Thank you, Tom, Eric, and Elaine, for your time and guidance!
I agree that it is a difficult statistical problem. In fact, I have been in contact with chemical engineers that also come to this position. As Tom suggested, I too am not sure if this exercise that I am developing will achieve what I hope with my students but I am thinking that it will produce some valuable discussion and I can only hope that it moves us to more meaningful discussion. Should I have some initial successes here, I may move into learning LAMMPS, etc. for a more thoroughly representative model.
For now though, the convex hull solution that Tom suggested is nearly exactly what I was envisioning in my original question. I have found the python script and here are some of my initial observations. After selecting atoms of interest, I use the command "convexhul sel sharp true eachChain true" to generate the desired surfaces. In some cases, no volume measurement is given (e.g. 0.000 (10 holes)) but the area is still calculated. In other cases, a value is produced (e.g. 0.000007629 (8 holes)) and the are is still calculated. I have not figured out a rhyme or reason. (As an aside, this process is far superior than a solution that I was trying in parallel, with blender and making faces, then calculating volumes.) So, can you provide any advice here as I am not sure what the units of that volume are nor what "holes" is referring to? Perhaps this is the sequence that I am selecting the atoms in.
I have also spent a little bit of time working through the original suggestions. Here are my observations:
So, here is where I have currently landed: 1) I open the ORCA output file and inspect the convergences steps which I will encourage my students to view as the liquid water moving around. 2) I switch the convergence tab next to the energy in the plot that pops up. Holding the down arrow key, we are able to watch the water molecules move around throughout the convergence. 3) I select the appropriate water molecules of interest and use "surface sel enclose sel." 4) volume measurements are meaningful 5) with the surface selected, I hold the arrow keys in the convergence tab and I was very presently surprised to see that the surface moved with the changing structure, and the volume change updates automatically!
Another observation about the convex hulls that are generated: Once these are produced, they are fixed in space as I move through the various convergence steps. This has the advantage of being able to see where the water molecules were originally located. I wonder thought: is it trivial to have the option of leaving then convex hull in place (as I described above) and also have the option of tracking those atoms over the throughout convergence?
Thank you again!
Dan
-----Original Message----- From: Tom Goddard goddard@sonic.net Sent: Wednesday, December 21, 2022 7:59 PM To: ChimeraX Users Help chimerax-users@cgl.ucsf.edu Cc: Harrison, Daniel P harrisondp@vmi.edu Subject: Re: [chimerax-users] using atomic coordinates to measure volume
Hi Dan,
Quantifying the packing density of molecules in a liquid mixture looks like a difficult statistical problem. The two things I can suggest to get some idea of density is to measure distances (water-water, ethanol-water, ethanol-ethanol). ChimeraX docs explain how. And on the ChimeraX recipes web site is a Python script that can make a convex hull surface using the center points of specified atoms. And ChimeraX can compute the volume in that surface. So you could for instance look at volumes of tetrahedra define by 4 packed waters or water-ethanol combinations.
Not sure how any of this will elucidate partial molar volume, where mixing water and ethanol occupies less volume then the water and ethanol separately.
Tom
On Dec 21, 2022, at 11:15 AM, Eric Pettersen via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:
I think you could select the water molecules of interest and the solute and then "surface sel enclose sel" to put a surface around them and then "measure volume" to report the volume enclosed by the surface. If the waters and solute are in separate models, I think you would need to use the "combine" command to put them into one model because I don't think a surface computation works across multiple models.
--Eric
Eric Pettersen UCSF Computer Graphics Lab
On Dec 21, 2022, at 4:59 AM, Harrison, Daniel P via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:
Hello Tom,
Ah yes. The second case of liquid state calculation with random packing is closer to what I am working with.
Here is my current process:
- Assuming the crystal structure of ice is a good model for liquid water, I took the crystal structure of ice and symmetry generated a 60-water molecule xyz coordinate file.
- I took the xyz file to perform a very low level, PM3, optimization calculation with ORCA and the structure is relatively stable. (Note I will do higher level QM/MM calculations once I have fleshed out this method further).
- Next, I replaced a water molecule in the center of this 60-molecule cluster with either methanol, ethanol, isopropanol, and tbutanol. With these I performed low-level optimization calculations.
- Pleasingly, the water molecules reorient themselves and collapse
around the alcohols to varying degrees. (As an aside, the optimization trajectory steps produce a nice "video" that one could interpret as liquid structure in motion.) 5) My objective is to quantify this collapse by having my students select some number oxygen atoms around the area of interest (i.e. around the alcohol) and calculate the volume made up by this "random" orientation of liquid water molecules with UCS ChimeraX. This is where I am failing.
Does this make sense? Do you have any suggestions either capturing the information I'm looking to acquire or on my process?
Thank you greatly for your consideration and help!
Dan
P.S. In the long run, I am hoping to add another dimension to this project by adding a virtual reality element with ChimeraX but that is down the road.
-----Original Message----- From: Tom Goddard goddard@sonic.net Sent: Tuesday, December 20, 2022 4:32 PM To: Harrison, Daniel P harrisondp@vmi.edu Cc: ChimeraX Users Help chimerax-users@cgl.ucsf.edu Subject: Re: [chimerax-users] using atomic coordinates to measure volume
Hi Daniel,
Is the idea for your partial molar volume calculation that you have crystals and want to find the fraction of the volume of a crystal unit cell that is occupied by the atoms? Or are you trying to do a liquid state calculation where the atom positions would be packed randomly?
Tom
On Dec 20, 2022, at 1:23 PM, Harrison, Daniel P via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:
Thank you, Elaine!
I will take a look through your suggestions and let you know if I have any further questions.
My "maximum" volume, I was envisioning that a number of plane-faces would need to be generated for the calculations and envisioned that that can be accomplished any number of ways. Maximum was referring to the space inside of the planes made by the largest polyhedron possible from all of these planes, though I admit that I am not an expert in this area.
Best wishes and thank you again, Dan
-----Original Message----- From: Elaine Meng meng@cgl.ucsf.edu Sent: Monday, December 19, 2022 11:42 AM To: Harrison, Daniel P harrisondp@vmi.edu Cc: chimerax-users@cgl.ucsf.edu Subject: Re: [chimerax-users] using atomic coordinates to measure volume
Hi Dan, I don't understand what you mean by "maximum" volume. You can calculate a molecular surface that encloses the atoms (solvent-excluded surface based on their VDW radii) and then measure the volume enclosed by that surface.
"surface" command and solvent-excluded surface diagram https://rbvi.ucsf.edu/chimerax/docs/user/commands/surface.html
For example, if the set of atoms is selected, command
surface sel enclose sel measure volume #1
...assuming the structure and surface are opened as model #1. If some other model number, substitute. There is also a graphical interface to "measure volume," menu: Tools... Volume Data... Measure Volume and Area. For atomic models in angstrom coordinates, the reported volume will be angstroms cubed.
<https://rbvi.ucsf.edu/chimerax/docs/user/commands/measure.html#volu me
https://rbvi.ucsf.edu/chimerax/docs/user/tools/measurevolume.html
ChimeraX only measures volume inside an enclosed surface. It does not measure the volume of a pocket that is open to the exterior.
For pocket volumes, especially if open to exterior, you could try CASTp server, or 3Vee server to give a map that you could open in ChimeraX, show as isosurface, and then measure volume as described above. http://sts.bioe.uic.edu/castp/calculation.html http://3vee.molmovdb.org/
I hope this helps, Elaine
Elaine C. Meng, Ph.D. UCSF Chimera(X) team Department of Pharmaceutical Chemistry University of California, San Francisco
On Dec 18, 2022, at 6:02 PM, Harrison, Daniel P via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:
Hello:
I would like to calculate the maximum volume made by selection of atoms in a given structure. Can you provide guidance on how to go about this process? It seems like there should be a simple procedure in ChimeraX but I have failed to locate it.
My specific problem: I would like to calculate the volume of the void created by ice under various conditions so that my physical chemistry students can quantify and visually understand partial molal volume non-ideality.
I would greatly appreciate any assistance you can provide.
Dan
ChimeraX-users mailing list ChimeraX-users@cgl.ucsf.edu Manage subscription: https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users
ChimeraX-users mailing list ChimeraX-users@cgl.ucsf.edu Manage subscription: https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users