Our research focuses on the development of computational methods to study molecular recognition (ligand binding) and large-scale conformational changes in bio-molecules.  Molecular dynamics (MD) simulations have been successfully employed to model biological phenomena. But MD is fundamentally limited by the time scales that can be accessed (the time-scale problem). When I was at ETH Zurich as assistant professor in the group of Prof. Michele Parrinello (2006-2009), I crucially contributed to the development of 3 widely used methods to overcome the time-scale problem and compute free energy surfaces, namely Metadynamics, Parallel-Tempering Metadynamics and the path-like collective variables method (PCV). These methods are able to efficiently compute the free energy landscape associated with complex events, such as folding, lingand-target association and large-scale conformational changes. As the leader of the Computational Biophysics group at CNIO and more recently Professor at UCL, I continued the development of computational methods, including a very efficient method to compute the kinetics associated with complex biological events (TS-PPTIS), and applied these methods to better understand the mode of action of anticancer drugs. A recent highlight of our research was the clarification of the complex mode of action of the first allosteric inhibitor (SSR) of the Fibroblast Growth Factor Receptor. The free energy simulations have shown that SSR induces a previously unknown conformational change in the extracellular portion of FGFR. The accurate structural information obtained from the simulations has been used to design SSR derivatives now in pre-clinical development as anti-cancer agents. Furthermore, by using PT-metaD we reconciled different views on the mode of action of oncogenic and drug-resistance mutations in the Epidermal Growth Factor Receptor.