An Integrated Approach to Identify Clinically Relevant Long Non-Coding RNAs (lncRNAs) in Cancer

• Activity Number: 5
• Type: Invited
• Date/Time: Sunday, August 4, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Xiaole Shirley Liu*+
• Companies: Dana-Farber and Harvard University
• Keywords: exon arrays ; lncRNA ; model-based ; tumor profiling ; clinical outcome ; copy number variations

Abstract

Despite growing appreciation of the number and importance of long non-coding RNAs (lncRNAs) in normal physiology and diseases, the knowledge on cancer-related lncRNAs is still limited. We utilized probes from array-based expression profiles that are uniquely mapped to annotated lncRNAs to investigate the expression of over 10,000 lncRNAs in 946 tumor samples over four different cancer types. By integrating the expression profiles of these lncRNAs with clinical outcome and somatic copy number alterations, we identified lncRNAs that are associated with specific tumor subtypes and clinical prognosis, and predicted those that are potential drivers of cancer progression. We further validated our computational predictions by experimentally confirming the oncogenic functions of two novel lncRNAs in prostate cancer. Our results demonstrate that public genomic data contain a wealth of information on disease-relevant non-coding RNAs, and integrating orthogonal datasets and clinical information could facilitate the discovery of clinically-relevant lncRNAs and help prioritize their functional studies.