FAQ: Preparing and Debugging OpenFE Simulations

This FAQ contains a list of issues you may encounter when preparing and running simulations with OpenFE. This list will be updated as feedback is provided by benchmark partners. If you find your issue is not currently addressed then please raise an issue on the industry benchmarking repository.

1. Ligand Alignment

Ligands in a series should be pre-aligned when passed to OpenFE tools. This means that regions of the ligand that should be part of a common core, should have the same coordinates.

In these benchmarks we will use the Kartograf atom mapper. Any ligand regions which do not align in 3D space will be unmapped and treated as unique atoms in the alchemical topology.

2. Protein Structures Ready for Simulation

OpenFE does not run a very long equilibration during the RBFE calculation. As such, a protein structure with clashes will likely fail with the Protocol reporting “NaN” errors.

You should make sure that your protein structure is ready for MD simulation before passing it to OpenFE.

3. All My Complex Edges Fail But Not The Ligands

This indicates a likely issue with the protein structure, either a clash with the ligand, or a clash within the protein structure itself.

To check this we suggest running a conventional MD simulation and seeing if the system remains stable. You can use the OpenFE MD Protocol for this purpose.

4. My Ligand Transformations Are Failing

This likely indicates a bad mapping.

We would encourage you to visualize your mappings. If you see unusual atom assignments, please get in touch with the OpenFE team and we will guide you through how to assign alternative mappings.