🚀 go-pugleaf

In article <3t9qb4$g6c@news.ox.ac.uk>, dnicker@sable.ox.ac.uk (Darren Nickerson) writes:
|> Dear "O" gurus,
|> 
|> In the absence of the resources offered by the "O" WWW server (it is currently
|> unreachable due to hardware trouble I understand), I am turning to you for a

actually, it's the O-info mail server that's down; the WWW page works (I think)

|> quick bit of advice on a fairly straightforward problem.
|> 
|> I have a model refined to R=0.20 at 1.8A, in which all but 9 of the approx 300
|> residues have been built.  I see "patchy" density for the remaining residues,
|> and I would like to build them in in the hopes that the improved phases will
|> reveal the entire chain.  I would like to build 2 unconnected sections of 
|> 3 residues, but I'm not sure of the best way to accomplish this in "O".  The
|> model was built with Frodo, and only recently have I discovered the "joy of O".
|> I have never built in O at all, so any pointers to useful info would be much
|> appreciated.  As well, quick suggestions as to potential troubles in building
|> such a "disconnected" model would be helpful.
|> 
|> Thanks in advance,
|> 
|> -Darren Nickerson

well, since it has to be quick'n'dirty, I would do the following
(alwyn would tell you to learn the baton commands):

- mutate_insert all 9 missing residues (e.g. as Alanines)
- use lego_loop to get coordinates for their CA atoms (select
  the loop that passes to most of your patches of density)
- move_atom the CAs into the density
- use lego_auto_mc to put the main-chain coords in
- use lego_auto_sc for the side-chain atoms
- use move_zone (or rsr_rigid) on individual residues to
  fit them better into the density
- regularise the geometry with refi_zone
- then I would throw this into a 4000K slow cool (with
  all nine residues as ala so they can move around a lot
  if they want to)
- after that, look at the density & decide what to do
  (put in the real side chains or delete again) for each
  of the nine trouble makers
- if you have NCS, i would constrain it & average the maps
  (2Fo-Fc and Fo-Fc) to see if the density gets better

the O tutorial teaches you how to model insertions (chapter 8);
the manual explains individual commands

on the other hand, from R & resolution i assume that your model
is already fairly complete (waters etc); the current map is therefore
a kind of omit map for the missing bits; if there is no good
density yet, perhaps the residues are disordered
putting something in may lead to *apparently* better phases/maps
due to model bias, without really improving the model, so
beware

--cd