15/05/2025
WHO: Prof. Rachel Kolodny, Haifa University
WHEN: Thursday, May 15th 2025 at 12:00
WHERE: BUILDING 503 (Computer Science) AUDITORIUM
Title: Predicting gene sequences with AI to study codon usage patterns
Abstract:
We use AI models to study a problem in evolutionary molecular biology that has potential impact in biotechnology. I will start with describing the problem of predicting the codons used by different organisms to encode proteins. These codons are under evolutionary selective pressure that optimizes multiple, overlapping signals that are only partially understood. Our approach is to train AI models that can predict codons given their amino acid sequence in four organisms (two eukaryotes and two bacteria), and study the extent to which we can successfully learn patterns in naturally occurring codons to improve predictions. Finally, I will also describe our observations regarding using information encoded in homologous proteins. Using contemporary AI methods offers a new perspective on codon usage patterns, which is important both for understanding evolutionary processes, but also as a novel tool for biotechnology to optimize codon usage in endogenous and heterologous proteins.
Joint work with Tomer Sidi, Shir Bahiri-Elizur, and Tamir Tuller (published PNAS 2024)
Short Bio:
Rachel Kolodny is a Professor of Computer Science in the University of Haifa. Before joining the department of Computer Science in 2006, she completed her PhD at the Stanford school of engineering, working with Leo Guibas and Chemistry Nobel laureate Michael Levitt, and her post-doctoral training with Prof. Barry Honig at the Columbia medical school. In the years 2022-2023, she served as departmental chair.
