Yes in the first couple of rounds of refinement it refines very well for a
3.2 angstrom structure (Now at 30%/24% Rfree/R). Everything Packs
contiguously except for a "donut" hole in between six dimers that are
related by symmetry. Trying to put a molecule there disrupts the symmetry
and leads to clashes. I have a synchrotron trip next week, hopefully this
should help clear things up a bit.


On Wed, Mar 19, 2014 at 12:17 AM, Eleanor Dodson
<[email protected]>wrote:

> I think you have solved it! That is an excellent LLG and if you can't see
> anything else in the map, then there s prob. not another molecule.
> Does it refine? If you look at the maps following refinement any missing
> features should become more obvious.
> Solvent content of 65% is not uncommon.
> Eleanor
>
>
> On 19 Mar 2014, at 03:46, Yarrow Madrona wrote:
>
> Hello CCP4 Users,
>
> I recently collected data in-house on an Raxis IV and am trying to solve a
> 3.2 angstrom structure.
>
> I have obtained only "partial solutions" using Phaser and would like some
> help. I believe I only have two molecules in the ASU instead of three as
> suggested by the mathew's calculation. I believe I have two molecules in
> the ASU with a space group of P312 despite a high solvent content. I have
> outlined by line of reasoning below.
>
> 1. Indexes as primitive hexagonal
>
> 2. Self rotation function (MolRep) gives six peaks for chi = 180. (I'm
> assuming chi is equivalent to kappa for Molrep) supporting the 2 fold axis
> in the P312 space group. See this link,
> https://www.dropbox.com/s/sj7c28pvuz7fei2/031414_21_rf.pdf
>
> 3. Scaling in P3 gives 6 molecules/ASU predicted by Mathew's calculation.
> Phaser gives solutions for only 4 molecules.
>
> 4. Scaling in P312 gives 3 molecules/ASU predicted by Mathew's
> calculation. Phaser gives solutions for only 2 molecules.
>
> Mathews calculation for data scaled in P312:
>
> *For estimated molecular weight   44000.*
>
> *Nmol/asym  Matthews Coeff  %solvent       P(3.20)     P(tot)*
>
> *____________________________________________________________*
>
> *  1         6.84            82.03         0.00         0.00*
>
> *  2         3.42            64.07         0.18         0.13*
>
> *  3         2.28            46.10         0.81         0.86*
>
> *  4         1.71            28.13         0.01         0.01*
>
> *  5         1.37            10.17         0.00         0.00*
>
> *____________________________________________________________*
>
> *Phaser Stats:*
>
> Partial Solution for data scaled in P312:
>
> RFZ=11.7 TFZ=27.4 PAK=0 LLG=528 TFZ==18.8 RF++ TFZ=52.1 PAK=0 LLG=2374
> TFZ==30.3
>
> 6. No peaks in patterson map (No translational symmetry).
>
> 5. Very strong 2fo-fc density for two ligands in each monomer (Heme and 4
> bromo-phenyl Immidazole) despite not including them in the search model.
>
> 6. There is only one "black hole" where it would be possible place another
> subunit but there is not much interpretable density and the symmetry of the
> space group would be broken if this was done. Six Dimers are arranged
> around this hole. I can post a picture if anyone wants to see it.
>
> 6. Early refinement of the "partial solution" gives an Rwork/Rfee ~
> 24%/31% for a 3.2 angstrom data set. Probably over parameterized judging by
> the gap in R/Rfree but still better than I would guess if I had only 2/3 of
> the ASU composition.
>
> *My belief is that there really is only two molecules in the ASU and that
> there just happens to be a very large solvent channel giving a 65% solvent
> content.*
>
> *I would like help in determining whether this is likely or if I have
> missed something. Thank you for your help in advance!*
>
> *-Yarrow*
>
>
> Post Doctoral Scholar
>
> UCSF
>
> Genentech Hall, Rm N551
>
> 600 16th St., San Francisco, CA 94158-2517
>
>
>

Reply via email to