Research Lines.
Our group focuses on the development of novel computational approaches to understand the structure and dynamics of complex biological systems.
Deep Learning & AI
Geometric Algebra, Flow Matching, and diffusion models for biomolecular modeling.
Macromolecular Simulations
Elastic Network Models, Normal Mode Analysis, Dynamic Relaxation, and coarse-grained simulations
Integrative Modeling
Combining multiresolution experimental data (X-ray crystallography, cryo-EM, SAXS) with computational models.
Protein Modeling
Developing tools for predicting protein-protein interactions, loop closure modeling, and knowledge-based potentials.
AI Protein Design
We develop generative deep learning models, built on geometric algebra, flow matching and diffusion, to design new proteins from scratch. Instead of only predicting how known sequences fold, these models create novel structures and sequences tailored to a target shape or function.
Normal Mode Analysis
Visualizing the collective motions of protein structures using our highly efficient Normal Mode Analysis (NMA) tools. These simulations reveal intrinsic flexibility essential for biological function without the computational cost of full MD simulations.
Flexible Fitting into Cryo-EM
Our hybrid methods bridge the resolution gap by flexibly fitting high-resolution atomic structures into lower-resolution electron microscopy density maps, providing detailed atomistic models of large molecular machines.
Ab initio Loop Modeling
Explore an interactively rendered 3D prediction generated by our fast loop-closure modeling tool. Accurate all-atom loop predictions are critical for structural refinement and antibody design.