多种机器学习模型构建阿尔兹海默病失巢凋亡相关预测模型OACSTPCD
Construction of the anoikis-related prediction model for Alzheimer's disease based on various machine learning models
阿尔兹海默病(AD)是最常见的神经退行性疾病.失巢凋亡(Anoikis)是一种新型的程序性细胞死亡方式,可导致多种疾病的发展.本研究旨在探讨失巢凋亡相关基因(ANRGs)在AD中的作用,并建立预测模型.基于GSE33000 数据集筛选到1 666 个AD与正常脑组织相比差异表达的基因,与 53 个ANRGs取交集,得到 10 个基因.利用上述基因,对 310 例AD患者进行无监督聚类,将其分为 3 个亚型,进一步分析不同亚型间的免疫微环境差异.之后,采用WGCNA算法筛选与AD相关的特征基因,选取 4 种机器学习算法(RF,GLM,SVM和XGB),构建AD罹患风险的预测模型,并在 3 个外部队列中进行验证(GSE5281,GSE29378,GSE122063).最后,基于XGB模型中的 5 个AD特征基因(TM6SF1,SMYD3,OXCT1,MAP1B和 ITP-KB),成功构建一个列线图,为AD的临床预测提供参考.
Alzheimer's disease(AD)is the most common neurodegenerative disease.Anoikis is a new type of programmed cell death that can lead to the development of many diseases.The purpose of this study is to investigate the role of anoikis-related genes(ANRGs)in AD and establish a prediction model.Based on GSE33000,1 666 differentially expressed genes are screened,and 10 genes are obtained by intersection with 53 ANRGs.Using the above genes,310 patients with AD are classified into three subtypes by unsupervised clustering,and the differences of immune microenvironment among different subtypes are further analyzed.After that,WGCNA algorithm is used to screen the characteristic genes associated with AD,and combined with four machine learning models(RF,GLM,SVM and XGB),the AD risk prediction model is constructed and verified in three external cohorts(GSE5281,GSE29378 and GSE122063).Finally,we successfully construct a nomogram based on five AD characteristic genes(TM6SF1,SMYD3,OXCT1,MAP1B and ITPKB)of the XGB model to provide reference for clinical prediction of AD.
范钰;陈婷婷;陈钢;张永康
山西医科大学 第五临床医学院,太原 030012山西医科大学 基础医学院,太原 030001山西医科大学 第五临床医学院,太原 030012||山西省人民医院 中医科,太原 030012
临床医学
阿尔兹海默病失巢凋亡分子分型机器学习预测模型
Alzheimer's diseaseAnoikisMolecular clustersMachine learningPrediction model
《生物信息学》 2024 (003)
192-203 / 12
2022年张永康全国名老中医药家传承工作室建设项目(国中医药办人教函[2022]75号);山西省中医药管理局张永康学术经验研究项目(No.2019ZYYC041).
评论