Personalized cancer drug repurposing using single cell RNA-Seq data

Abstract Human cancers are highly heterogeneous. Studying tumor and tumor-microenvironment (TME), particularly those that are likely to develop resistance, is essential for the effective development of personalized anti-cancer therapy. In recent years the development of new single-cell genomics technologies, such as single-cell RNA- Seq (scRNA-Seq), has made this analysis possible at the single-cell level. Drug repurposing is an effective strategy for identifying and developing new uses of drugs outside the scope of their original medical application. This strategy offers various advantages over developing an entirely new drug for cancer therapeutics, such as the lower risk of failure, significantly shorter time frame for drug development, and much less cost. The “reverse matching” approach to search for drugs from large pharmacogenomics databases such as the LINCS L1000 system has been shown as an effective approach to apply to the genomics data. However, conventional computational methods using the reverse matching principle have been relying on bulk samples, overlooking the complexity of the tumor ecosystem. To fill in the gap, we have recently developed a comprehensive drug repositioning computational framework called ASGARD (version 1.0), which utilizes the single cell RNA-Seq (scRNA-Seq) transcriptomics data as the input. In this project, we will improve ASGARD version 1.0 to the next version, by fully exploiting the features related to tumorigenesis. We will consider genetic information, in addition to gene expression from scRNA-Seq data; we will also include interactions between tumor cells and TME cells to formulate the new and improved drug scores, for drug ranking. After benchmarking, we will apply ASGARD version 2.0 to repurpose drugs for primary liver cancers. We will validate these drug candidates using a combination of literature evidence, in vitro experiments and electronic medical record (EMR) data.