Abstract

: Predicting the functional consequences of genetic variation remains a central challenge in human genetics, with direct implications for understanding disease mechanisms and improving clinical decision-making. In this talk, I will discuss my lab’s recent advances in developing effective genomic language models that learn functional constraints directly from DNA sequences alone, and showcase how these models are enabling new insights into variant interpretation and analysis of complex traits.

Speaker Bio

: Yun S. Song is a Professor of EECS and Statistics at the University of California, Berkeley. He was originally trained in mathematics and theoretical physics, with degrees from MIT and Stanford University, but transitioned into biology during his postdoc at the University of Oxford. Over the past two decades, he has conducted interdisciplinary research on diverse computational biology problems. His recent research interests include developing AI/machine learning models for biological sequences (DNA, RNA, and proteins), single-cell genomics, metagenomics, and immunology. His awards and honors include the NIH Pathway to Independence Award, the NSF CAREER Award, an Alfred P. Sloan Research Fellowship, a Packard Fellowship for Science and Engineering, and the Chan Zuckerberg Biohub Investigator Award.