A co-expression modules based gene selection for cancer recognition

作者：Lu, Xinguo^*; Deng, Yong; Huang, Lei; Feng, Bingtao; Liao, Bo

来源：Journal of Theoretical Biology, 2014, 362: 75-82.

DOI：10.1016/j.jtbi.2014.01.005

• Gene expression data
• WGCNA
• Cancer recognition

Gene expression profiles are used to recognize patient samples for cancer diagnosis and therapy. Gene selection is crucial to high recognition performance. In usual gene selection methods the genes are considered as independent individuals and the correlation among genes is not used efficiently. In this description, a co-expression modules based gene selection method for cancer recognition is proposed. First in the cancer dataset a weighted correlation network is constructed according to the correlation between each pair of genes, different modules from this network are identified and the significant modules are selected for following exploration. Second, based on these informative modules information gain is applied to selecting the feature genes for cancer recognition. Then using LOOCV, the experiments with different classification algorithms are conducted and the results show that the proposed method makes better classification accuracy than traditional gene selection methods. At last via gene ontology enrichment analysis the biological significance of the co-expressed genes in specific modules was verified.

• 出版日期2014-12-7
• 单位湖南大学; 中国人民解放军国防科学技术大学