About the Speaker
Francesco Paolo Casale is a Senior Research Group Leader at Human Technopole in Milan, Italy, also affiliated with Helmholtz Munich and the Faculty of Informatics at the Technical University of Munich in Garching, Germany. After training in Physics at the University of Naples Federico II, he completed a PhD in human genetics and statistical modeling at the University of Cambridge and EMBL-EBI, followed by postdoctoral work in AI for biology at Microsoft Research in Boston. He then led an interdisciplinary team at insitro in the San Francisco Bay Area, applying machine learning and human genetics to target discovery. Since 2022, he has been a Principal Investigator at Helmholtz Munich, where he led a research program developing AI methods that integrate human genetics and multimodal data to build mechanistic models of disease. In September 2026, he joined Human Technopole as a Senior Research Group Leader.
Talk Abstract
Understanding how genetic variation leads to human disease requires phenotypes that capture biology at multiple scales, from tissue morphology to an individual’s overall health. In this talk, I will discuss how artificial intelligence can help derive such quantitative phenotypes from complex biomedical data and connect them to disease genetics.
I will first show how representations learned from medical images can be combined with genetic association studies to identify and interpret genetic effects on tissue and organ morphology. Using examples from histology and retinal imaging, I will illustrate how these approaches can reveal disease-relevant genetic loci and provide interpretable views of their phenotypic consequences in tissues and organs [1, 2].
I will then move to representations of human health, presenting RisQ, a multimodal framework that learns shared structure in disease risk across hundreds of diseases and prediction horizons from biobank-scale data. These representations support disease prediction while providing quantitative phenotypes for genetic discovery [3].
Together, these examples illustrate how AI can expand the phenotypic space available to human genetics and help connect genetic associations to interpretable models of disease biology.
[1] Chaudhary S, et al. HistoGWAS: an AI-enabled framework for automated genetic analysis of tissue phenotypes in histology cohorts. Genome Biology. 2026. [2] Shilova L, et al. REECAP: Contrastive learning of retinal aging reveals genetic loci linking morphology to eye disease. medRxiv. 2025. [3] Hager P, et al. Learning the shared structure of human health across diseases, modalities, and time. medRxiv. 2026.
We are looking forward to the talk!