Table_4_Screening of crosstalk and pyroptosis-related genes linking periodontitis and osteoporosis based on bioinformatics and machine learning.xlsx

Background and objectiveThis study aimed to identify crosstalk genes between periodontitis (PD) and osteoporosis (OP) and potential relationships between crosstalk and pyroptosis-related genes.
MethodsPD and OP datasets were downloaded from the GEO database and were performed differential expression analysis to obtain DEGs. Overlapping DEGs got crosstalk genes linking PD and OP. Pyroptosis-related genes were obtained from literature reviews. Pearson coefficients were used to calculate crosstalk and pyroptosis-related gene correlations in the PD and OP datasets. Paired genes were obtained from the intersection of correlated genes in PD and OP. PINA and STRING databases were used to conduct the crosstalk-bridge-pyroptosis genes PPI network. The clusters in which crosstalk and pyroptosis-related genes were mainly concentrated were defined as key clusters. The key clusters’ hub genes and the included paired genes were identified as key crosstalk-pyroptosis genes. Using ROC curve analysis and XGBoost screened key genes. PPI subnetwork, gene–biological process and gene-pathway networks were constructed based on key genes. In addition, immune infiltration was analyzed on the PD dataset using the CIBERSORT algorithm.
ResultsA total of 69 crosstalk genes were obtained. 13 paired genes and hub genes TNF and EGFR in the key clusters (cluster2, cluster8) were identified as key crosstalk-pyroptosis genes. ROC and XGBoost showed that PRKCB, GSDMD, ARMCX3, and CASP3 were more accurate in predicting disease than other key crosstalk-pyroptosis genes while better classifying properties as a whole. KEGG analysis showed that PRKCB, GSDMD, ARMCX3, and CASP3 were involved in neutrophil extracellular trap formation and MAPK signaling pathway pathways. Immune infiltration results showed that all four key genes positively correlated with plasma cells and negatively correlated with T cells follicular helper, macrophages M2, and DCs.
ConclusionThis study shows a joint mechanism between PD and OP through crosstalk and pyroptosis-related genes. The key genes PRKCB, GSDMD, ARMCX3, and CASP3 are involved in the neutrophil extracellular trap formation and MAPK signaling pathway, affecting both diseases. These findings may point the way to future research.

研究背景与目的：本研究旨在明确牙周炎（periodontitis, PD）与骨质疏松症（osteoporosis, OP）之间的串扰基因，以及串扰基因与焦亡（pyroptosis）相关基因间的潜在关联。
研究方法：本研究从GEO数据库下载牙周炎与骨质疏松症的数据集，进行差异表达分析以获得差异表达基因（differentially expressed genes, DEGs）。取两组数据集的重叠差异表达基因，作为连接牙周炎与骨质疏松症的串扰基因。焦亡相关基因通过文献调研获取。采用Pearson相关系数计算牙周炎、骨质疏松症数据集中串扰基因与焦亡相关基因的表达相关性，取两个数据集中相关性基因的交集作为配对基因。借助PINA与STRING数据库构建串扰-桥接-焦亡基因的蛋白质相互作用（protein-protein interaction, PPI）网络。将串扰基因与焦亡相关基因主要富集的模块定义为关键模块。提取关键模块中的核心基因（hub genes）及其中包含的配对基因，作为关键串扰-焦亡基因。采用ROC曲线分析与XGBoost算法筛选关键基因。基于关键基因构建PPI子网、基因-生物学过程网络及基因-通路网络。此外，使用CIBERSORT算法对牙周炎数据集进行免疫浸润分析。
研究结果：本研究共获得69个串扰基因。在关键模块（cluster2、cluster8）中，13个配对基因以及核心基因肿瘤坏死因子（tumor necrosis factor, TNF）与表皮生长因子受体（epidermal growth factor receptor, EGFR）被鉴定为关键串扰-焦亡基因。ROC曲线与XGBoost分析显示，PRKCB、GSDMD、ARMCX3及CASP3相较于其他关键串扰-焦亡基因，对疾病的预测准确性更高，且整体分类性能更优。京都基因与基因组百科全书（Kyoto Encyclopedia of Genes and Genomes, KEGG）富集分析结果显示，上述4个关键基因参与中性粒细胞胞外陷阱形成与丝裂原活化蛋白激酶（mitogen-activated protein kinase, MAPK）信号通路。免疫浸润分析结果表明，4个关键基因均与浆细胞呈正相关，与滤泡辅助性T细胞、M2型巨噬细胞及树突状细胞（dendritic cells, DCs）呈负相关。
研究结论：本研究揭示了牙周炎与骨质疏松症通过串扰基因及焦亡相关基因实现的共同发病机制。关键基因PRKCB、GSDMD、ARMCX3及CASP3通过参与中性粒细胞胞外陷阱形成与MAPK信号通路，同时影响两种疾病的发生发展。本研究结果可为未来相关研究提供方向。