MDQM9-nc dataset

This dataset contains MD simulations of 12,530 non-cyclic molecules from the QM9 dataset in vacuum and room temperature using the GAFF force field. We carried out these simulations using the openmmforcefields and OpenMM packages. All initial conditions were generated by energy minimizing the QM9 geometry in the corresponding GAFF force field. We sampled different molecules proportionally to their number of heavy atoms with a median sampling time of 36.5 ns. Moreover, for 100 molecules from the test set (10 %), we run longer 100 ns Replica Exchange (RE) simulations.