Dear Ivan,
I have had a quick look at the files you sent:
-in the screen shots, the C10 of your F3P residue is at almost the same position as a nitrogen. This causes severe clashes and you need to delete
one of them (probably the C10). It is normal that upon formation of a covalent link, an atom of one of the molecules has to go, depending on the chemistry of the reaction.
-you have a cif file for the F3P molecule, but no cif file describing the link between the F3P and the protein. I do not know whether the latest
versions of Coot and Refmac can generate this definition automatically, but the versions I worked with require a definition of the link. If your link is a standard trans peptide bond, you could try adding the following link card to the header of your pdb using
a text editor:
LINK C F3P A 0 N GLY A 1 TRANS
Be careful, you have to cut and paste the line as is, the pdb format is extremely picky about the number of spaces in these cards. In this case
you also have to rename your O1 to O, your C3 to C and your C2 to CA to use the standard nomenclature for peptide bonds. You also have to update your cif file to use these new atom names. I have not tested it myself, so I cannot guarantee that it works. In
this case you do not have to provide a cif file with link definition, since you use the standard peptide link definition already available in ccp4.
-otherwise you have to generate your own cif with link definitions, which can be done be Jligand. The link below has a tutorial. You may want to
go through it to learn how to create a link:
http://www.ysbl.york.ac.uk/mxstat/JLigand/tutorial_link.html
in summary:
-you need to delete a redundant atom , probably C10
-you need cif files for both the F3P
AND the link with glycine. You may even have to merge the two cif files.
-you need a link card in the header of your pdb. Either the makelink option in Coot (see my previous mail), or cutting and pasting it with a text
editor.
Good luck!
Herman
Von: xaravich ivan [mailto:[log in to unmask]]
Gesendet: Freitag, 17. Juli 2015 20:17
An: Schreuder, Herman R&D/DE; Greg Warren
Cc: ccp4bb
Betreff: Re: [ccp4bb] AW: [ccp4bb] Coot Smiles would not build
Hi Greg, Herman all others in CCP4 who can help,
Thank you for your suggestions which I followed, but I am still not able to make the link between the F3P molecule and the first residue of protein chain GLycine.
Here is how I have modified the PDB after merging the molecules.
CRYST1 16.950 44.250 66.410 90.00 90.00 90.00 P 21 21 21
SCALE1 0.058997 0.000000 0.000000 0.00000
SCALE2 0.000000 0.022599 0.000000 0.00000
SCALE3 0.000000 0.000000 0.015058 0.00000
HETATM 585 C1 F3P A 0 -3.340 -3.808 27.614 1.00 20.00 C
HETATM 586 C2 F3P A 0 -3.613 -2.393 27.654 1.00 20.00 C
HETATM 587 F F3P A 0 -2.442 -1.672 27.870 1.00 20.00 F
HETATM 588 C3 F3P A 0 -4.477 -1.785 26.669 1.00 20.00 C
HETATM 589 C10 F3P A 0 -3.926 -0.760 25.891 1.00 20.00 C
HETATM 590 O1 F3P A 0 -5.640 -2.341 26.364 1.00 20.00 O
ATOM 1 N GLY A 1 -3.846 -0.932 25.879 1.00 5.57 N
ATOM 2 CA GLY A 1 -4.553 -0.083 24.938 1.00 5.14 C
ATOM 3 C GLY A 1 -4.433 -0.577 23.509 1.00 4.49 C
ATOM 4 O GLY A 1 -3.832 -1.619 23.250 1.00 5.72 O
I have attached the .cif file as well as the PDB of the F3P molecule as well as the images that I am getting when I am trying to link the two molecules as suggested my Herman and Greg.
We do not have something similar in PDB unfortunately.
As you have noticed the molecule diffracted to 0.95 Angs and I got several datasets at 1.02-1.05 Angs and the density is clean.All the amino acid residues are built like a charm.
Yet it is taking me ages to complete this structure!!!
I thought there should be a separates protocol describing the steps in CCP4/COOT on how to do this type of covalent linking on any molecule (as many people are trying to build new molecules (which are not always
found in the PDB with such restraint information) that so many protein molecules are modified for drugs and other purposes.
Once I get a solution, I will write up everything and then the CCP4/COOT developers could edit and use it as an addendum in the "How to".
Thanks again in advance for your suggestions and the time you are taking to read through the problem. I sincerely appreciate it.
Ivan
On Thu, Jul 16, 2015 at 11:36 PM, <[log in to unmask]> wrote:
Hi Ivan,
Giving your “ligand” the same residue number and chain ID as the first residue of your protein definitively will cause problems like being unable to refine it. There
is nothing against giving the “ligand” residue number 0 (zero) what I would do.
Also, if I understand correctly, your “ligand” is covalently linked to the N-terminus. In this case you have to specify the link in your pdb file. In coot you can do
this by clicking 2 atoms (Extensions-Modelling-Make Link). This has to be done after the “ligand” and protein have been merged. You also have to provide a cif file describing this link. This is non-trivial, but should all be described in the Refmac documentation.
I would also check whether this modification already exist in the pdb, so you could use existing definitions.
Good luck!
Herman
Von: CCP4 bulletin board [mailto:[log in to unmask]]
Im Auftrag von xaravich ivan
Gesendet: Freitag, 17. Juli 2015 04:59
An: [log in to unmask]
Betreff: Re: [ccp4bb] Coot Smiles would not build
Hi everyone,
I am definitely on my way to becoming crazy!!! I took the pdb created by Greg and I prepared my own .cif file in CCP4 as well as used the .cif file prepared by Greg but once i merge the molecule after placing it in the density. it somehow doesnot refine. I
merge the restraints into one cif file and still it does not let me refine even in COOT when I am using Calculate>Model/fit/refine>Real space refine zone, right after I place the molecule. I am deleting the hydrogens though before I place the molecule into
the density (before refinement image). After running Refmac5 I lose most of the density as you see in the image and the molecule is misplaced.
Another problem is that this molecule is basically placed at the N-terminal of residue 1 of chainA so it also has the number 1. I do not know if this is creating the problem. I chnaged the chain name into Chain A, as this is a residue in chain A and renumbered
it to the last residue of the peptide instead of renumbering all the aminoacids and still no luck.
What should I do, I am running out of ideas.
Am I doing the steps right?
1) I generate or get a pdb of the ligand molecule (basically it is not ligand in that sense!!)
2) Generate restraints by ProDrg
3) Open the pdb of the molecule, and import cif file and then open the PDB protein molecule and the corresponding map
4) Fiting the ligand molecule into the protein Map
5) Merge molecule in COOT
I can't explain how much I would appreciate your help and suggestions.
Thanks,
Ivan
On Wed, Jul 15, 2015 at 9:41 AM, Paul Emsley <[log in to unmask]> wrote:
On 11/07/2015 01:02, xaravich ivan wrote:
Hi everyone
I am using SMILES to build this Fluopropyl group and SMILES would always give me this error, after I put the correct string.
The image and string I built and the error image is attached.
Am I missing something?
For the record, the problem occurred because you were running coot in a directory in which it was unable to write files and directories. (This is not recommended.)
Paul.