Abstract 6575: LinkedOmicsKB: A web portal to explore pan-cancer molecular and phenotype associations

Abstract

Abstract Pan-cancer multi-omics data produced from integrated genomic, epigenomic, transcriptomic, proteomic, and post-translational modification (PTM) profiling of a wide variety of cancer types holds great potential for understanding cancer biology and generating therapeutic hypotheses. To realize this potential and make analysis and visualization of these complex and interconnected data easily accessible to cancer biologists and clinicians, we have developed a web portal, LinkedOmicsKB. The web portal provides access to a harmonized proteogenomic dataset of over 1000 patient samples covering 10 cancer cohorts from the Clinical Proteomic Tumor Analysis Consortium (CPTAC). We calculated associations between methylation, copy number variation (CNV), RNA, protein, and phosphosite data for each gene and further correlated the proteogenomics data with clinical and computed molecular phenotypes. All results are stored in a MongoDB database and visualizations are provided for exploring pan-cancer multi-omics relationships as well as individual statistics. We demonstrate the utility of LinkedOmicsKB to provide insights into clinical phenotypes, somatic mutations, and understudied genes. In the pan-cancer CPTAC data, overall survival was correlated with proteins involved in protein hydroxylation, including PLOD1 and PLOD2. Additionally, two phosphorylation sites on the tumor suppressor MIG-6 were associated with worse survival. These sites were also associated with hypoxia, MAPK, and EGFR pathway activity scores, suggesting a relationship between the signaling in these pathways and cancer prognosis. STK17B is an understudied kinase that regulates apoptosis. We found STK17B was upregulated in 6 cancer types at the protein level but only two at the RNA level. The protein abundance of STK17B was highly associated with immune scores and JAK/STAT signaling, supporting a role for STK17B in the immune response. We identified 3 phosphorylation sites on STK17B, which were associated with EGF pathway activity scores and immune-related scores. LinkedOmicsKB is a valuable tool that can be used to generate biological and clinical insights into any gene, phosphosite, mutation, or phenotype. Citation Format: Sara R. Savage, Yuxing Liao, Yongchao Dou, Zhiao Shi, Xinpei Yi, Wen Jiang, Jonathan T. Lei, Bing Zhang. LinkedOmicsKB: A web portal to explore pan-cancer molecular and phenotype associations [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 6575.