Machine Learning at the Flatiron Institute Seminar: Sonya Hanson

Title: Applications of Machine Learning in Structural Biology

Abstract: Structural biology is a branch of molecular biology that focuses on understanding the three-dimensional structures of biological macromolecules —- proteins and nucleic acids —- and their dynamics. These structures are fundamental to all cellular processes, making structural biology essential for advancements in drug discovery, the understanding of disease, and development in biotechnology. Within the Structural and Molecular Biophysics (SMBp) team at the Flatiron Institute, we develop cutting-edge tools and methods in cryo-electron microscopy (cryo-EM) image analysis and molecular dynamics (MD) simulations. In recent years there has been a push in machine learning (ML) methods applied to these fields. One of the most significant advances is AlphaFold, which predicts protein structures from sequences with remarkable accuracy. The impact and practical applications for this tool in our own work will be discussed. Additionally, ML has driven a surge of innovative methods in cryo-EM, our own work in this area is mostly inspired by advances in computer vision applied to cryo-EM. Finally, ML is being used to try to reduce the computational costs of MD simulations, and we will review both our own ideas in this space and the state-of-the-art in the field at this time. Overall, while the impact of ML on structural biology is already clear there are many future areas for development, especially when it comes to expanding the time and length scales of our understanding of the molecular foundations of life.