This chapter briefly describes how to set up some common types of molecular systems in QUANTA for use in CNX calculations. These examples can serve as templates for setting up your own work.
Setting up a CNX system for a simple protein
Setting up an CNX system for a protein, solvent, ion, and nucleic acid (see Setting up a system for a nucleic acid, solvent, ion, and ligand).
To complete these exercises, copy these files from the directory $QNT_ROOT/tutorial/files into your working directory:
This section describes how to set up a simple protein system and perform CNX calculations on it.
1. From the File menu, select Import. The Import File Librarian is displayed
2. Select the Protein Data Bank PDB file 1a2p.pdb to import.
3. This protein has some residues with multiple conformations (disorder). You will be asked if you wish to "Handle Alternate Location as Disorder?" Click Yes. Unless you have turned on the preference that causes the CRYST card to be read automatically, you are prompted whether you want to set up symmetry. Click Yes.
4. You are then asked to confirm that the spacegroup P32 is correct. Click Yes. (The reason for these prompts is that some third-party programs that output PDB files do not write the CRYST line correctly, so it is not always safe to read it blindly.)
5. The file 1a2p.msf is created and opened.
6. If you import a PDB or other format file that does not contain correct cell and spacegroup information, this can be added to the MSF using the menu function Edit | Symmetry->Define.
7. From the Edit menu, select Split and Clean. You should see the molecule colored according to whether it is protein, solvent, nucleic acid, or other. Note the three red atoms. These are zinc ions. Click Save to separate MSFs. You are warned that "Bonding exists between different classes." This warning is triggered because the zinc ions were assumed to be covalently bonded to adjacent protein residues. Click Continue to break these bonds. You now have three MSF files each containing one sort of molecule, enabling the most appropriate tools to be easily applied to each for correcting the structure.
8. Click Clean Options. This allows control of the hydrogen model to be used. Choose None and click OK.
9. Click Clean Protein. This adds hydrogens when that option is chosen, sets up standard names and atom types, and rebuilds any sidechains that have missing atoms. You are asked about the disorder in the protein. Choose to Keep all Disorder and click OK.
10. Click Clean Solvent. This command serves a similar function for solvent as Clean Protein does for protein.
11. Ligand molecules would normally fall within the class "Other". The Molecular Editor is available from the Split and Clean palette to fix bond order or any other details of ligand molecular structure. In this case, the zinc ions are the only member of the "Other" class, and do not require additional attention.
12. Click Finish to exit the Split and Clean function.
This section describes the steps for minimizing the protein using CNX:
1. From the Applications menu select CNX Interface. The CNX Interface palette is displayed.
2. From this palette, select Parameter Set.... Select Use Engh and Huber parameters and click OK.
3. From this palette, select Set-up Structure Factors.... The Select Intensities file librarian is displayed listing the structure factor files.
4. Select the intensity file 1a2p.cv. Set the High Resolution Limit to 1.5. Click Open.
5. Select the intensity file dummy.fob; make sure the weight factor WA is set to 0. Leave the other fields at their default values and click Open.
6. Select the Set CNX Host... tool. Select a local host from the scrolling list and click OK. The CNX Job Characteristics dialog is displayed; accept the default values and click OK.
7. Select the Refine Structure... tool. The Positional Refinement Settings dialog box is displayed. In this dialog box, of the three types of refinement that you can perform, select only Perform Minimization Refinement, with Number of Steps set to 20. Select Initial B-factor Correction: Anisotropic and Bulk Solvent Correction: Atom Mask. Select the action at the bottom of the page; Wait for Calculation to complete. Leave all other fields at their default values and click OK.
8. The CNX job runs on the selected host from the file cnx.bat. You will be notified in the Textport when cnx.bat is started and when it is finished.
9. Select Update Coordinates.... Select the refined coordinate file, 1a2p_Protein_refine.pdb. Click Open. The three zinc ions need new segment names. Choose 'D' when prompted regarding this issue for the first ion. Use 'E' and 'F' for the second and third ion, respectively, when prompted. In the 'Save data into msf files' dialog box, all three molecules should be checked (default). Choose to Create a new generation of the file and click OK. The coordinates on the screen and in the active MSFs are updated. If you watch the screen carefully you can see the atomic positions shift as the coordinates are updated.
Several files should be generated from this example:
The PSF (.xpsf) and parameters files (*.xprm) from Quanta, can be used with other CNX scripts, when combined with the input pdb file that is written by the CNX interface (*_input.pdb). Frequently, a topology file and a parameter file are needed to define a ligand for standalone CNX. Quanta can be very useful in this case. The Molecular Editor will write a CHARMm-style RTF file if requested in the dialog that occurs upon exit. That RTF can be converted to a CNX-style topology file (*.top) by using the appropriate command in the CNX Interface/CNX Utilities.. That topology file, combined with the parameter file (*.xprm) that the CNX Interface writes, supply the ligand information needed for refinement in standalone CNX.
You could encounter several problems when running this sort of calculation. This section discusses three common ones.
QUANTA probably cannot find where the executable for CNX is located. Check with your systems administrator or check your installation guide for the locations of cnx.bat and cnx.exe. cnx.bat should be in the $HYD_EXE directory, and it should be referenced in the $HYD_LIB/applcomm.db file for any machines on which you want to run CNX. The cnx.bat file is a simple C-shell script that executes cnx.exe.
This message box is displayed during PSF generation if any atoms have too many attached atoms. This could happen if the geometry is poor and the distance-search bonding algorithm is used.
This problem should not occur for a protein-only system, since the Special Protein algorithm should be used automatically.
Using the Edit | Bond Options function, you can also disable bonding between different segments. This is useful if you have solvent in close proximity to other structures within the same MSF.
Bad bonding can occur if connectivity has been incorrectly specified in an external file. In such cases, switch to an algorithm other than Stored Connectivity plus Distance Search. If bad bonds persist, manual editing is straightforward: the message box allows you to color the problematic bonds red and the rest of the molecule white. Choose this and focus in on the trouble spots. Use the Break Bond tool on the Modeling palette to eliminate the problem bonds. Then you can use File | Save As to make sure the corrected connectivity is saved for future use.
"Unknown atom types" refers to the situation where one or more atoms have not been assigned a valid CHARMm atom type. This should never happen for a protein that has been typed on import using the correct dictionary file or cleaned with the Split and Clean application. It can occur for non-standard groups. Methods for correcting the problem include:
This section presents an example of how to:
Import the pdb coordinate file 102d.pdb, and proceed with Setting up the structure for CNX as for the previous example up to and including Step 6.
1. From the Edit menu, select Split and Clean. You should see the molecule colored according to whether each molecule is solvent, nucleic acid, or other. Click Save to Separate MSFs.
2. Click Clean Options. This allows control of the hydrogen model to be used. Choose Polar and click OK.
3. Within the Split and Clean application, click Clean Solvent. This adds hydrogens to the solvent and corrects the atom types if necessary. For the purposes of crystallographic refinement, ideally the X-SOLVATE application should be used to place water molecules in the electron density. The CHARMm HBUILD function is also useful, since it checks for hydrogen-bonding partners as it places the hydrogens.
4. Click Clean Nucleics. This adds hydrogens and corrects atom types in the DNA.
5. Setting up a ligand generally requires manual intervention to obtain the best chemical representation. If the initial structure has very good geometry, then the Auto Edit tool on the Split and Clean palette may be able to perform the tasks automatically. In this case, click Molecular Editor and choose 102d_Other.msf. Select the Change Bond tool. A total of 16 bonds should be changed from single bonds. At both ends of the molecule, click each bond in the two rings three times to make these systems aromatic. Each bond should have bond order of 1.5 as indicated by double lines, one of which is solid and one of which is dashed. Also, set the bond between each of the four nitrogen atoms and its adjacent carbon atom to have bond order 1.5 (dashed and solid bonds). Click Add Polar Hydrogens.
6. Click Save and Exit. In the Save Options palette that appears, choose Reassign atom types and Reassign atom charge. Click OK. The Adjust Total Charge palette now opens. Enter the desired total charge as 2.04 and select Do not smooth. Click OK.
7. Click Finish to exit the Split and Clean application. To check that the structures are correctly set up, perform a PSF mode CHARMm calculation. The missing parameter box that opens indicates those parameters that were estimated within QUANTA. Quit this tool to finish the energy calculation.
This section describes the steps for minimizing the protein using CNX.
1. From the Applications menu select CNX Interface. The CNX interface palette is displayed.
2. From this palette, select Parameter Set.... Select Standard Parameters and click OK.
3. From this palette, select Set-up Structure Factors. The Select Intensities file librarian is displayed listing the structure factor files.
4. Select the reflection file 102d.cv. Set the High Resolution Limit to 2.2. Click Open.
5. Select Generate CNX PSF to write the necessary structure and parameter files for your molecular system.
6. Select the Set CNX Host... tool. Select a local host from the scrolling list and click OK. The CNX Job Characteristics dialog is displayed; accept the default values and click OK.
7. Select the Refine Structure... tool. The Positional Refinement Settings dialog box is displayed. In this dialog box, of the three types of refinement that you can perform, select only Perform Minimization Refinement, with Number of Steps set to 20. Select Initial B-factor Correction: Anisotropic and Bulk Solvent Correction: Atom Mask. Select Generate maps post-refinement. Select the action at the bottom of the page; Wait for Calculation to complete. Leave all other fields at their default values and click OK.
8. The CNX job runs on the selected host from the file cnx.bat. You will be notified in the Textport when cnx.bat is started and when it is finished.
9. Select Update Coordinates... Select the refined coordinate file, 102d_Nucleic_refine.pdb. Click Open. In the Save data into msf files dialog box, all three molecules should be checked (default). Choose to Create a new generation of the file and click OK. The coordinates on the screen and in the active MSFs are updated. If you watch the screen carefully you can see the atomic positions shift as the coordinates are updated.
Several files should be generated from this example: