Re: [Jmol-users] Color structural alignment by deviation "RMSD"?

[Theis, K. W. <kt...@we...>]

Eric,

A complete description of a structural alignment should include the set of pairs of atoms that were aligned. A complete description of an RMSD should include the set of pairs of atoms whose distances were used to calculate the RMSD. It is non-trivial (maybe impossible) to figure out from a superposition which corresponding atoms pairs were used in the superposition.

The two sets are not necessarily the same (i.e. to figure out the side chain RMSD, you probably first superpose the two structures by optimizing the fit of the main chain or C-alpha atoms, and then measure the distance of corresponding side chain atoms).

For the DALI case, they give you a structural alignment of the two sequences. It would require a bit of programming to extract the corresponding pairs of residue numbers.

For the structure prediction case, the sequences of the two structures are the same. Here, the question is whether all residues were used in the superposition, or those that are too divergent in structure were left out (in some interactive superposition algorithm).

The problem of superposing a given set of atom pairs has a unique best solution (Kabsch matrix). The decision which set of atom pairs to superpose is up to you, and depends what aspect of the two structures you are interested in. So you need the set of corresponding atoms pairs before you can solve your problem computationally.

Karsten

--
Karsten Theis
Chemical & Physical Sciences Department
Westfield State College

________________________________
From: Eric Martz <em...@mi...>
Sent: Friday, February 26, 2021 6:02 PM
To: Jmol Users LIst <jmo...@li...>
Subject: [Jmol-users] Color structural alignment by deviation "RMSD"?


Caution External Email: This email originated outside of WSU. Do not click links, open attachments, or respond if it appears to be suspicious.

For years, I have been frustrated by not knowing any way to color backbones in a structural alignment of two proteins by deviation in Jmol.

Jmol's compare command aligns two proteins, but I don't know any way to color one of the backbone alpha carbons by deviation from the other. Am I missing something?

Swiss-PdbViewer offers multiple methods to do structural alignments, and after aligning, will color the protein that was moved by deviation from the static reference structure. However, it offers no way to put the deviation values in e.g. the temperature column of the PDB file so they could be used for coloring by Jmol. (It does offer such a function, but one that works only when the two proteins have identical sequences.)

PyMOL can also align, and color by deviation ("RMSD") with its command colorByRMSD, but it is unclear to me whether the deviation values can be saved to a disk PDB file with those values in the B-factor/temperature column (which could then be used by Jmol). I have submitted a support request about this to Schrödinger.

If you'd like to see examples of structural alignments colored by deviation, take a look here: https://pymolwiki.org/index.php/ColorByRMSD

Thanks for any and all help, -Eric