Re: [phenixbb] sequence independent model building possible?

Dear All, Thanks for all the valuable inputs. I was able to build side chains for the core of the protein. Using this sequence information, I was able to identify sequence family from Pfam. It seems like it has phosphate-binding domain (I could also see a blob which looks like a phosphate). I am using Multiple sequence alignment from the closest Pfam hits to built sidechains for the rest of the protein. The latest scores are, R/Rfree: 28/37; LLG=3681; TFZ=39.2 I hope to solve it very soon. Thanks, Kaushik P.S: 1) I had searched Dali too. These hits (with less than 2 A RMSD), when used as phasing model in Phaser failed to yield a solution. The top hits probably belong to the same fold but seems to have internal domain movements. However, the core was conserved (which I recognise now as phosphate-binding domain). 2) Had also searched PDB for unit cell with similar dimensions without much help. Will checkout Nearest-Cell v2.0. Thanks for the this. 3) ~20% of the structure is loop for which backbone isn't built. I hope to model this soon. Thanks again for all the help. On Sat, Feb 6, 2016 at 3:01 AM, Jon Agirre wrote:

Dear Kaushik,

if you're suspecting you've crystallised something else, perhaps you could try running your crystal parameters through the nearest-cell server ( http://app.strubi.ox.ac.uk/nearest-cell/nearest-cell.cgi), which will scan the PDB for crystals that match the input one.

Good luck,

Jon

On 5 February 2016 at 20:06, Christian Roth wrote:

...

Hi, besides the already excellent suggestions, you might want to try if density modification (NCs, solvent flattening, histogram matching) improves your map a bit further. If you can assign enough residues you improve your maps than even further step by step. On top your stretches are than definitely long enough for a blast search.

Christian On 5 Feb 2016 06:02, "Kaushik Hatti" wrote:

...

Hello,

Is abinitio model building possible for a map with poly alanine model at 1.9A resolution?

We thought we had crystallised our protein of interest X, collected data at 1.9 A and all attempts to solve protein X (which has many homologs) through MR failed. All attempts to re-crystallise the same protein also failed.

Now, we think the initial protein which got crystallised could be a contaminant (we don't have any crystals left from this batch to check for the sequence of the crystallised protein). Through various methods (and a bit of luck) we have arrived at a decent map with LLG : 3600 and TFZ: 22 and R/Rfree : 37/41 (for a poly alanine model).

I believe these scores indicate right fold. As I still don't know the sequence information, is it possible to build sidechains directly from the map (I could only identify a couple of residues and the model largely remains PolyAla)? Autobuild with Rebuild-in-place didn't help in identifying any more residues.

I have also searched PDB database for similar structures. But, none of those are either from our expression system (E. coli) or organism of our protein of interest. Neither did I find any similar sequences from E. coli or our organism of interest.

Any leads/suggestions would be helpful. Thanks, Kaushik, MRN Murthy lab, MBU, IISc, India

-- Stupidity is everyone’s birthright. However, only the learned exercise it! --Kaushik (28Oct2014)

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]

-- Dr Jon Agirre York Structural Biology Laboratory / Department of Chemistry University of York, Heslington, YO10 5DD, York, England http://www.york.ac.uk/chemistry/research/ysbl/people/staff/jagirre/ +44 (0) 1904 32 8253

-- Stupidity is everyone’s birthright. However, only the learned exercise it! --Kaushik (28Oct2014)