Transcriptomics-based Screening and Novel compound for Hepatocellular Carcinoma

HCC is the sixth most common cancer and the third leading cause of cancer-related death worldwide. It has a high mortality rate, yet lacks effective treatment options. We used our Gene expression profile Predictor on chemical Structures (GPS) platform to discover novel and selective anti-HCC drug candidates. GPS is a deep learning-based drug discovery system for the screening of a large compound library and de novo designing of novel compounds that can reverse transcriptional phenotype. To achieve this, a previously defined HCC signature was queried against the GPS-generated transcriptomic profiles for compounds in the ZINC library with almost seven million drug-like compounds. Top-ranked compounds were nominated for testing their cytotoxicities in HCC cell lines in vitro, and their efficacy in vivo. Overall design: Here we include samples and processed data (HCC_CDX_TPM.csv,HCC_CDX_RAWCOUNT.csv) for Huh7 xenograft samples treated with our novel compound MSU45302 or DMSO. In addition, we also provide raw and processed data (HUH_4852_TPM.csv,HUH_4852_RAWCOUNT.csv) for Huh7 cellline treated with compound PB56874852 (two concentrations 4µM and 10µM) and DMSO. We designed MSU45302 by substituting specific structural components of PB56874852, optimizing its properties for enhanced efficacy.