*** For details on how to be removed from this list visit the *** *** CCP4 home page http://www.ccp4.ac.uk ***
Hi Sam, Is this data from any HOT beamline? It is observed that Hg derivatives (HG bound to CYS) are very sensitive to radiation damage. If you have already refined the structure and ready to submit do not read the lines below. Also, this mail has nothing to do with your refinement problems. If not... Now assuming your map is of reasonably good quality, you can calculate anomalous difference Fourier map with the first-data (say Peak) in your MAD experiment together with the current experimental phases. Do the same using last data (say remote). If you see very high anomalous difference density peaks with data-first and no density with data-last (this will be weaker even if there is no decay)...then YOU MAY BE LUCKY!! That means you have enough decay to exploit. Assuming the version of SOLVE you are using is recent you should see be able to locate the refined f' values. If it is very high (I do not know whether the program has any limits) then again this indicates you have specific decay. EASY TO FOOL THE PROGRAM NOT THE BOSS (VERY SMART) Now rerun the SOLVE MAD-job with first data (use as small wedge as possible) and the last data with high f' values (say somewhere between -40 to -80) OR run SIRAS with first data as derivative and last data as native. If you have decay you should be able to see much better map . You are essentially adding decay contribution to dispersive deference (See Pseudo mad section of Acta D. 2005, vol 61, pp 1289-1298). Some other tests: Do you remember whether the errors you used in HKL2000 had much higher values for the first data compared to the last data (particularly at low resolution?) Also see discussion about the chi2 plot in the same paper. Hope you will get some bonus electrons. HAPPY HOLIDAYS Ramu > *** For details on how to be removed from this list visit the *** > *** CCP4 home page http://www.ccp4.ac.uk *** > > > > Dear Everybody, > > I have a 2.7A MAD data o f Hg derivative. By SOLVE I got the phase and by > RESOLVE I got initial model (50%) which I modelled upto 75% (poly-ALA/Gly) > but some of the fragments are not contineous. C2 space group and 2 > molecules > in the asymmetric init, 400 residues per monomer. > In order to extend and complete the model (poly ALA) I tired to refine the > model by REFMAC5 (CCP4i) and CNS. > > In REFMAC5 I used "no restrain" and varied weight from 0.3 (default) to > 0.05. In all cases I got the output PDB where mainchain residues are not > regularized as seen both in O and COOT. In O I could at least see CA > trace, > but in COOT CA trace is brocken (maximum 4-5 resides contineous). > > Thereafter I used "restrain" and in library gave a monomer Alanine, then I > got a output PDB which is better than before. I can get CA trace in both O > and COOT. But some of the mainchain atoms and CB are not regularized. I > included Hg atom in the input PDB. But in the output PDB I donot get Hg. > > I would appreciate if somebody advice me how to overcome this. > > Then I tried CNS "model_map.inp" to get the map for improving the model. > (a) generate.inp gave me output (generate.pdb) but I still miss the Hg > atoms > in the output. > (b) then I added Hg atoms manually in generate.pdb and ran "minimize.inp". > In the output (minimize.pdb) after minimize I am also not getting Hg > atoms. > (c) However, taking the minimize.pdb (with protein residues poly ALA/Gly > only) I am trying to run "model_map.inp". I would like to know how I can > get > HL co-efficients from .SCA file (created by HKL2000) as it requires as > follows. > > {* Hendrickson-Lattman coefficients A,B,C,D *} > {* required for the "mlhl" target and phase combined > or observed maps *} > {+ table: rows=1 "HL coefficients" cols=4 "A" "B" "C" "D" +} > {===>} obs_pa="pa"; > {===>} obs_pb="pb"; > {===>} obs_pc="pc"; > {===>} obs_pd="pd"; > > Then I would like to know how can I calculate > xmin, xmax, ymin, ymax, zmin, zmax > as it requires as follows. > > {* limits in orthogonal angstroms for box mode or fractional coordinates > for fract mode *} > {+ table: rows=3 "x" "y" "z" cols=2 "minimum" "maximum" +} > {===>} xmin=""; > {===>} xmax=""; > {===>} ymin=""; > {===>} ymax=""; > {===>} zmin=""; > {===>} zmax=""; > > > If I run "model_map.inp" without above parameters, then I get the map but > that does not cover the model and model chains (with helices and strands) > are also not going through the electron density. > > d) I assessed in the model that there are 3 Cys residues per monomer, so > total 6 Cys residues per Asymmetric unit. > When I checked generate.pdb in COOT after (a) above, I see there are > straight lines (25 to 50A) between CYS residues of two monomers in the > Asymmetric unit. Is COOT assuming something here between Cys residues of > two > monomers in the Asymmetric unit or I needed to do something so that I can > avoid these kinds of straight line btween Cys residues. > > I would appreciate comments and suggestions to short out above problems. > > Thanks > Sam > > _________________________________________________________________ > Your Hotmail address already works to sign into Windows Live Messenger! > Get > it now > http://clk.atdmt.com/MSN/go/msnnkwme0020000001msn/direct/01/?href=http://get.live.com/messenger/overview >
