基于加权基因共表达网络分析和机器学习筛选椎间盘退行性变中的标志线粒体基因

Screening for marker mitochondrial genes in pathogenesis of intervertebral disc degeneration based on integrated weighted gene co-expression network analysis and advanced machine learning

目的采用加权基因共表达网络分析(WGCNA)和机器学习筛选椎间盘退行性变(IDD)中的标志线粒体基因。方法从基因表达汇编(GEO)数据库下载椎间盘相关芯片数据,采用WGCNA对芯片GSE70362数据集中IDD的相关基因模块与线粒体基因取交集,联合机器学习[支持向量机递归特征消除(SVM-RFE)、最小绝对收缩和选择算子(LASSO)回归和随机森林(RF)算法]进一步筛选IDD标志线粒体基因,采用受试者工作特征(ROC)曲线验证标志线粒体基因的有效性。下载GSE15227数据集对标志线粒体基因进行外部验证。利用Cibersort算法评估IDD标志线粒体基因与免疫浸润的生物信息学关联。结果通过WGCNA联合机器学习算法共筛选出4个IDD标志线粒体基因,分别为BNIP3、ISCU、MCUB及SPTLC2。ROC曲线显示BNIP3、ISCU、MCUB、SPTLC2的曲线下面积(AUC)分别为0.562,0.780,0.766,0.702。通过外部数据集验证发现只有BNIP3的表达差异具有统计学意义(P<0.05)。免疫细胞差异分析显示,与正常对照组相比,IDD患者的单核细胞数量较少,差异具有统计学意义(P<0.05)。BNIP3与调节性T细胞(Tregs)呈正相关(r=0.43),与滤泡辅助性T细胞呈负相关(r=-0.40)。结论线粒体基因BNIP3、ISCU、MCUB和SPTLC2可作为新型生物标志物,BNIP3可能通过调控免疫细胞参与IDD的发生。以上基因可作为基于线粒体稳态策略治疗IDD的潜在靶点。

Objective To screen marker mitochondrial genes in intervertebral disc degeneration(IDD)using weighted gene co-expression network analysis(WGCNA)and advanced machine learning.Methods Intervertebral disc degeneration microarray data were downloaded from the Gene Expression Omnibus(GEO)database,and WGCNA was employed to intersect the IDD related gene modules and mitochondrial genes in the GSE70362 dataset.Advanced machine learning approaches(support vector machine-recursive feature elimination[SVM-RFE],least absolute shrinkage and selection operator[LASSO]regression and random forest[RF]algorithm)were used to screen the marker mitochondrial genes in IDD further.The effectiveness of the marker mitochondrial genes was validated using receiver operating characteristic(ROC)curve.External validation was conducted using the GSE15227 dataset.Additionally,the Cibersort algorithm was utilized to evaluate the correlation between marker mitochondrial genes and immune cell infiltration.Results Four marker mitochondrial genes(BNIP3,ISCU,MCUB and SPTLC2)were identified by integrated WGCNA and machine learning algorithms.The ROC curve showed an area under the curve(AUC)for BNIP3,ISCU,MCUB and SPTLC2 as 0.562,0.780,0.766 and 0.702,respectively.External dataset validation revealed that only the expression of BNIP3 exhibited statistically significant difference(P<0.05).Immunocyte differential analysis showed that compared to the normal control group,the number of monocytes in IDD patients was fewer,and the difference was statistically significant(P<0.05).BNIP3 was positively correlated with regulatory T cells(Tregs,r=0.43)and negatively correlated with follicular helper T cells(r=-0.40).Conclusions Mitochondrial gene BNIP3 ISCU,MCUB and SPTLC2 can serve as novel biomarkers,and BNIP3 may participate in the occurrence of IDD by regulating immune cells.The above genes can serve as potential targets for the treatment of IDD based on mitochondrial homeostasis strategies.