The rise of machine learning in structural biology and molecular engineering

Proteins are the fundamental molecular building blocks of life, dictating the spatial and temporal occurrence of most biological functions. They ultimately encode the genetic information of every organism and translate it into the physical and chemical mechanisms at the basis of most of the molecular processes sustaining living organisms. They have evolved to do so by specifically interacting with other proteins, nucleic acids, metabolites, membranes in order to form supramolecular complexes whose structural architecture is at basis of their biological function. Predicting their structure and the way they form functional interacting networks remains however a major challenge. Understanding the principles governing these physicochemical interactions is moreover at the basis of our ability to engineer and design molecular entities able to deliver controllable functionalities for therapeutic intervention. Since the introduction of AlphaFold, machine learning methods have been evolving at an unprecedented pace and are now pervasive to address any of these problems in protein science. In this lecture, I will introduce the latest concepts and advances in this domain along with the recent developments done in my laboratory in the domain of geometric deep learning applied to integrative structural biology and protein design and engineering. These advances are anticipated to enhance our comprehension of fundamental biology and to facilitate the development of more effective therapies.

Speaker

ISA Visiting Fellow - Matteo Dal Peraro

Ecole Polytechnique Fédérale de Lausanne – EPFL, Switzerland

Visit Prof. Dal Peraro's web page

PhD students and researchers who are interested may request an attendance certificate by writing to segreteria.isa@unibo.it specifying their birthplace and date of birth.

The delivery of the attendance certificate requires the attendance of at least 70% of the lecture.