Host: The Japanese Society of Toxicology
Name : The 49th Annual Meeting of the Japanese Society of Toxicology
Date : June 30, 2022 - July 02, 2022
Although genome-wide association studies (GWAS) have revealed numerous genetic variants associated with complex diseases, discovering therapeutic targets and biomarkers need translation of GWAS findings into biological knowledge. We have recently proposed a novel research strategy, "Polygenicity in a dish," in which polygenic risk scores (PRS) constructed from GWAS results are analyzed in detail in human cells and organoid models. Our drug-induced liver injury (DILI) study showed that this strategy worked well. Based on the hypothesis that the DILI risk can be explained by the sum of very tiny effects of many SNPs, we constructed PRS models by collaboration with the international consortiums for DILI (iDILIC and DILIN). Clinical trial data and in vitro survival assays with primary human hepatocytes and iPS cell-derived liver organoids showed that the genetic DILI risk was associated with the DILI vulnerability of many DILI-inducible drugs. Interestingly, in vitro transcriptome analysis and survival assays demonstrated that higher genetic DILI risk was related to biological pathways such as oxidative stress responses, indicating the validity of our "Polygenicity in a dish" strategy. Besides, we have also proposed an AI model, MENTR, to predict non-coding RNA expressions (especially enhancer RNAs from an activated enhancer) from only DNA sequence patterns, inspired by the conventional sequence motif analysis. MENTR has succeeded in interpreting rare variants associated with complex diseases, such as asthma and atopic dermatitis, by linking cell-type-specific enhancer RNA expressions. In this talk, we will introduce the latest trends in utilizing statistical analysis and machine learning techniques.
