Chao-Yang Gong, Rong Tang, Yong-Qiang Shi, Tai-Cong Liu, Kai-Sheng Zhou, Wei Nan, Hai-Hong Zhang
Lanzhou University Second Hospital, Lanzhou, China.
Orthopaedics Key Laboratory of Gansu Province, Lanzhou, China.
Front Cell Dev Biol. 2021 Feb 23;9:644220. doi: 10.3389/fcell.2021.644220. eCollection 2021.
In this study, we identified eight survival-related metabolic genes in differentially expressed metabolic genes by univariate Cox regression analysis based on the therapeutically applicable research to generate effective treatments ( = 84) data set and genotype tissue expression data set ( = 396). We also constructed a six metabolic gene signature to predict the overall survival of osteosarcoma (OS) patients using least absolute shrinkage and selection operator (Lasso) Cox regression analysis. Our results show that the six metabolic gene signature showed good performance in predicting survival of OS patients and was also an independent prognostic factor. Stratified correlation analysis showed that the metabolic gene signature accurately predicted survival outcomes in high-risk and low-risk OS patients. The six metabolic gene signature was also verified to perform well in predicting survival of OS patients in an independent cohort (GSE21257). Then, using univariate Cox regression and Lasso Cox regression analyses, we identified an eight metabolism-related long noncoding RNA (lncRNA) signature that accurately predicts overall survival of OS patients. Gene set variation analysis showed that the apical surface and bile acid metabolism, epithelial mesenchymal transition, and P53 pathway were activated in the high-risk group based on the eight metabolism-related lncRNA signature. Furthermore, we constructed a competing endogenous RNA (ceRNA) network and conducted immunization score analysis based on the eight metabolism-related lncRNA signature. These results showed that the six metabolic gene signature and eight metabolism-related lncRNA signature have good performance in predicting the survival outcomes of OS patients.
在本研究中，我们基于治疗适用性研究生成有效治疗方法（=84）数据集和基因型组织表达数据集（=396），通过单变量Cox回归分析在差异表达的代谢基因中鉴定出八个与生存相关的代谢基因。我们还使用最小绝对收缩和选择算子（Lasso）Cox回归分析构建了一个六个代谢基因特征，以预测骨肉瘤（OS）患者的总生存期。我们的结果表明，六个代谢基因特征在预测OS患者的生存方面表现良好，并且也是一个独立的预后因素。分层相关性分析表明，代谢基因特征准确预测了高危和低危OS患者的生存结果。六个代谢基因特征在独立队列（GSE21257）中预测OS患者生存方面也被验证表现良好。然后，通过单变量Cox回归和Lasso Cox回归分析，我们鉴定出一个准确预测OS患者总生存期的八个与代谢相关的长链非编码RNA（lncRNA）特征。基因集变异分析表明，基于八个与代谢相关的lncRNA特征，高危组中顶端表面和胆汁酸代谢、上皮间质转化和P53通路被激活。此外，我们构建了一个竞争性内源RNA（ceRNA）网络，并基于八个与代谢相关的lncRNA特征进行了免疫评分分析。这些结果表明，六个代谢基因特征和八个与代谢相关的lncRNA特征在预测OS患者的生存结果方面表现良好。
