We combine computational and experimental methods to understand the structure, interactions and evolution of proteins. ​ A fundamental goal is to encode our understanding of proteins into computational and experimental methods that enables the design self-assembling protein systems. We also want to understand the process of self-assembly by inferring self-assembly pathways from simulated protein structures and time-resolved experimental data.

We develop methods based on artificial intelligence for protein structure prediction and design. We also develop methods to simulate the evolution of protein structure. In the experimental domain we develop high-throughput methods for characterizing the biophysical properties of proteins in cells and study folding of proteins in vivo.