Dear Pooja,
A few remarks:
- Matthews does not show 4 chains in the asymmetric unit, it suggests 4 chains. However, in reality it can be more or less chains, although rare, 75% solvent (2 chains) is not unheard of.
- An initial Rfree of 38% is ok, 32% after refinement is a bit high
- Disordered loops do exist and you may have to live with them (not be able to build them).
- To correct for anisotropy, I suggest the staraniso server from global phasing.
- Unless you ran your molecular replacement in P1, Zanuda will confirm the space group you choose for MR. So run MR in P1 (if your symmetry is not too high) and run Zanuda again. (Have someone) look critical at the assigned space group and assess whether another choice might also be possible.
Good luck!
Herman
Von: CCP4 bulletin board [mailto:[log in to unmask]] Im Auftrag von Pooja Kesari
Gesendet: Donnerstag, 26. Januar 2017 15:12
An: [log in to unmask]
Betreff: Re: [ccp4bb] Bad density for chains
We have a 2.6 A structure showing four chains in an asymmetric unit. Our protein is 360 residues around 40 kDa . Mattews shows four chain in an assymetric unit (solvent 49% mattews coeff 2.44). The template has about 60% homologous with our protein. The molecular replacement against this template gave an initial free R of 38. We did chain tracing and found that we have good density (2Fo-Fc) for chain A and B but poor density for C and D.
1. The density for a particular stretch of 10 amino acids (disordered loop region) is absent in all the chains. We could not found density for this flexible loop region in any of the already known structures. Any suggestion on how can we build this region?
2. We did not find density for most of the loop regions in chain C and D which were well traced in chain A and B. How can we improving the density for these two chains based on chain A and B (Density modification)?
3. We analysed the data using phenix xtriage and found that our data shows severe anisotropy. Any suggestion of anisotropy correction?
Pointless and Ctruncate analyses didn't show twinning or NCS. I have checked the space group using Zanuda. We are stuck at a free value of 32.
On Thu, Jan 26, 2017 at 4:47 PM, Eleanor Dodson <[log in to unmask]> wrote:
This is a bit too vague to help much.
How did you solve the structure?
Eleanor
On 26 January 2017 at 03:50, Pooja Kesari <[log in to unmask]> wrote:
Dear All,
Thank you all for reply.
We have checked the data for twinning.
Our protein is 360 residues around 40 kDa protein.
We have tried TLS refinement.
chain A and B don't superimpose well with chain C and D. (A and B chains also share slight difference )
Since we don't have proper density for some regions chain C and D, we are not sure whether these chain have similar or different conformations.
We tried anisotropy correction and the model refined a bit.
On Wed, Jan 25, 2017 at 10:32 AM, Debanu <[log in to unmask]> wrote:
Hi Pooja,
Are you positive you have the correct space group and there are no other issues like twinning, etc?
If sure, did you define NCS groups in refinement? TLS refinement? Try different refinement programs?
How big is the molecule? Was it solved by MR or experimental phasing?
You can try superimposing A/B on C/D and refinement with tight NCS then adjust NCS restraints during model adjustments based on local differences or also see if phenix autobuild helps.
Best,
Debanu
--
Debanu Das
Accelero Biostructures
On Jan 24, 2017, at 8:42 PM, Pooja Kesari <[log in to unmask]> wrote:Dear All,
I have a 2.6 A resolution structure having four chains in an asymmetric unit.
The chain A and B have density for almost all residues however we don't have proper residue density in chain C and D.What can be tried to build chain C and D ?
Many Thanks
Pooja
--
Thanks & Regards,
Pooja KesariResearch Scholar
Department Of Biotechnology
Indian Institute of Technology Roorkee
INDIA
--
Thanks & Regards,
Pooja KesariResearch Scholar
Department Of Biotechnology
Indian Institute of Technology Roorkee
INDIA