This presentation introduces efforts in predicting novel drug target genes using gene network analysis powered by Fugaku. While gene network analysis has traditionally been conducted, it has not sufficiently considered the variability in individual patients' responses to drugs and pathogens. Although AI-based methods are often evaluated solely on their predictive capabilities, this study expands on the concept of explainable AI using Bayesian networks. Leveraging Fugaku's computational power, we present an approach to explore novel drug candidate genes in a manner comprehensible to humans.