用于识别与晚期肾病相关关键基因的预测模型的构建与验证

Construction and Validation of Predictive Model to Identify Critical Genes Associated with Advanced Kidney Disease.

作者信息

Xin Guangda, Zhou Guangyu, Zhang Wenlong, Zhang Xiaofei

机构信息

Department of Nephrology, China-Japan Union Hospital of Jilin University, Changchun, China.

Department of Matological and Oncological, China-Japan Union Hospital of Jilin University, Changchun, China.

出版信息

Int J Genomics. 2020 Nov 12;2020:7524057. doi: 10.1155/2020/7524057. eCollection 2020.

BACKGROUND

Chronic kidney disease (CKD) is characterized by progressive renal function loss, which may finally lead to end-stage renal disease (ESRD). The study is aimed at identifying crucial genes related to CKD progressive and constructing a disease prediction model to investigate risk factors.

METHODS

GSE97709 and GSE37171 datasets were downloaded from the GEO database including peripheral blood samples from subjects with CKD, ESRD, and healthy controls. Differential expressed genes (DEGs) were identified and functional enrichment analysis. Machine learning algorithm-based prediction model was constructed to identify crucial functional feature genes related to ESRD.

RESULTS

A total of 76 DEGs were screened from CDK vs. normal samples while 10,114 DEGs were identified from ESRD vs. CDK samples. For numerous genes related to ESRD, several GO biological terms and 141 signaling pathways were identified including markedly upregulated olfactory transduction and downregulated platelet activation pathway. The DEGs were clustering in three modules according to WGCNA access, namely, ME1, ME2, and ME3. By construction of the XGBoost model and dataset validation, we screened cohorts of genes associated with progressive CKD, such as , , and . represented the highest score ( score = 21) in predictive model.

CONCLUSION

Our results demonstrated that , , , and might be critical genes in CKD progression.

背景

慢性肾脏病（CKD）的特征是肾功能进行性丧失，最终可能导致终末期肾病（ESRD）。本研究旨在鉴定与CKD进展相关的关键基因，并构建疾病预测模型以研究危险因素。

方法

从基因表达综合数据库（GEO数据库）下载GSE97709和GSE37171数据集，其中包括CKD患者、ESRD患者的外周血样本以及健康对照。鉴定差异表达基因（DEG）并进行功能富集分析。构建基于机器学习算法的预测模型，以鉴定与ESRD相关的关键功能特征基因。

结果

从CKD与正常样本中筛选出共76个DEG，而从ESRD与CKD样本中鉴定出10114个DEG。对于众多与ESRD相关的基因，鉴定出了几个基因本体（GO）生物学术语和141条信号通路，包括嗅觉转导明显上调和血小板激活途径下调。根据加权基因共表达网络分析（WGCNA）算法，这些DEG聚类为三个模块，即ME1、ME2和ME3。通过构建XGBoost模型和数据集验证，我们筛选出了与CKD进展相关的基因队列，如[此处原文缺失具体基因名称]、[此处原文缺失具体基因名称]和[此处原文缺失具体基因名称]。[此处原文缺失具体基因名称]在预测模型中得分最高（得分=21）。