Lectures and seminars AI-driven design of proteins to study evolution and encode dynamics
Deep learning has rapidly transformed protein science, enabling the design of complex proteins with unprecedented control over three-dimensional structure. Beyond practical applications, protein design also offers a powerful framework for addressing fundamental questions in molecular evolution.
The André lab uses de novo protein design to explore whether stable, folded proteins can arise from the restricted sets of amino acids plausibly available in prebiotic chemistry. In parallel, they develop deep learning–based approaches that incorporate phylogenetic principles to guide de novo protein design and to improve methods for ancient sequence reconstruction. While AI-designed proteins typically adopt well-defined structures, they often lack the conformational dynamics that characterize natural proteins. To address this, they have developed design strategies that enable proteins to sample multiple conformational states, including fold-switching behavior, and they characterize these dynamic properties using nuclear magnetic resonance spectroscopy.
Professor Ingemar André is Professor of Biochemistry and Structural Biology at Lund University. His research lies at the intersection of computational biology, artificial intelligence and experimental biochemistry, with a focus on understanding how proteins fold, interact, evolve and assemble. His group develops AI and computational methods for protein structure prediction and design, alongside experimental approaches to investigate protein properties and self-assembly.
Host
Ana Teixeira, Department of Physiology and Pharmacology (ana.teixeira@ki.se)