10/21(火) OUBICセミナー（CiDER共催）

Abstract

Antimicrobial resistance (AMR) represents one of the most pressing global health threats, with resistant pathogens causing millions of deaths annually and compromising the effectiveness of modern medicine. As traditional laboratory-based susceptibility testing becomes increasingly inadequate for rapid clinical decision-making, genomics-based approaches powered by artificial intelligence and machine learning offer promising alternatives for AMR prediction and surveillance. We will explore the current landscape of bioinformatics tools and AI methodologies applied to genomic data for AMR detection and prediction. We will examine the fundamental principles underlying classification models used in this domain. Key considerations for establishing robust datasets will be discussed, including feature and class encoding. Despite significant advances, current AI approaches face substantial limitations including model interpretability, generalizability across diverse bacterial populations, and the challenge of rare resistance mechanisms. We will critically assess these constraints and explore emerging solutions. We will conclude by outlining future research directions that could enhance the clinical translation of AI-driven AMR prediction tools, ultimately supporting more effective antimicrobial stewardship and improved patient outcomes. ）

Abstract

One of the major challenges in controlling viral infectious diseases lies in the rapid evolution of viruses, which enables them to swiftly alter their virological characteristics. During the COVID-19 pandemic, for example, successive variants with enhanced immune-escape capabilities and increased fitness emerged, making epidemic control extremely difficult. Advancing our understanding—and ultimately our ability to predict—viral evolution and epidemic dynamics could pave the way for more effective strategies to control infectious diseases.
In this seminar, I will introduce CoVFit, an AI model that predicts viral fitness, and PLANT, an AI model that predicts viral antigenicity. Through these examples, I will explore the future of AI for infectious diseases.

10月21日(火) 9：00-10：00