Short Trajectory MD with Analysis for FE-NES
In this tutorial, we will run the Short Trajectory MD with Analysis and FE-NES for Binding Affinity Prediction Floes with six selected ligands of the Tyk2 receptor.
Run the Short Trajectory MD with Analysis Floe
For more detailed information on how to run the Short Trajectory MD with Analysis Floe, please refer to the Short Trajectory MD with Analysis How-To Guide.
Floe Inputs
The ligand and protein input datasets used in this tutorial are:
Running the Floe
On the Orion Floe page, locate the Short Trajectory MD with Analysis Floe. You can either find it from the list of floes in the MD Affinity package or search for STMD or STMDA in the search bar. Click “Launch Floe” to open the Job Form. In this tutorial, keep the default parameter values. For more information about the parameters, please refer to the Short Trajectory MD with Analysis How-To Guide. Specify the Protein Input Dataset and Ligand Input Dataset. We suggest modifying the output dataset names so that you can recognize your data later. Here, the prefix Tyk2_6ligs_ is added to the default output dataset names.
Figure 1. The Job Form for the Short Trajectory MD with Analysis Floe using the Tyk2_6ligs input dataset.
Hit the “Start Job” button to begin the floe.
Understanding Results and Output Datasets
The results from the STMD Floe can be accessed through the job output in the Jobs tab on the Floe page. In the Jobs tab, select your STMD job to reach the job details. The Details panel includes information about the run. The Reports section includes any reports generated by the floe. The MD Trajectory Analysis Floe Report is shown in Figure 2.
Figure 2. The MD Trajectory Analysis Floe Report for the Short Trajectory MD with Analysis Floe using the Tyk2_6ligs input dataset.
The Results section includes output datasets and collections. Two datasets are written: the optional FreeForm output dataset (Tyk2_6ligs_freeform_output) and the equilibrium output dataset (Tyk2_6ligs_md_output_with_analysis). The equilibrium output dataset will be used as the Existing End-State Simulations input for the FE-NES for Binding Affinity Prediction Floe later in this tutorial.
These datasets can be activated for further analysis in the 3D & Analyze page.
Run the FE-NES for Binding Affinity Prediction Floe
For more detailed information on how to run the FE-NES for Binding Affinity Prediction Floe, please refer to the FE-NES tutorial.
Floe Inputs
The STMD Floe produced the Tyk2_6ligs_md_output_with_analysis. This is the minimal input for the FE-NES for Binding Affinity Prediction Floe.
In addition to the collected input datasets, you can optionally provide a file containing the experimental ligand affinities for the comparison between experimental and calculated affinities:
Running the Floe
On the Orion Floe page, locate the FE-NES for Binding Affinity Prediction Floe. You can either find it from the list of floes in the MD Affinity package or search for “FE-NES” in the search bar. Click “Launch Floe” to open the Job Form. In this tutorial, keep the default parameter values. For more information about the parameters, please refer to the FE-NES tutorial. Specify the inputs and the Ligand Affinity Experimental File.
Figure 3. The Job Form for the FE-NES for Binding Affinity Prediction Floe.
Hit the “Start Job” button to begin the floe.
Understanding Results and Output Datasets
The results from the FE-NES Floe can be accessed through the job output in the Jobs tab on the Floe page. In the Jobs tab, select your FE-NES job to open the job details. The Details panel contains information about the run. The Reports section lists the reports generated by the Floe. This run produces three reports: the Relative Affinity Predictions Report, the Edge Map Diagram with Predictions, and the Affinity Predictions Report. The Relative Affinity Predictions Report is shown in Figure 4.
Figure 4. The Relative Affinity Predictions Report for the FE-NES for Binding Affinity Prediction Floe using the Tyk2_6ligs input dataset.
The Relative Affinity Predictions Report contains several tiles. Each tile shows an edge and its predicted RBFE, calculated using the Bennett acceptance ratio method for FE-NES.
An example of the Edge Map Diagram with Predictions is shown in Figure 5.
Figure 5. The interactive Edge Map Diagram with Predictions.
Of the seven edges connecting the six Tyk2 ligands, one is classified as having medium trustworthiness. The Edge Map Diagram with Predictions is interactive. Click an edge to open its individual edge report. For example, click the edge connecting ejm_42 and ejm_54 to open a report similar to Figure 6. This report shows poor forward and reverse work distribution overlap for both the bound and unbound calculations, indicating that the results are less trustworthy.
Figure 6. An example of a mediocre RBFE calculation.
The Affinity Predictions Report contains two main sections. In the first section, experimental and predicted affinities are compared in a graph of \(\Delta G_{Experimental}\) versus \(\Delta G_{Predicted}\). This graph is available when an experimental affinity file has been provided and the ligand edge map is sufficiently connected to support affinity predictions.
Figure 7. The experimental versus predicted affinities for Tyk2_6ligs.
In addition, the graph data are provided in tables, along with statistical metrics such as correlation metrics and linear models. The second section compares the experimental and predicted relative binding affinities using a graph of \(\Delta\Delta G_{Experimental}\) versus \(\Delta\Delta G_{Predicted}\), together with tables of the statistical metrics.
Figure 8. The experimental versus predicted relative binding affinities for Tyk2_6ligs.
In addition to the FE-NES and Affinity Floe Reports, the results are uploaded
to Orion as .tar.gz files that you can download locally. These archives
contain the FE-NES and affinity data in .csv format for further analysis,
along with the .html reports.