Structural phase transitions in crystals induced by pressure or temperature are complex phenomena of great importance. Besides fundamental interest, e.g. in Earth and planetary science, they also allow us to prepare new materials with unique properties. Most of these transitions are reconstructive, thermodynamically first order, and involve breaking and creation of chemical bonds. In past two decades a great progress was achieved in prediction of crystal structures, but a reliable prediction of structural transformations still lags much behind. The main problem is kinetics which ultimately decides about the outcome of the transformation. Understanding kinetics requires a detailed information about free-energy barriers involved in the microscopic transformation pathways. An important ingredient is also nucleation whose understanding in solid-solid transitions is still very incomplete. One of effective approaches to study these phenomena is metadynamics [1,2,3]. Another important ingredient are machine-learning based potentials which nowadays allow to simulate 105 or more atoms with high accuracy. In the talk I will present the progress made in simulations of structural transformations, illustrate the main achievements and outline possible future directions. I will first present the simulations of the B1/B2 structural transformation in NaCl and show the crossover from the collective Burgers mechanism to homogenous nucleation via the Watanabe -Toledano mechanism upon growing system size [4]. The next example will be pressure-induced densification of coesite where several kinetic pathways compete and different outcomes were observed in experiments [5]. Finally, I will show a recent study of structural transformations from diamond to post-diamond phases in carbon at TPa pressures, with specific focus on exploring transformation pathways to theoretically predicted but yet experimentally unobserved BC8 phase [6]. Our results demonstrate that metadynamics with generic collective variables is able to efficiently uncover non-trivial kinetic pathways in pressure-induced structural transitions, including complex crystalline systems with rich polymorphism. [1] A. Laio and M. Parrinello, PNAS 99, 12562-12566 [2002] [2] R. Martoňák, A. Laio, and M. Parrinello, Phys. Rev. Lett. 90, 075503 (2003) [3] R. Martoňák, D. Donadio, A.R. Oganov and M. Parrinello, Nat. Mater. 5, 623 (2006) [4] Matej Badin and Roman Martoňák, Phys. Rev. Lett. 127, 105701 (2021) [5] David Vrba and Roman Martoňák, J. Chem. Phys. 163, 114503 (2025). [6] Roman Martoňák, Sergey Galitskiy, Azat Tipeev, Joseph M. Gonzalez, Ivan I. Oleynik, arXiv:2509.00423 [cond-mat.mtrl-sci].