Potency Predictions for RNA-Binding Small Molecules Using RB-FEP

Tutorial Created with Software Release: 2025-3
Topics: Free Energy Perturbation (FEP), Small Molecule Drug Discovery
Products Used: FEP+, Maestro

Tutorial files

3.6 MB
This tutorial is written for use with a 3-button mouse with a scroll wheel.
Words found in the Glossary of Terms are shown like this: Workspacethe 3D display area in the center of the main window, where molecular structures are displayed

 

Tip: You can hover over a glossary term to display its definition. You can click on an image to expand it in the page.
Abstract:

 

This tutorial shows you how to run and analyze relative binding free energy perturbation calculations for a series of flavin analogs binding to an RNA receptor using FEP+.

 

Tutorial Content

  1. Introduction RB-FEP

  1. Creating Projects and Importing Structures

  1. Running FEP+ Calculations

  1. Analyzing the Results of the FEP+ Calculation

  1. Conclusion and References

  1. Glossary of Terms

1. Introduction to RB-FEP

Accurate prediction of ligand-receptor binding affinities is a critical part of hit-to-lead and lead optimization stages in the structure-based drug discovery process. Computational prediction of ligand binding affinity involves initial determination of the binding mode and subsequent evaluation of strength of the binding interactions. Binding affinity can be predicted using free energy perturbation (FEP) calculations, which is a rigorous first-principles or physics-based method that estimates the difference in binding affinity (∆∆G) between distinct chemical compounds. Schrödinger’s implementation of FEP is known as FEP+.

The free energy perturbation approach can be used in different ways to answer various scientific questions. In this tutorial, you will use relative binding free energy perturbation (RB-FEP) method that allows you to calculate a ligand’s predicted binding affinity in relation to that of a similar compound. It uses molecular dynamics and statistical mechanics to compute free energy differences between congeneric molecules. For an accessible introduction to FEP+, see the following videos on the fundamental concepts and the value of FEP predictions.

This tutorial uses the prepared 6DN3 RNA structure, which is a flavin mononucleotide (FMN) riboswitch bound to a flavin analog ligand BRX1555. If you are following along with a different structure, please make sure that your structure is prepared as shown in the Preparing Nucleic Acid Structures tutorial and the Preparing Protein and Ligand Structures for FEP+ documentation page, and fulfils the Requirements on Experimental Structures and Binding Affinity Data for FEP+.

In this tutorial, you will learn how to run and analyze relative binding free energy perturbation calculations for a series of flavin analogs binding to the 6DN3 RNA structure using the FEP+ panel. These flavin analogs constitute a congeneric series of ligands i.e. they have a common core scaffold (the isoalloxazine ring system) but possess different substituents at one or more positions.

2. Creating Projects and Importing Structures

At the start of the session, change the file path to your chosen Working Directorythe location where files are saved in  Maestro to make file navigation easier. Each session in Maestro begins with a default Scratch Projecta temporary project in which work is not saved, closing a scratch project removes all current work and begins a new scratch project, which is not saved. A Maestro project stores all your data and has a .prj extension. A project may contain numerous entries corresponding to imported structures, as well as the output of modeling-related tasks. Once a project is saved, the project is automatically saved each time a change is made.

Structures can be built in Maestro or can be imported using File > Import Structures (or drag-and-dropped), and are added to the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion and Project Tabledisplays the contents of a project and is also an interface for performing operations on selected entries, viewing properties, and organizing structures and data. The Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion is located to the left of the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed. The Project Tabledisplays the contents of a project and is also an interface for performing operations on selected entries, viewing properties, and organizing structures and data can be accessed by Ctrl+T (Cmd+T) or Window > Project Table if you would like to see an expanded view of your project data.

  1. Double-click the Maestro icon.

Figure 2-1. Change Working Directory option.

  1. Go to File > Change Working Directory.
  2. Find your directory, and click Choose.
  3. Pre-generated files are included for running jobs or examining output. Download the zip file here: https://www.schrodinger.com/sites/default/files/s3/release/current/Tutorials/zip/fep_nucleic_acid.zip
  4. After downloading the zip file, unzip the contents in your Working Directorythe location where files are saved for ease of access throughout the tutorial.

Figure 2-2. The Open Project panel.

  1. Go to File > Open Project.
  2. Navigate to and choose RNA_FEP_tutorial.prjzip in your Working Directorythe location where files are saved.
  3. Click Open.
    • Structures are added to the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion.
  4. In the Save scratch project warning box, click OK.

Figure 2-3. Saving the project in the Save Project panel.

  1. Go to File > Save Project As.
  2. Change the File name to RNA_RB-FEP.
  3. Click Save.
    • The project is now named RNA_RB-FEP.prj.

3. Running FEP+ Calculations

In this section, you will learn how to perform RB-FEP calculations to accurately quantify the binding affinities. It is important to note that running FEP+ calculations can be computationally demanding, requiring significant time and resources, including access to a compatible GPU. In order to save time, all the input files for running the calculation and examining the output have been included in the zip archive of the tutorial. For an in-depth guide to setting up RB-FEP calculations, also see the BACE1 Inhibitor Design Using Free Energy Perturbation tutorial.

The quality of FEP+ results strongly depends on getting the initial ligand pose right. There is no universally optimal protocol for determining the best starting poses for FEP+, as differences in ligand chemistry and pocket geometry between systems bring distinct challenges. We recommend starting by using the new FEP+ Pose Builder (Beta), as it has outperformed other methods on benchmarks, and incorporates many of the best practices shown in this tutorial that would otherwise have to be done manually. The Pose Generation for FEP documentation page provides a comprehensive overview of alternative approaches. To save time, the prepared and aligned ligands are included in the zip archive. If you are following along with a different set of compounds, make sure that you have prepared them using the LigPrep panel. For more information, please refer to Preparing Protein and Ligand Structures for FEP+ and the LigPrep panel documentation.

3.1 Generate FEP+ ready poses

In this tutorial, you will use the new FEP+ Pose Builder (Beta), which incorporates our best practice recommendations and has outperformed other pose generation methods on benchmarks.

Figure 3-1. The FEP+ Pose Builder in Tasks.

  1. Go to Tasks > Browse > FEP+ > FEP+ Pose Builder.
    • FEP+ Pose Builder (Beta) panel opens.

Figure 3-2. Specifying the ligands for generating poses.

The FEP+ Pose Builder guides you through the necessary settings top to bottom.

 

  1. Expand and select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entries (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries ligprep_flavin_analogs-out group in the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion.
  2. In the Specify Ligands section, for Ligands to process, choose Project Table (15 selected).
  3. Click Load.  

As the ligands have been prepared using the LigPrep panel, you should disable the preparation step.

 

  1. Toggle off Prepare Ligands and Filter Prepared Ligands sections.

Figure 3-3. Specifying the reference ligand for alignment.

You will use the co-crystal ligand pose as the reference for alignment.

 

  1. In the ligprep_flavin_analogs-out group in the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion, select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entries (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries BRX1555-6DN3_natural.
  2. In the Align Ligands section in the panel, for Reference Ligands, choose Project Table (1 selected) and click Load.
  3. For Reference selection, choose Fixed Reference.

Figure 3-4. Specifying the receptor for resolving clashes.

 

Finally, it’s time to specify the receptor structure to be used. Note that the ligands will be aligned to the reference ligand first and the receptor pocket is only used to detect and resolve clashes.

 

  1. Expand 6DN3 - prepared_split_by_structure group and select(1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entries (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries 6DN3 - prepared_RNA.
  2. In the Resolve Clashes section in the panel, for Receptor, choose Project Table (1 selected) and click Load.
  3. Change the Job name to FepPoseBuilder_flavin_analogs.
  4. Optional: Click Run.
    • This job takes ~3 minutes.
    • To save time, you will use the pre-generated result file for performing FEP+ calculations.
  5. Close the FEP+ Pose Builder (Beta) panel.

3.2 Set up the RB-FEP calculation

Figure 3-5. Opening the FEP+ panel.

  1. Go to Tasks > Browse > Free Energy Perturbation > FEP+.
    • The FEP+ panel opens.

Figure 3-6. The FEP+ panel.

For this tutorial, we have prepared an fmp file which already includes the experimental reference data. For instructions on how to load affinity data from a file, see the BACE1 Inhibitor Design Using Free Energy Perturbation tutorial.

 

  1. For Import Structures or perturbation map from, choose File.
  2. Next to File name, click Browse.
    • The Select Input File dialog box opens.
  3. Select FEP_Poses_flavin_analogs.fmp.
  4. Click Open.
    • A new group titled FEP+: FEP_Poses_flavin_analogs.fmp is added to the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion.
    • The receptor is includedthe entry is represented in the Workspace, the circle in the In column is blue in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.

 

Note: In the panel, a warning symbol is displayed next to “Receptor” and “Total ligands”. Clicking the warning symbol next to “Receptor” opens the Protein Reliability Report, which is not applicable as the receptor is an RNA.

Further, hovering over the warning symbol next to “Total ligands” reveals the number of ligands with missing torsion parameters and steric clashes with the receptor.

 

  1. Click Next.
    • The FEP+ Review panel opens in the Overview tab displaying information about the quality of each ligand.

Figure 3-7. The Overview tab in the FEP+ panel.

  1. Hover over the warning symbol to learn about the corresponding issue.

 

Note: The green check mark in the Quality column indicates there are no issues with the ligand structure. The warning symbol indicates that the ligand has steric clashes with the receptor or crystal waters, is not docked into the receptor, has missing atoms or does not have explicit torsional parameters in the force field.

 

In this example, 9 ligands have missing torsion parameters and 1 ligand has steric clashes with the receptor.

The missing torsion parameters can be generated by running the Force Field Builder during the FEP+ calculations as shown below. Regarding the steric clash - the alignment used in this tutorial was generated prior to the introduction of the FEP+ Pose Builder panel described above. Hence during Ligand Alignment the receptor was not included and one of the ligands, BRX1675, resulted in steric clashes with the RNA bases. In this particular case, FEP+ was able to resolve the clashes as it runs an initial minimization stage in the beginning of the simulations. As we now have the FEP+ Pose Builder, which is able to take into account the receptor structure, we recommend using that workflow to minimize steric clashes before initiating the FEP workflow.

Figure 3-8. Generating the perturbation map.

As FEP+ calculates the relative binding free energy differences between pairs of ligands, for which you need to generate edges in the Map tab.

 

  1. Go to the Map tab.
  2. Click Generate Map.
    • Map Options dialog box opens.
  3. Keeping all the settings to default, click Generate Map.
    • This takes a few seconds.
  4. Click Settings in the toolbar at the bottom of the panel.
    • The FEP+ Panel Settings dialog box opens.

The more similar two ligands are, the more accurate the free energy prediction between them is. In cases where ligands are too dissimilar to directly draw an edge between them, it can be helpful to create an intermediate ligand between them. Each edge in the map increases the computational effort for the calculation, but increases its accuracy because the cycle closure correction can be applied. Creating a good FEP map is then a matter of balancing computational costs and accuracy. For more in-depth instructions on constructing FEP+ maps including a guided hands-on example, see the Free Energy Calculations for Drug Design with FEP+ online course.

Figure 3-9. The FEP+ Panel settings dialog box with modified settings.

 

You can run the parametrization of the missing ligand torsions as part of the FEP+ job.

 

  1. In the Force Field section, enable Generate missing parameters with Force Field Builder.  

Since RNA receptors are more flexible, our current best practice for obtaining reliable results is increasing the simulation time to 20 ns.

 

  1. In the Simulation Parameters section, set the simulation times to 20.00 ns.
  2. Click match.
    • This matches the simulation time for the Complex, Solvent and Vacuum legs of the simulation.

 

Adding salt to the system is essential for the stability of the highly-charged nucleic acid system. Note that this will also add neutralizing counterions to the system.

 

  1. In the System Builder section, check Add salt (0.150 M).
  2. Click OK.

Figure 3-10. FEP+ job set up.

  1. Change the Job name to 6DN3_RNA_fep.
  2. Click the Job settings cog.
  3. Select the appropriate hosts for running the Master Job, Force Field Builder Job and a GPU Host for the Subjob.
  4. Optional: Click Run.
    • The FEP+ jobs are time and resource intensive.
    • You can find the results of this simulation in the tutorial files.
  5. Close the FEP+ panel.

4. Analyzing the results of the FEP+ calculation

Note that these results may have been generated using a different software version than you are using, resulting in a slightly different map topology, starting ligand alignment, and affinity prediction.

Figure 4-1. Importing the FEP+ results.

The main results of an FEP+ job are contained in the <jobname>-out.fmp file, which you can load directly into the FEP+ panel.

 

  1. Go to Tasks > Browse > Free Energy Perturbation > FEP+.
  2. For Import Structures or perturbation map from, choose File.
  3. Next to File name, click Browse.
  4. Navigate to and select 6DN3_RNA_fep-out.fmp in your Working Directorythe location where files are saved.
  5. Click Open.
  6. Click Next.
    • The FEP+ Review panel opens in the Overview tab.

 

Note: The Quality column in the overview again shows the warning about missing torsion parameters for a few of the ligands, but this is just because the example output does not include the custom parameters that were generated as part of the FEP+ job. These results were obtained on fully parametrized ligands. See this KB article on how to merge the new custom force field parameters with your existing one.

The Overview tab summarizes the key results of the FEP+ calculation for each ligand: You can see the predicted binding affinity including error bars calculated from the cycle closure correction and compare it to the experimental binding affinity.

 

Some of the ligands in the data set are present with multiple protonation and/or tautomeric states, so you need to apply the Groups correction to the results. This will use Epik to determine the solvent pKas of different states of the same ligand and use those to determine the complex pKas and correct the predicted dGs using the method described in this publication.

Figure 4-2. Applying Groups correction to the results.

You can apply the Groups correction at a pH at which the experimental assay was run. The experimental assay for this ligand series was run at a pH of 8.0 (see publication), so this pH is used to estimate the ligand populations.

 

  1. Go to the Map tab.
  2. Click the Groups tab heading.
  3. For Use pH, set the value to 8.0.
  4. Click Auto-populate.
    • This takes ~2 minutes.
  5. For Show group information in, choose Map and other data displays.  

Note: See the FEP+ Panel – Map Tab documentation page for more information on what Groups corrections are and how they are calculated.

You can now go through the available statistics to determine how predictive the FEP+ model is on your system before using it prospectively.

Figure 4-3. The FEP+ Correlation Plot.

A scatterplot is a great way to visualize the correlation statistics between predictions and the experimental affinity data.

 

  1. Click on Plot at the bottom of the panel.
    • The Correlation Plot (FEP+) opens.

In the plot, you can hover over the points to see which ligands they represent. The X-axis shows the experimental affinities, while the Y-axis shows the FEP+ predictions. The dark gray band in the plot shows the area where prediction and reference are within 1 kcal/mol of each other, the light gray band represents an error between 1-2 kcal/mol.

 

On the right-hand side, you can see various statistical metrics for the data set. The root-mean-square error (RMSE) is a good measure of how well the predictions are with respect to the experimental affinities. Unlike the R2, the RMSEs are independent of the dynamic range of the dataset and hence are more reliable metrics for the goodness of the predictions. Our FEP+ best practice is to proceed with prospective FEP+ calculations for models showing less than ~1.3 kcal/mol in retrospective validations. For a more detailed discussion of these metrics, see the Correlation Plot (FEP+) panel documentation.

 

In this example, the model’s predictions show good correlation with the experimental reference - RMSEs for both edgewise and pairwise predictions are less than 1.3 kcal/mol.

Figure 4-4. The Analysis tab of the FEP+ panel.

The Analysis tab provides a deeper look at the data underlying the prediction, as well as some technical metrics that can be used to gauge simulation quality.

 

  1. Go to the Analysis tab.

 

Note: The Edge Analysis View…  tool here is not yet supported for nucleic acid receptors. One way to currently get some of the analysis from the Edge Analysis panel is to use the <jobname>-out.fmpdb file to visualize the edge trajectories.

The table in the analysis tab lists all edges from the FEP+ map, showing the predicted free energy difference between the corresponding ligand pair from that edge alone or taking the cycle closure correction into account. The Energy convergence, Ligand RMSD, REST exchange and cycle closure convergence columns group these technical metrics for simulation quality into the categories Good, Fair, or Bad. See the FEP+ Panel – Analysis tab panel documentation for a detailed explanation of all metrics in this table.

 

You should look out for bad edges, which may be due to low ligand similarity, a bad initial pose, or receptor reorganization. In this case, we have already removed all bad edges from the map to clean up the results. Note that a bad edge can impact the quality of the results for ligands not directly connected to that edge due to the cycle closure correction, so removing them from the map is a good practice.

Figure 4-5. The Analysis tab showing the filtered edges.

You can filter the table to show edges containing particular ligands based on their name or a substructure. Use this option to investigate whether the predictions of ligands connected to the X-ray structure ligands (6DN1, 6DN2, 6DN3) are predicted well .

 

  1. In the Search field, type 6DN.
    • The table gets updated.

 

Note: This filter applies to all other tabs in the FEP+ panel.

While some edges are predicted better than others, all of them show Good or Fair metrics in terms of their Energy and Cycle Closure convergence, Ligand RMSD and REST sampling. Hence, the FEP+ model should be sufficiently predictive for prospective use without further optimization.

 

When digging into the results for your own system, we recommend having a look at the FEP+ documentation pages on identifying problematic edges, investigating outliers, and troubleshooting other common issues.

5. Conclusion and References

In this tutorial, you learned how to run and analyze relative binding free energy perturbation calculations for a series of flavin analogs binding to the 6DN3 RNA structure using the FEP+ panel.

6. Glossary of Terms

Entries - a simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion

Included - the entry is represented in the Workspace, the circle in the In column is blue

Project Table - displays the contents of a project and is also an interface for performing operations on selected entries, viewing properties, and organizing structures and data

Recent actions - This is a list of your recent actions, which you can use to reopen a panel, displayed below the Browse row. (Right-click to delete.)

Scratch Project - a temporary project in which work is not saved, closing a scratch project removes all current work and begins a new scratch project

Selected - (1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entries (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries

Working Directory - the location where files are saved

Workspace - the 3D display area in the center of the main window, where molecular structures are displayed