Hi,
>> 9. The selenomethionine dataset was solved using MIRAS in SHARP/autoSHARP.
>> The experimentally phased electron density yields contiguous tracts of
>> density in the right place, unbiased density indicates a good solution.
>> Model building was conducted in P212121, initially into the experimental
>> maps and later with refinement in refmac using HL-coefficient-based
>> restraints. In some regions, sequence can easily be deduced from clean
>> electron density (for the resolution). In other regions, side chains are
>> missing, and in others, density is completely inconsistent with the
>> connectivity of the chains and highly conserved structural elements. As
>> occurs sometimes with twinned data, many loops cannot be modelled at all,
>> and the Rfree does not drop below 0.41 with an Rfactor of 0.34. The
>> result
>> is a model that is about 60-70% complete. Refinement was performed with
>> and
>> without B-factor refinement.
It is not uncommon to have some severe density truncations in density
modification from difficult/poor data/phases. Since you have a partial
model, I would recommend a procedure we've used over the last years quite
a few times: feed-back of the model into density modification (see
e.g. http://scripts.iucr.org/cgi-bin/paper?S090904950302394X). It is
easily done in the SHARP/Sushi interface:
- copy your latest/best model into sharpfiles/datafiles/model_xyz.pdb
- start a new density modification run with this initial model as an
additional phase information: our SOLOMON script will then start
not just from the SHARP phases, but from the SHARP+model
phases. There are some decisions you might need to do
* how many cycles to use this model: if you're worried about bias,
just use it for a single cycle. Usually, I would expect a model
that is based on some initial experimental map to be fairly
unbiased - unless you used some homologuous structure as an
initial model.
I usually use the model for say 10 cycles and doing 40 cycles of
density modification.
* you might want to redo the optimisation of the solvent content:
since you now start with a better initial map (hopefully), this
might give you a very different solvent content - initially poor
regions that were flattened might pop up.
- have a look at the resulting map (FBshasol/PHIBshasol column if you
don't use the tools in Sushi) and try to correct your model and -
maybe - complete it.
- go back to the start with a now improved model.
I tend to do this 3-4 times at least (sometimes not a lot seems to be
happening at the first 1-2 cycles, but often it suddenly picks up).
If you want to help your refinement (e.g. in REFMAC) with HLs from
SHARP: you want to use FP/SIGFP and HLA-D from the eden.mtz file (or
FPsha/SIGFPsha and HLA-D from a eden_flat_*.mtz file). These are in
synch with each other. Also: remember to build MSE residues and adjust
your f' value.
If you want to improve the HA refinement in SHARP you could use phase
information from a model or from a density modification run throughout
the HA refinement and residual map calculation (but _not_ the
phasing): if you're still missing a few Se or there are some alternate
conformations for the Se, this might be a good way of picking them up.
>> 11. phenix.autobuild is able to build a polyalanine model that covers
>> about
>> 25% of the molecule.
I've had some very nice results recently with Kevin Cowtans BUCCANEER
(latest binary from his WWW page): it seems especially good in
building a lot of connected regions.
If you use such an automatic built model in the iterative procedure
mentioned above, you don't really have to worry about bias I guess.
Cheers
Clemens
--
***************************************************************
* Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com
*
* Global Phasing Ltd.
* Sheraton House, Castle Park
* Cambridge CB3 0AX, UK
*--------------------------------------------------------------
* BUSTER Development Group (http://www.globalphasing.com)
***************************************************************
|
