I'm going to try to start writing down some of our discussions
during our Thursday Chimera meetings. My first stab at the
current issues may be found at
https://www.cgl.ucsf.edu:8000/intranet/chimeradevelopment/2002-04-18.html
There is also a link from the intranet page to the list of available pages
(currently one :-). If I've left off an issue, please let me know and
I will revise.
Conrad
I've implemented the interface changes that standardize capitalization and
so forth that we agreed on in one of our meetings a while back (http://www.
cgl.ucsf.edu/home/meng/frameprotocol.html). The only substantive change
was the reorganization of the Help menu. The remaining changes were
principally case changes and adding/subtracting colons and whatnot.
--Eric
Hi Ed,
You're welcome!
>* Why can you "surface ligand", but not "select ligand"?
Most commands accept an atom specification, but ligand/main/etc. are
surface categories, not atom specs. Of course, that is just a
restatement of your observation that "ligand" does not work as an
atom spec. The surface categories were developed since there needed
to be reasonable groupings for surface calculation, a whole different
issue than whose surface is actually displayed. For example, if
you surface a molecule such as biotin that is considered "ligand," the
surface actually encloses the biotin, but if you surface some random
residue in the protein, Chimera "knows" it is part of "main" and only
shows surface where the residue is at the surface of the main protein.
Thus surface *creation* would use the categories, but then *display*
of the surface could use either categories as a whole, or atom specs,
which could be subparts of categories.
When using all those commands that use atom specs, I agree it might
be convenient to also allow ligand/main/etc. to be atom specifications.
The programmers on the team would need to think about this a bit more,
in case there are ramifications. For now, you could use the atom spec
@/surfaceCategory=ligand
>* Automatic chain break detection would be really helpful. Currently,
>Chimera connects all the residues in a chain, even when it's clear from the
>bond length that there must be a chain break.
Check out the command "longbond" - it should do the trick. There should not
be a lot of long bonds in PDB standard-compliant files. They will occur
in segments of missing density, but their presence does inform you that
there is no chain break in reality, just missing density. Also, it is
pretty easy to just use longbond if you want them to go away. The
command can take parameters, but in practice I have never needed to
specify them; I just use "longbond" with no arguments.
Hope this helps!
Elaine
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Elaine C. Meng, R.Ph., Ph.D.
Babbitt Lab and Computer Graphics Lab
meng(a)cgl.ucsf.edu
http://www.cgl.ucsf.edu/home/meng/index.html
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
The email below makes some good points. I can answer that "ligand" etc.
are surface categories and not atom specs, but it does beg the question
of why we couldn't make them usable as atom specs. The possible answer
to the second item is that so far, we have focused more on handling
correct PDB files, not various pathological cases. However, I am not
sure what is supposed to happen in a correct PDB file when part of a chain
is lacking density (e.g. a chain with residues 1-20 lacking density for
residues in the middle...is there supposed to be a TER card in that case?).
Comments on these issues welcome. Thanks,
Elaine
--------------------------------------------------------------------
>From boas(a)stanford.edu Thu May 2 14:10:18 2002
Date: Thu, 02 May 2002 14:10:13 -0700
To: Elaine Meng <meng(a)cgl.ucsf.EDU>
From: "F. Edward Boas" <boas(a)stanford.edu>
Subject: Re: Chimera issues/questions
Hi Elaine,
I wasn't expecting such a quick reply ... thanks! I have a couple more
suggestions:
* Why can you "surface ligand", but not "select ligand"?
* Automatic chain break detection would be really helpful. Currently,
Chimera connects all the residues in a chain, even when it's clear from the
bond length that there must be a chain break.
Thank you!
Cheers,
Ed
________________________________________________________________________
F. Edward Boas Phone: 650-787-1688
P.O. Box 17890 Email: boas(a)post.harvard.edu
Stanford, CA 94309 Web: http://www.stanford.edu/~boas
I've changed chimera-dev from a plain mailing list to a "mailman"
mailing list, just to see how well mailman works. You can visit
http://www.cgl.ucsf.edu/mailman/listinfo
to see a web interface. Let me know what you think.
Conrad
PS This is only the second list I've created, so some things
may not be working properly.
