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

[Robert H. <ha...@st...>]

Eric, don't worry -- you aren't missing anything. I looked into doing
alignments several years ago and decided to stick with the original plan of
leaving things like this to more specialized programs that could do a
better job of it. Remember, PyMOL is also not an alignment program. It's an
image creating tool. So maybe it does some approximation of an alignment --
and maybe it is actually very good. I don't know. More probably it's
"pretty enough" for what people want to see.

In the end, a structure alignment is just a list of coordinates, after all,
and perhaps a distance matrix. Aren't there web services now that can do
this via an API? For example
https://academic.oup.com/nar/article/47/W1/W5/5486273  ?

Anyway, that paper does a nice job of illustrating the technical issues
involved in sequence+structure alignment, which is what you are after, I
think. Then the question simply comes down to how to access an API and how
to translate the output data into something Jmol can display. This is the
Jmol strategy.

True, Jmol does implement the DSSP algorithm exactly, but that is a very
specific algorithm that doesn't have a lot of variants and was graciously
donated to Jmol with full code reference. True sequence+structure alignment
is a much more complicated process. Multiple algorithms, multiple
iterations; lots of adjustable parameters.

So, take a look and see what you can find. Maybe there is a great service
out there that would like to work with us. And if you are satisfied with
whatever algorithm PyMOL uses, and Jmol doesn't read the PSE file, send it
to me so I can discover why not. You shouldn't need any sort of export.
Atom positions and coloring are pretty basic aspects in PyMOL, so Jmol
should be able to read that just fine. Then you can make your own PDB
files! :)

Bob