This is only a suggestion but sometimes it helps.

First idea - try as hard as you can to extend the native data - at 3.6A
resolution all model building is extremely hard!

To work with what you have I would do this - others may have different
ideas.
Combine both data sets into one mtz file, and include your SHARP phasing
PHIB FOM HLA etc (not solvent flattened yet)

run DM with averaging for the homodimer, using the native F to 3.6A and
the SHARP phases - this will extend your phases to 3.6A

You will have to be a bit clever to only average over the homo-dimer and
not the second component. You will need a mask covering one molecule (A
say) and the transformation matrix which converts A to B. There is
documentation to explain this

With luck that should improve the map.

 Then I would refine your existing model, with the phase restraints
HLADM etc using REFMAC. This will give you combined output phases and
the resultant maps SHOULD improve..

But as I said - at 3.6A you will not get a very clear ma and it will be
tricky to build extra residues. The Se markers should help..

 Eleanor



Raja Dey wrote:
>
> Hi,
> It would be of great help if you can give some
> suggestion on the following problem.
>
>
> I have two data sets of a
> complex molecule. One is native (3.6A) and the other is Se-der
> (4.5A). In one complex molecule I have one homo-dimer in complex with
> DNA( 2 times 90 aa + 2 times 15 na = ~MW 28000, call it as component
> 1) plus a second long helical protein(130 aa, call it as component
> 2). I have two such complex molecule in the AU (space gr. P65).
> I got a very good solution by
> PHASER(using native data) for the component 1, but component 2 has
> some ambiguity.
> I used the solution of
> component 1 and Se-der data to create Ano-diff Fourier map from model
> phase. I got clear peaks at 3.5 sigma exactly at the Met-S positions.
> I pulled out 11 Se sites, 9 of which are precisely exists at the
> right positions looked upon the model component 1). Other 2 are
> expecting on the two copy of component 2.
> What I did next?
> 1. I used CNS to create both
> SAD phase from Se and Model phase from component 1 and then combined
> them. Used the combine HL co-eff to run density modification and
> create 2Fo-Fc map. Very poor quality, but you can see some features.
> 2. Used only SAD phase from Se
> sites and run density modification followed by 2Fo-Fc map
> generation.
> Useless map.
> 3. Run SHARP/AutoSHARP with Se
> sites upto DM. Then run solvent flattening with model component 1.
> No good map.
> Do you know any other way to
> go from here? Running rigid.inp, anneal.inp, minimize.inp and
> bgroup.inp sequentially with the component 1 I reached R/R(free)
> 40%/45%.
> Any suggestion is well
> appreciated.
> Regards...
> Raja
>
>
> From Chandigarh to Chennai - find friends all over India. Go to http://in.promos.yahoo.com/groups/citygroups/
>