Crystal Structure Prediction

1. Double-click the Materials Science icon
   • (No icon? See Starting Maestro)

Figure 2-1. Change Working Directory option.

2. Go to File > Change Working Directory
3. Find your directory, and click Choose
4. Pre-generated files are included for running jobs or examining output. Download the zip file here: schrodinger.com/sites/default/files/s3/release/current/Tutorials/zip/csp.zip
5. 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. Save Project panel.

6. Go to File > Save Project As
7. Change the File name to csp_tutorial, click Save
   • The project is now named csp_tutorial.prj

Figure 2-3. Drawing acetaminophen.

We will proceed to draw the systems that we will use in the subsequent sections using the 2D Sketcher.

8. Go to Edit > 2D Sketcher
   • The 2D Sketcher panel opens
   • Learn more about the 2D Sketcher in the Introduction to Materials Science Maestro tutorial
9. Draw acetaminophen
   • Acetaminophen is shown in the figure
10. Click Save as New
11. Name the entry acetaminophen
12. Click OK

Figure 2-4. Drawing nitrofurantoin.

13. Click the piece of paper to clear the sketcher
14. Draw nitrofurantoin
15. Click Save as New
16. Name the entry nitrofurantoin
17. Click OK
18. Close the 2D Sketcher

Figure 3-1. Opening the Crystal Structure Prediction panel.

1. With the acetaminophen entry 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 Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries and includedthe entry is represented in the Workspace, the circle in the In column is blue, go to Tasks > Materials > Polymorph Prediction > Crystal Structure Prediction
   • The Crystal Structure Prediction panel opens

Figure 3-2. Running the Crystal Structure Prediction panel to generate conformers.

2. Change the Task to Generate conformers
   • Running a conformers job first is best practices to ensure the conformers are what you’d expect before moving to the much longer packing and ranking calculation
3. Change the Output to Z’=1 crystals
4. Click Load to load the molecule
   • Acetaminophen will now be loaded into the panel
5. Uncheck Generate missing parameters
   • In this example, the Generate missing parameters option is removed because no torsions are missing. However, if the molecule has missing torsions, users are recommended to use the Force Field Builder panel to generate customized parameters, save the customized parameters, and then select the folder with the customized parameters via "Use customized version". Users have the option to select "Generate missing parameters," which will run a Force Field Builder job during the CSP confgen workflow. However, we do not recommend this approach. It is always best practice to run Force Field Builder separately and verify the quality of the parameters before starting a CSP job.
6. Change the Job name to csp_conformers
7. Adjust the job settings () as needed
   • This job requires a CPU host and should complete in a couple of minutes on 8 CPUs
   • Job Server is required for submission of Crystal Structure Prediction calculations
8. Click Run or import the output_confgen.maegz file from the provided tutorial files
   • The count of generated conformers may not match the number found in the supplied file
9. Close the Crystal Structure Prediction panel

Figure 3-3. Running the Crystal Structure Prediction panel on the acetaminophen conformers.

10. When the job is finished, a new entry titled output_confgen is automatically  included in the entry list
11. With the output_confgen group 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 Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries, use the Workflow Action Menu (WAM) button () or go to Tasks > Materials > Polymorph Prediction > Crystal Structure Prediction
    • The Crystal Structure Prediction panel opens
12. Change the Task to Perform packing search & ranking
13. Click Load to load the conformers
    • 6 conformers of acetaminophen will now be loaded into the panel
14. Change the Job name to csp_acetaminophen
15. Adjust the job settings () as needed
    • This job requires a CPU and GPU host. The job can be completed in five days using 32 CPUs and 1 GPU. However, this calculation runs thousands of subjobs. Specifying 10,000 CPUs is ideal but will use many resources at once. A balance needs to be achieved between wall time and the use of resources.
    • Job Server is required for submission of Crystal Structure Prediction calculations.
16. If you would like to run the job, click Run, otherwise, we will proceed with pre-generated results in the next section
17. Close the Crystal Structure Prediction panel

The calculation will pack the conformers into unit cells by using various space groups and then relaxing the structures and filtering out the highest-energy structures. The relaxation part does a series of relaxations, each of which is followed by a filter. The first uses molecular mechanics (MM) (cheapest), then machine learning force field (MLFF) (medium), then density functional theory (DFT) (most expensive).Checking the "include DFT ranking" box will return the DFT ranking results. Fewer crystals require a DFT run compared to those needing MM.

Figure 4-1. Opening the Crystal Structure Prediction Viewer panel.

1. Open the CSP Viewer from Tasks > Materials > Polymorph Prediction > Crystal Structure Prediction Analysis
   • The Crystal Structure Prediction Viewer panel opens
2. Click Load Structures and navigate to where you downloaded the tutorial files and choose Section_04 > output_dft_relaxation.maegz file
   • The results will populate into the viewer panel

Figure 4-2. Clicking and viewing the data points.

3. Click the data points to view information about the structure

Each dot corresponds to one crystal structure and appears as a new entry in the entry list and the packing arrangement is shown in the workspace

Figure 4-3. Viewing the results in data table form.

4. Go to the Data Table tab
5. Click Relative Energy to sort the table by relative energy

Scroll across the data table to view information about each crystal structure. There are two known experimental crystal structures and both are found here, highlighted in blue: one has a density of 1.382 and spacegroup Pbca (corresponding to the first row), while the other has a density of 1.349 and spacegroup P121/n1 (corresponding to the second row).

Figure 4-4. Exporting the conformers.

The conformers extracted from the Z'=1 results are required as input for the subsequent Z'=2 calculation workflow.

6. Click Export
7. Select Extract Conformers

Figure 4-5. Naming the conformer file.

8. Name the file conformers_extract
9. Click Save

Figure 4-6. The extracted conformers.

10. Close the Crystal Structure Prediction Viewer panel

The extracted conformers are added to the entry list.

Figure 5-1. Opening the Crystal Structure Prediction Viewer panel.

1. With the Z’=1 acetaminophen extracted conformers 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 Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries, go to Tasks > Materials > Polymorph Prediction > Crystal Structure Prediction
   • The Crystal Structure Prediction panel opens
2. Change the Task to Perform packing search & ranking
3. Change the Output to Z’=2 crystals
4. Click Load to load the conformers
   • 3 conformers of acetaminophen will now be loaded into the panel
5. Click Space Groups
   • The Advanced Settings panel will open and select space groups can be selected
6. Select Pca21 and Pna21
   • The sampling was limited to space groups with observed experimental structures to improve computational efficiency.
   • In prospective studies, all possible space groups should be thoroughly sampled
7. Click Save

Figure 5-2. Running the crystal structure prediction calculation.

8. Change the Job name to csp_acetaminophen_Z2
9. Adjust the job settings () as needed
   • This job requires a CPU and GPU host. This calculation runs thousands of subjobs. Specifying 10,000 CPUs is ideal but will use many resources at once. A balance needs to be achieved between wall time and the use of resources.
   • Job Server is required for submission of Crystal Structure Prediction calculations.
10. If you would like to run the job, click Run, otherwise, we will proceed with pre-generated results
11. Close the Crystal Structure Prediction panel

Figure 5-3. Viewing the crystal structure prediction results.

11. Open the CSP Viewer from Tasks > Materials > Polymorph Prediction > Crystal Structure Prediction Analysis
    • The Crystal Structure Prediction Viewer panel opens
12. Click Load Structures and navigate to where you downloaded the tutorial files and choose Section_05 > output_dft_relaxation.maegz file
    • The results will populate into the viewer panel
13. Click the data points to view information about the structure and view the system in the workspace

Each dot corresponds to one crystal structure and appears as a new entry in the entry list and the packing arrangement is shown in the workspace.

Figure 5-4. Sorting the results by relative energy.

14. Go to the Data Table tab
15. Click Relative Energy to sort the table by relative energy

Scroll across the data table to view information about each crystal structure. There are two known experimental crystal structures and both are found here and highlighted in blue.

Figure 6-1. Opening the Crystal Structure Prediction Viewer panel.

1. With the nitrofurantoin entry 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 Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries and includedthe entry is represented in the Workspace, the circle in the In column is blue, go to Tasks > Materials > Polymorph Prediction > Crystal Structure Prediction
   • The Crystal Structure Prediction panel opens
2. Change the Task to Generate conformers
3. Change the Output Z’=1 crystals
4. Click Load to load the molecule
   • Nitrofurantoin will now be loaded into the panel
5. Change the Job name to csp_conformers_nitrofurantoin
6. Adjust the job settings () as needed
   • This job requires a CPU host and should complete in about an hour on 8 CPUs
   • Job Server is required for submission of Crystal Structure Prediction calculations
7. Click Run or import the Section_06 > output_confgen.maegz file from the provided tutorial files
   • The count of generated conformers may not match the number found in the supplied file

Figure 6-2. Clicking and viewing the data points.

8. When the job is finished or imported, a new entry titled output_confgen is automatically included in the entry list
9. With the output_confgen group 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 Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries, use the Workflow Action Menu (WAM) button () or go to Tasks > Materials > Polymorph Prediction > Crystal Structure Prediction
   • The Crystal Structure Prediction panel opens
10. Change the Task to Perform packing search & ranking
11. Click Load to load the conformers
    • 27 conformers of nitrofurantoin will now be loaded into the panel
12. Change the Job name to csp_nitrofurantoin_Z1
13. Adjust the job settings () as needed
    • This job requires a CPU and GPU host. This calculation runs thousands of subjobs. Specifying 10,000 CPUs is ideal but will use many resources at once. A balance needs to be achieved between wall time and the use of resources.
14. Click Run or we will import pregenerated files in the next step

Figure 6-3. Viewing the crystal structure prediction results.

15. Open the CSP Viewer from Tasks > Materials > Polymorph Prediction > Crystal Structure Prediction Analysis
    • The Crystal Structure Prediction Viewer panel opens
16. Click Load Structures and navigate to where you downloaded the tutorial files and choose Section_06 > anhydrate > output_dft_relaxation.maegz file
    • The results will populate into the viewer panel
17. Click the data points to view information about the structure and view the system in the workspace

Each dot corresponds to one crystal structure and appears as a new entry in the entry list and the packing arrangement is shown in the workspace.

Figure 6-4. Viewing the data table.

18. Go to the Data Table tab
19. Click Relative Energy to sort the table by relative energy

Scroll across the data table to view information about each crystal structure. There are two known experimental crystal structures and both are found here and highlighted in blue. Form alpha is represented by the first highlighted row, and form beta is represented by the second.

20. Click Export
21. Select Extract Conformers
22. Name the file conformers_anhydrate
23. Click Save
    • An entry group with two entries has been added to the entry list
24. Close the Crystal Structure Prediction Viewer

Figure 6-5. Setting up a monohydrate calculation.

25. With the conformers_anhydrate 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 Entry List (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries, go to Tasks > Materials > Polymorph Prediction > Crystal Structure Prediction
    • The Crystal Structure Prediction panel opens
26. Change the Task to Perform packing search & ranking for monohydrate
27. Change the Output to Monohydrates/Solvates/Co-crystals
    • Solvent and coformer options are available once the output is changed. We will use the default water setting.
28. Click Load to load the conformers
    • 2 conformers of nitrofurantoin will now be loaded into the panel
29. Change the Job name to csp_nitrofurantoin_monohydrate
30. Adjust the job settings () as needed
    • This job requires a CPU and GPU host. This calculation runs thousands of subjobs. Specifying 10,000 CPUs is ideal but will use many resources at once. A balance needs to be achieved between wall time and the use of resources.
    • Job Server is required for submission of Crystal Structure Prediction calculations.
31. If you would like to run the job, click Run, otherwise, we will proceed with pre-generated results
32. Close the Crystal Structure Prediction panel

Figure 6-6. Viewing the monohydrate results.

16. Open the CSP Viewer from Tasks > Materials > Polymorph Prediction > Crystal Structure Prediction Analysis
    • The Crystal Structure Prediction Viewer panel opens
17. Click Load Structures and navigate to where you downloaded the tutorial files and choose Section_06 > monohydrate > output_dft_relaxation.maegz file
    • The results will populate into the viewer panel
18. Click the data points to view information about the structure and view the system in the workspace

Each dot corresponds to one crystal structure and appears as a new entry in the entry list and the packing arrangement is shown in the workspace.

Figure 6-7. Viewing the data table.

19. Go to the Data Table tab
20. Click on the rows to view the system in the workspace
    • You will see water molecules have been added to the system
    • Scroll across the data table to view information about each crystal structure

Scroll across the data table to view information about each crystal structure. There are two known experimental crystal structures and both are found here and highlighted in blue. Monohydrate II is represented by the first highlighted row, and monohydrate I is represented by the second.

21. Close the panel once finished