基于HARP-Node2vec和GCN的多标签分类模型用于LncRNA相关疾病预测  

A multi-label classification model based on HARP-Node2vec and GCN for LncRNA-related diseases prediction

作　　者：王波 周振宇 廉佐政[1] 徐靖然 牛翔慧 WANG Bo;ZHOU Zhen-yu;LIAN Zuo-zheng;XU Jing-ran;NIU Xiang-hui(College of Computer and Control Engineering,Qiqihar University,Heilongjiang Qiqihar 161006,China)

机构地区：[1]齐齐哈尔大学计算机与控制工程学院,黑龙江齐齐哈尔161006

出　　处：《齐齐哈尔大学学报（自然科学版）》2024年第6期26-29,41,共5页Journal of Qiqihar University(Natural Science Edition)

基　　金：黑龙江省教育厅基本科研业务费面上项目(145209125);齐齐哈尔大学研究生创新科研项目(YJSCX2021014)。

摘　　要：LncRNA在多种生命活动中发挥着重要作用并且与人类所患疾病有着密切的联系。首先将筛选出来的LncRNA、miRNA以及Disease相关节点构成LncRNA-miRNA-Disease网络,然后经过HARP算法进行多次折叠缩小网络,将缩小的网络输入到Node2vec算法以获得节点的嵌入向量,最后用GCN标签分类器预测与LncRNA相关的多种疾病。在五倍交叉验证中获得0.8800的AUC值,表明具有良好的预测性能。Researchers have found that LncRNA play important roles in various biological processes and are closely associated with diseases in humans.In this study,the selected LncRNA,miRNA,and Disease-related nodes are used to construct an LncRNA-miRNA-Disease network.The HARP algorithm is then applied to recursively coarsen the network,and the reduced network is fed into the Node2vec algorithm to obtain node embedding vectors.Finally,a GCN label classifier is used to predict a variety of diseases associated with LncRNA.The model achieves AUC values of 0.8800 in five-fold cross-validation,respectively,indicating that it has good predictive performance.

关 键 词：LncRNA HARP 图卷积网络 关联预测 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]