FEP+ Checklist

Before applying FEP+ to a project, it is helpful to assess the suitability of this calculation to your research. Use the checklist to gauge whether your project meets sufficient criteria to move forward with FEP+. Having some missing checkpoints can be tolerated, but a decreasing amount of checkpoints correlates to more difficulty in achieving success with applying FEP+ to a research program.

Selection of target and compounds

  • For retrospective testing, include some compounds that have known binding affinity from a reliable biological assay

  • There is at least one high-quality crystallographic structure that corresponds to the biological assay

    → If a structure from a homology model has to be used, make sure the structure is validated through retrospective FEP+ calculations

  • The location of the binding site is well established and the binding mode of the ligands is conserved across the series

  • Any idea compounds explored are structurally-similar to one another and to known compounds

FEP+ preparation

Protein preparation:

  • Run the Protein Preparation Wizard on the complex. The protonation state of the residues in the binding pocket may be adjusted from a consideration of the interactions with ligand and waters.

Ligand preparation:

  • Run LigPrep to enumerate stereoisomers, tautomers and identify correct protonation state.

  • Run Force Field Builder to ensure that all ligand torsions are defined and update f16_utt.dat file

    → For fused rings, even torsions that show up as defined may need to be refit taking into account the fused nature of the ring. This can be done with FF Builder in the Advanced mode.

FEP map input:

  • Create file/project containing the protein followed by all the ligands for the map

  • Align all ligands to the crystal structure ligand

    → NOTE: good ligand alignment is very important for correct atom mapping

    → After alignment, carefully check if ligands clash with protein

FEP map and files preparation:

  • Add binding affinity data for known compounds

  • Ensure reasonable edges, add/remove if applicable

    → If possible link novel compounds to known compounds (either directly or through an intermediate compound)

  • Add intermediates, if applicable

FEP+ job submission on a cluster:

  • Either submit jobs from the GUI, or write out files and copy to cluster to submit calculations

FEP+ analysis

  • Copy the jobname_out.fmp and jobname_out.fmpdb to your desktop

  • Import the jobname_out.fmp into the FEP+ panel

  • Ensure low hysteresis, omit edges, if applicable and save updated jobname_updated_out.fmp file

  • If the calculations contained compounds with binding affinity data, open the Correlation Plot (FEP+) Panel from the FEP+ Panel to view the statistics (ensure low MUE and RMSE).

  • Check the Simulation Interaction Diagram (SID) reports or trajectories for further debugging or understanding protein-ligand interactions, ligand flexibility etc.

FEP+ troubleshooting

  • Check torsion sampling in the SID reports. For example, are asymmetrically substituted rings well sampled in both the complex and solvent calculations?

    • Try applying custom cores to increase the sampling

  • Check the Convergence Plots and Sliding Window (small amplitude, sinusoidal shape)

    • Try increasing the length of the simulations (i.e., 10 ns)
    • Ensure that there is adequate ligand sampling

  • Check for water sampling involving buried crystal structure waters (look at the trajectories to see if the included waters exchange with bulk solvent)

  • Write out the PDF for the SID report

    • Check that there is good replica exchange for problematic edges
    • If there is not, try increasing the number of lambda windows and/or increasing the sampling

For more information, see Troubleshooting Common Issues.