Hi Wei,
First a note on style: Your reply style leaves something to be desired.
I had some difficulty in understanding the answers to my questions.
Generally speaking, if the response style is interleaved, you should
follow that style, rather than top posting. Also, you should quote the
response, rather than just re-quoting your original post.
On 08/10/13 03:23, Wei Shi wrote:
Thank you so much for the reply!
You are welcome.
Would you please let me know how to use Coot's ligand builder or where
I could find more information for Coot's ligand builder?
You use the ligand builder by using: Calculate -> Ligand Bulider.
I don't know which is the best way to build the ligand to fit the
density in my case. I got the molecular replacement solution using the
apo-protein strucutre, but found out that there might be a
conformational change in some part of the protein, so the density for
that part is bad and also the density for the ligand is not quite
continuous....So, I deleted the part that have bad density and do a
refinement and then, place the ligand structural model in pymol. I
first got the 3D structure of the ligand from 3DMET,
You did this because the ligand type is not already in the PDB library
(or refmac dictionary). Is that the case?
which a database for small molecules, and load that PDB onto PRODRG2
website
generally that's a bad idea. You should use the mdl mol file if you can
get it.
If you have the SMILES string you can use CCP4's libcheck via Coot's
File->SMILES.
and get a new PDB as described in more detail below. And when I opened
the protein with the ligand built back in, and found out that in some
part, the ligand needs to be moved in order to fit the density well,
so I was trying to move the ligand in coot to fit the density.
I see. You can use Torsion General without needing a dictionary, I
believe.
It seems that the reason coot is not working is because that I used
the PDB which is not compatible with the LIB file?
It seems that way to me.
Now when I loaded the PDB file of the ligand which I entered the
PRODRG2 website instead of the PDB generated on that site, plus mtz
and LIB file onto coot, I could move the ligand.
Clever old prodrg. It has done the atom name matching. Generally
though, the output pdb file should be used with the output dictionary.
"move the ligand" - in what way?
Before I used the pdb file generated in the PRODRG2 website (no H's
PDB file, since on the website it says that 'to get the ligands pdb
file (to build into your structure using a modelling program such as
Coot) use the"pdb no H's).
I doubt that my web site says such a thing, but no matter...
Now, when I was trying to move the ligand, Coot didn't give the
previous error message, but it is just very slow and didn't quite work
the way I expected,
What did you do? What did coot do? What did you expect?
and I will play with it....
And, for the CIF, on the PRODRG2 website, it says CIF topology
(general
<http://davapc1.bioch.dundee.ac.uk/cgi-bin/prodrg/run.html#DRGMAC.LIB>
[D]
<http://davapc1.bioch.dundee.ac.uk/cgi-bin/prodrg/run/5252D9EEC5880000191A/DRGMAC.LIB?m=1>
or PHENIX
<http://davapc1.bioch.dundee.ac.uk/cgi-bin/prodrg/run.html#DRGPHE.LIB>
[D]
<http://davapc1.bioch.dundee.ac.uk/cgi-bin/prodrg/run/5252D9EEC5880000191A/DRGPHE.LIB?m=1>)
and I downloaded the general one to load onto Coot.
OK, I am unfamiliar with that output. I would have thought that the
output of a dictionary generator that is compatible with Phenix is
compatible with Coot too.
You mentioned that 'I'd use the CIF dictionary that had hydrogens-
the one that is compatible with Refmac'. I am wondering how you
usually generate the CIF file?
For the current stable release, at least, I use a these methods.
1) CCP4's libcheck via Coot's File -> SMILES
2) CCP4's cprodrg via Coot's ligand builder
I'd also recommend looking at Phenix.elbow.
And, I am doing all the things I mentioned above and in the previous
email because I read in the following link that "If you have a ligand
you will need to read the "Refmac" dictionary for that ligand into coot".
http://www.ucl.ac.uk/~rmhasek/coot.html
<http://www.ucl.ac.uk/%7Ermhasek/coot.html>
I see. I suspect that that web page was written before Coot's ligand
bulider. I wouldn't do it that way today.
Let me know what you think might a better strategy for me to build
back in the ligand to fit the density....
If this is an as-yet unknown ligand:
Calculate -> Ligand Builder
then either import the mol file or sketch the ligand
Apply [3D molecule appears in graphics]
Then, in the graphics window, rotate, translate the molecule as needed
and edit chi angles as needed.
or
Calculate -> Other Modelling Tools -> Find Ligand [possibly using the
"flexible" option]
Then refine it.
It should take a minute or two.
Thank you so much!
welcome.
Regards,
Paul.
