A flexible penalized integrated analysis of mRNA and miRNA expression levels as biomarkers for endometrial cancer classification

DOI: 10.5176/2251-3833_GHC16.33

Authors: Yang Xu, Donghai Dai, Kai Wang

Abstract: The expression levels of mRNAs and miRNAs in cancer tumors contain valuable information for cancer classification. The availability of The Cancer Genome Atlas (TCGA) data provides a unique opportunity for extracting such information. An integrative method that incorporates information from both mRNA and microRNA and relates them to clinical features is desirable. We use a flexible LASSO method for identifying mRNA and miRNA signatures for important clinical features in endometrial tumors. These features are pathological stage, histological type, and grade. This method identified mRNAs and miRNAs that are known to be risk factors in endometrial cancer. It also identified some novel ones which warrant further investigation. This approach is applicable to studies that involve more types of features.

Keywords: LASSO; endometrial cancer; stage; grade; type; microRNA; mRNA; cancer classification