Large-scale conformational transitions represent both a challenge and an opportunity for computational drug design. Exploring the conformational space of a druggable target with sufficient detail is computationally demanding. However, if it were possible to fully account for target flexibility, one could exploit this knowledge to rationally design more potent and more selective drug candidates. Here, we discuss how molecular dynamics together with free energy algorithms based on Metadynamics and Path Collective Variables can be used to study both large-scale conformational transitions and ligand binding to flexible targets. We show real-life examples of how these methods have been applied in the case of cyclin-dependent kinases, a family of flexible targets that shows promise in cancer therapy.