SSJNMF算法识别基因-药物共模块探讨

Identification of Gene-Drug Common Modules with SSJNMF Algorithm

为识别基因-药物共模块,更好地发现抗癌药物作用的分子机理,选择针对既定癌症的潜在药物靶标,在非负矩阵分解算法中加入基因、药物和基因-药物模块的数据信息,提出了SSJNMF算法。将该算法用于癌症药物敏感性基因组学数据集,识别了70个基因-药物共模块,与已发表的两种算法(NetNMF和JNMF)进行比较,结果表明,SSJNMF识别出的共模块具有更强的统计意义和生物意义。

To identify gene-drug common modules,and better discover the molecular mechanism of anticancer drug action,potential drug targets for specific cancers are selected.The SSJNMF algorithm is proposed by adding the data information of gene,drug and gene-drug module into the non-negative matrix decomposition algorithm.The proposed algorithm is applied to the cancer drug sensitivity genomics dataset and 70 gene-drug co-modules are identified and compared with two published algorithms,i.e.NetNMF and JNMF.The results show that the common modules identified by SSJNMF have greater statistical and biological significance.