| 引用本文: | 刘爽,冯雯,顾雪锋.基于未折叠蛋白反应基因特征的儿童脓毒症预测模型构建及分子亚型鉴定[J].生物信息学,2025,23(2):131-142. |
| LIU Shuang,FENG Wen,GU Xuefeng.Identification of molecular subtypes and construction of a predictive model for pediatric sepsis based on unfolded protein response genes[J].Chinese Journal of Bioinformatics,2025,23(2):131-142. |
|
| 摘要: |
| 通过分析来自GEO数据库的儿童脓毒症转录组数据,探究了未折叠蛋白反应在该疾病发生机制中的作用。首先利用随机森林和支持向量机算法筛选了诊断模型的候选基因包括EXOSC4,EIF2AK3,CEBPB,WIPI1,EXOSC6,EXTL2和SRPRB,使用多因素逻辑回归构建诊断模型,并利用外部数据集对其进行了验证。接着,分析了这些基因与免疫细胞浸润的相关性,发现它们与中性粒细胞的浸润具有强相关性。此外,通过一致性聚类将儿童脓毒症患者分成了三个亚型,比较了它们在临床特征和炎症因子表达方面的差异。最后,通过加权基因共表达网络分析筛选出每个亚型的核心基因,并发现这三个亚型在免疫系统、代谢和细胞死亡等生物学过程上存在显著差异。药物预测结果显示不同亚型的患者可能对不同种类的药物具有不同的敏感性。总之,这项研究为儿童脓毒症的诊断和精准治疗提供了新的思路。 |
| 关键词: 儿童脓毒症 未折叠蛋白反应 免疫浸润 诊断模型 机器学习 |
| DOI：10.12113/202401005 |
| 分类号:Q343.1+2 |
| 文献标识码:A |
| 基金项目: |
|
| Identification of molecular subtypes and construction of a predictive model for pediatric sepsis based on unfolded protein response genes |
|
LIU Shuang1,2, FENG Wen2, GU Xuefeng2
|
|
(1. School of Health Science and Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China; 2. School of pharmacy, Shanghai University of Medicine & Health Sciences, Shanghai 201318, China)
|
| Abstract: |
| By analyzing transcriptomic data of pediatric sepsis from the GEO database, the role of unfolded protein response in the pathogenesis of this disease was investigated. First, candidate genes for the diagnostic model were screened by using random forest and support vector machine recursive feature elimination algorithms, including EXOSC4, EIF2AK3, CEBPB, WIPI1, EXOSC6, EXTL2and SRPRB.A diagnostic model was constructed by multiple logistic regression and validated with three external datasets. Next, the correlation between these genes and immune cell infiltration was analyzed, revealing a strong correlation with neutrophil infiltration. Furthermore, patients with pediatric sepsis were divided into three subtypes by consensus clustering, and their differences in clinical features and expression of inflammatory factors were compared. Finally, core genes for each subtype were selected through weighted gene co-expression network analysis, and significant differences were found among these three subtypes in biological processes such as the immune system, metabolism, and cell death. Drug prediction results showed that patients with different subtypes may have different sensitivities to different types of drugs. In summary, this study provides new ideas for the diagnosis and precision treatment of pediatric sepsis. |
| Key words: Pediatric sepsis Unfolded protein response Prognostic model Machine learning |
