Molecular dynamics data associated with the publication: "Functional Dynamics of Substrate Recognition in TEM Beta-Lactamase" Trajectores were generated in GROMACS, and the carbon alpha coordinates were extracted and aligned with the JEDi analysis software. Details of the simulations and analysis are given in the publication. Data in apo.zip contains trajectories for 32 total trajectories of TEM-1, TEM-2, TEM-10, and TEM-52 beta-lactamase, each starting form different 8 crystal structures Data in holo.zip contains 16 trajectories of TEM-1, TEM-2, TEM-10, and TEM-52 beta-lactamase in complex with ampicillin, amoxicillin, cefotaxime, and ceftazidime each. Trajectories files are in comma delimited format, with rows representing degrees for freedom (789 total), and columns representing samples (10000 per trajectory file). Supervised Projective Learning for Orthogonal Completeness (SPLOC) software for analysis as performed in the publication can be found at: https://github.com/BioMolecularPhysicsGroup-UNCC/MachineLearning/tree/master/SPLOC