-
Notifications
You must be signed in to change notification settings - Fork 113
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Strange RMSD values #105
Comments
Hi @livaschar , this is unwanted but expected behavior. It is because of the way the reordering is working. Trying to match atoms with atoms can be very approximate. Especially when the atom environments are very similar. For example, if you try
you will see when you try to reorder, it is based on distance between atom pairs, and might not give the correct match. Unfortunately, the Hungarian algorithm requires that the structures are already aligned to be a good fit. Which in your case happens after your initial alignment, which is why you see a good fit here.
There is a missing But in essence your problem is you are trying to match your atoms on a very symmetric molecule, which is super hard to do. You might want to checkout |
Thanks for your quick reply, @charnley! The thing that I don't fully get is what the coordinates of the 'calculate_rmsd -p' command are. Which method is used to reorder the molecule? Hungarian, distance, etc., or is it something else and I got it wrong? I will appreciate it if you notify us when qml is up and running. Thank you in advance. |
@livaschar The default reorder algorithm is the Hungarian one, see for instance |
@nbehrnd so if the coordinates printed by '-p' flag are the ones coming from the Hungarian algorithm, shouldn't the two cases
produce the same results, since the 'sp_modified.xyz' is just the final modified version of the 'sp.xyz'? |
@livaschar If I compare the structures As mentioned by @charnley, the restrain to use atomic positions (only) to describe a structure and highly symmetrical structures to compare with each other show a limitation of an algorithm / the implementation of an algorithm. For me, |
I calculate the RMSD value for two xyz files (opt.xyz and sp.xyz) with all possible methods, and the smallest value I take for a specific example is 2.1.
Then I run ''calculate_rmsd -p opt.xyz sp.xyz'' and take the modified sp coordinates (sp_modified.xyz). When calculating the rmsd between opt.xyz and sp_modified.xyz, I take a value of 0.4.
Why is that? Shouldn't the calculation produce the same results?
xyz.zip
The text was updated successfully, but these errors were encountered: