1102281. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Guo G., Sun Y., Qian G., and Wang Q., “LIC Criterion for Optimal Subset Selection in Distributed Interval Estimation,” Journal of Applied Statistics 50, no. 9 (2023): 1900–1920. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 26. Shannon P., Markiel A., Ozier O., et al., “Cytoscape: A Software Environment for Integrated Models of Biomolecular Interaction Networks,” Genome Research 13, no. 11 (2003): 2498–2504. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 27. Riyaphan J., Pham D. C., Leong M. K., and Weng C. F., “In Silico Approaches to Identify Polyphenol Compounds as α‐Glucosidase and α‐Amylase Inhibitors Against Type‐II Diabetes,” Biomolecules 11, no. 12 (2021): 1877. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Lin L., Yu W., Yang R., et al., “Ubiquitin C‐Terminal Hydrolase L1 Promoted Pro‐Angiogenic Capacity of Periodontal Ligament Stem Cells via HIF‐1α/YAP Signaling in Periodontitis,” Stem Cell Res Ther 16, no. 1 (2025): 271. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Zheng X. Y., Mao C. Y., Qiao H., et al., “Plumbagin Suppresses Chronic Periodontitis in Rats via Down‐Regulation of TNF‐α, IL‐1β and IL‐6 Expression,” Acta Pharmacologica Sinica 38, no. 8 (2017): 1150–1160. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 30. Bullon P., Morillo J. M., Ramirez‐Tortosa M. C., Quiles J. L., Newman H. N., and Battino M., “Metabolic Syndrome and Periodontitis: Is Oxidative Stress a Common Link?,” Journal of Dental Research 88, no. 6 (2009): 503–518. [ DOI ] [ PubMed ] [ Google Scholar ] 31. Huang R. Y., Tseng F. Y., You J. J., et al., “Targeting Therapeutic Agent Against C3b/C4b, SB002, on the Inflammation‐Induced Bone Loss in Experimental Periodontitis,” Journal of Clinical Periodontology 50, no. 5 (2023): 657–670. [ DOI ] [ PubMed ] [ Google Scholar ] 32. Huang R. Y., Tseng F. Y., Cheng C. D., et al., “Complement Components C3b and C4b as Potential Reliable Site‐Specific Diagnostic Biomarkers for Periodontitis,” Journal of Periodontal Research 58, no. 5 (2023): 1020–1030. [ DOI ] [ PubMed ] [ Google Scholar ] 33. Desbats M. A., “The COQ2 Genotype Predicts the Severity of Coenzyme Q10 Deficiency,” Human Molecular Genetics 25, no. 19 (2016): 4256–4265. [ DOI ] [ PubMed ] [ Google Scholar ] 34. Kuo M. C., Lu Y. C., Tai C. H., et al., “COQ2 and SNCA Polymorphisms Interact With Environmental Factors to Modulate the Risk of Multiple System Atrophy and Subtype Disposition,” European Journal of Neurology 29, no. 10 (2022): 2956–2966. [ DOI ] [ PubMed ] [ Google Scholar ] 35. Bețiu A. M., Chamkha I., Gustafsson E., et al., “Cell‐Permeable Succinate Rescues Mitochondrial Respiration in Cellular Models of Amiodarone Toxicity,” International Journal of Molecular Sciences 22, no. 21 (2021): 11786. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 36. Subramanian A., Tamayo P., Mootha V. K., et al., “Gene Set Enrichment Analysis: A Knowledge‐Based Approach for Interpreting Genome‐Wide Expression Profiles,” Proceedings of the National Academy of Sciences of the United States of America 102, no. 43 (2005): 15545–15550. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 37. Saeys Y., Inza I., and Larrañaga P., “A Review of Feature Selection Techniques in Bioinformatics,” Bioinformatics 23, no. 19 (2007): 2507–2517. [ DOI ] [ PubMed ] [ Google Scholar ] 38. O'Neill L. A., Kishton R. J., and Rathmell J., “A guide to immunometabolism for immunologists,” Nature Reviews Immunology 16, no. 9 (2016): 553–565. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 39. Makowski L., Chaib M., and Rathmell J. C., “Immunometabolism: From Basic Mechanisms to Translation,” Immunological Reviews 295, no. 1 (2020): 5–14. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 40. Hotamisligil G. S., “Foundations of Immunometabolism and Implications for Metabolic Health and Disease,” Immunity 47, no. 3 (2017): 406–420. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 41. Yokota Y., Takaki K., Baba K., et al., “Amiodarone Inhibits the Toll‐Like Receptor 3‐Mediated Nuclear Factor κB Signaling Pathway by Blocking Organelle Acidification,” Biochemical and Biophysical Research Communications 708 (2024): 149801. [ DOI ] [ PubMed ] [ Google Scholar ] 42. Ahmad I. and Husain F., “0715 Bacterial Virulence, Biofilm and Quorum Sensing as Promising Targets for Anti‐pathogenic Drug Discovery and the Role of Natural Products,” New Look to Phytomedicine: Advancements in Herbal Products as Novel Drug Leads (2014), 43. [ Google Scholar ] 43. Feduska E. T., Thoma B. N., Torjman M. C., and Goldhammer J. E., “Acute Amiodarone Pulmonary Toxicity,” Journal of Cardiothoracic and Vascular Anesthesia 35, no. 5 (2021): 1485–1494. [ DOI ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Figure S1: Molecular docking pose and interaction details of the candidate drug with the key genes. (A) The predicted binding pattern of the candidate drug to COQ2. (B) The predicted binding pattern of the candidate drug to CFI. JCMM-30-e71141-s001.docx (448.6KB, docx) Table S1: The Lactic acid metabolism‐related genes. JCMM-30-e71141-s003.xlsx (13KB, xlsx) Table S2: The sequences of primes. JCMM-30-e71141-s005.docx (14.6KB, docx) Table S3: Gene set enrichment analysis for CFI. JCMM-30-e71141-s004.xlsx (33.7KB, xlsx) Table S4: Gene set enrichment analysis for COQ2. JCMM-30-e71141-s002.xlsx (32.1KB, xlsx) Data Availability Statement The data that support the findings of this study are openly available in Gene Expression Omnibus (GEO) at https://www.ncbi.nlm.nih.gov/geo/ , reference number GSE16134 , GSE173078 . Articles from Journal of Cellular and Molecular Medicine are provided here courtesy of Blackwell Publishing ACTIONS View on publisher site PDF (9.7 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Microbiome and Proteome Mapping in Periodontitis#870175
• Levilactobacillus Brevis Oral Health#1111640
• Diagnosis of Periodontal Disease Using Genetic Marker in Saliva#854430
• Clinical Study of the Presence of Trichomonas Tenax and Entamoeba Gingivalis in Patients With Periodontitis#915038
• Pan-tissue transcriptomic profiling of dairy cattle.#280985