Abstract Introduction: Decades of efforts have been made to have a better understanding about the human genome, which encodes ~20,000 protein-coding genes as well as an astounding number of transcriptionally active noncoding RNAs. Among these noncoding RNAs, long-noncoding RNAs (lncRNAs) have been increasingly recognized to play essential roles in tumorigenesis, representing a new focus in cancer research. Analogous to the identification of protein-coding cancer driver genes, it is a crucial task in cancer research to identify lncRNA drivers and elucidate their functions in various cancer types. Methods: We first downloaded TCGA RNA-seq BAM files (>60TB) across 20 cancer types from the UCSC Cancer Genomic Hub (CGHub) and quantified the lncRNA expression based on GENCODE lncRNA annotation. To identify lncRNA drivers, we focused on identifying the functional relevant lncRNA-protein pairs through their expression profiling data. To prevent acquiring more false positives from conventional correlation analysis, we applied the context likelihood of relatedness (CLR) algorithm on evaluating co-expressed lncRNA and protein pairs based on RNA-seq and reverse-phase protein array (RPPA) data. In addition, to further increase the clinical relevance of the potential lncRNA drivers, we conducted survival analysis and tumor/normal differential analysis to boost the prediction power of our systematic computational approach. Results: In this study, we integrated lncRNA expression, protein expression and patient survival data to nominate lncRNA drivers in 20 cancers. Among ~4 million lncRNA-protein pairs, we have identified 9 lncRNA drivers and 16 potential lncRNA-protein interaction pairs linked with them. To validate our findings, we selected 3 lncRNAs with high confidence and performed knockdown experiments in multiple cell lines, in which 2 lncRNAs (TMPO-AS1 and HCP5) showed functional relevance with protein-coding genes MSH6 and LCK respectively. Both of the pairs have the same direction of correlations (positive or negative) as we predicted. Our results indicate that the identified lncRNAs may functionally interact with their protein partners identified by our network-based approach, which will further provide the molecular foundation for investigating cancer mechanisms, identifying prognostic lncRNA markers and characterizing additional therapeutic targets. Citation Format: Jun Li, Zhicheng Zhou, Han Liang. A systematic network-based approach to prioritizing lncRNA drivers in cancers. [abstract]. In: Proceedings of the AACR Special Conference on Noncoding RNAs and Cancer: Mechanisms to Medicines ; 2015 Dec 4-7; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2016;76(6 Suppl):Abstract nr A36.