Molecular dynamics and AlphaFold dataset of fluorinated protein variants

This dataset contains AlphaFold predictions and molecular dynamics (MD) trajectories of 50 fluorinated protein variants. Experimentally measured fluorine NMR chemical shifts are provided in the CSV file. Unless otherwise stated, secondary shifts were calculated using a random coil / reference chemical shift of -61.82 ppm. Proteins were labelled with 4-trifluoromethyl-L-phenylalanine (tfmF), modelled as a tyrosine residue in AlphaFold and explicitly in MD simulations with the CHARMM36m (C36m) and ff15ipq force fields. For each variant, ColabFold and AlphaFold 3 predictions (five models) are included. For MD, each variant contains three independent trajectories lasting one microsecond (150 microseconds total over the entire dataset per force field). Each simulation folder contains distance (in Å) and angle (in degrees) descriptors, which quantify the interaction of tfmF with nearby aromatic residues to predict ring current effects. Each folder also contains the solvent-accessible surface area of the trifluoromethyl group calculated over each trajectory, and the RMSD with respect to the starting structure. General template folders for each force field (and an additional template for HRAS due to custom GDP/ligand parameters) are also provided to allow for reproduction of the setup and production simulations for fluorinated protein variants. 

Excel files summarising the average distance and angle descriptors for each variant and their experimental chemical shifts are provided (averaged over all AlphaFold models or across all three trajectories, discarding the first 200 ns as equilibration). 

The dataset also includes six long MD trajectories (lasting 20 microseconds) of two HemK variants (N-terminal domain, residues 1-73) simulated with the DES-Amber protein force field, as proof-of-concept of fluorinated probes (38tfmF for HemK) reporting on mutation-induced conformational changes. Wild-type (K12 strain) and a mutant (I26V/R34K/Q46R) were simulated.