Re: [phenixbb] PHENIX user mailing list <[email protected]>
Hi Maia,
I have a Ca2+ ion in a special position. the ocupancy should be 0.5, as the other half belongs to a symmetry-related molecule. However, refinement makes the occupansy 1. What should I do to prevent this? Should not it find automatically that?
phenix.refine uses the correct symmetry factor automatically (internally). Unless you have an actual partial occupancy, the occupancy factor in the pdb file should be 1, irrespective of the multiplicity of the special position. Let me know if this doesn't answer your question. Ralf
Thank you, Ralf and Nick for your help. I could see from the lof file that phenix.refine determined 2 sites at special positions, but I thought it did nothing about it, as the occupancies were brought back to 1. Thank you for your explanation that phenix.refine actually takes this into account. Maia Ralf W. Grosse-Kunstleve wrote:
Hi Maia,
I have a Ca2+ ion in a special position. the ocupancy should be 0.5, as the other half belongs to a symmetry-related molecule. However, refinement makes the occupansy 1. What should I do to prevent this? Should not it find automatically that?
phenix.refine uses the correct symmetry factor automatically (internally). Unless you have an actual partial occupancy, the occupancy factor in the pdb file should be 1, irrespective of the multiplicity of the special position.
Let me know if this doesn't answer your question.
Ralf _______________________________________________ phenixbb mailing list [email protected] http://www.phenix-online.org/mailman/listinfo/phenixbb
Ralf W. Grosse-Kunstleve wrote:
phenix.refine uses the correct symmetry factor automatically (internally). Unless you have an actual partial occupancy, the occupancy factor in the pdb file should be 1, irrespective of the multiplicity of the special position.
Doesn't that violate the mandate that the pdb file represent the contents of the AU? Unless I misunderstand Maia's structure, the Ca++ is shared between two asymmetric units, so each AU contains 0.5 of it? Ed
Ralf W. Grosse-Kunstleve wrote:
phenix.refine uses the correct symmetry factor automatically (internally). Unless you have an actual partial occupancy, the occupancy factor in the pdb file should be 1, irrespective of the multiplicity of the special position.
Doesn't that violate the mandate that the pdb file represent the contents of the AU? Unless I misunderstand Maia's structure, the Ca++ is shared between two asymmetric units, so each AU contains 0.5 of it?
That's the first time for me to look at the situation from that viewpoint... My thinking so far: Given the space group symmetry, the special-position information is in the coordinates. From that follows directly the multiplicity of the Wyckoff position, as listed in the International Tables. If the occupancy factor is convoluted with the multiplicity_of_the_special_position ------------------------------------ multiplicity_of_the_general_position quotient, the result is not an occupancy factor anymore since two distinct concepts are mixed. But ultimately, what should be in the PDB occupancy column is a matter of convention. Prompted by your message I looked at: http://www.wwpdb.org/documentation/format32/sect9.html This leaves me uncertain. Then: http://mmcif.pdb.org/dictionaries/mmcif_std.dic/Items/_atom_site.occupancy.h... Which says: The fraction of the atom type present at this site. The sum of the occupancies of all the atom types at this site may not significantly exceed 1.0 unless it is a dummy site. There is no mention of including the symmetry correction. Is anyone aware of other resources that define what to do about the symmetry correction factor for atoms on special positions? Ralf
"It depends on what the meaning of the word IS is" Here it depends on the meaning of the word "site". Every program I can recall, except for phenix.refine, has operated with the convention that every atom in the model is subject to symmetry expansion by all the equivalent positions. This convention has the result that an atom on a special position doubles (or triples, or ...) up and its occupancy has to be reduced to get the electron density right. (Obviously no one applies the same convention to electron density. Consistency is not a hallmark of humanity) I'm not saying that one way or the other is "right". I think its pretty clear that the traditional convention was created for the convenience of the programmer and many users have mistakenly left the occupancy at 1 with the result of inflated B factors. Your convention does not require that the user recognize or even understand special positions which is consistent with your goals of automation. The only problem, as usual, is interoperability. When one takes a phenix.refine model into Refmac or Shelxl the occupancy will be too high and this will surprise the users. Since interoperability is the wwPDB's business they should clarify what is expected to be in a PDB file. If it comes down to a vote based on what's already in the PDB, I think you'll lose. Dale Tronrud Ralf W. Grosse-Kunstleve wrote:
Ralf W. Grosse-Kunstleve wrote:
phenix.refine uses the correct symmetry factor automatically (internally). Unless you have an actual partial occupancy, the occupancy factor in the pdb file should be 1, irrespective of the multiplicity of the special position.
Doesn't that violate the mandate that the pdb file represent the contents of the AU? Unless I misunderstand Maia's structure, the Ca++ is shared between two asymmetric units, so each AU contains 0.5 of it?
That's the first time for me to look at the situation from that viewpoint...
My thinking so far:
Given the space group symmetry, the special-position information is in the coordinates. From that follows directly the multiplicity of the Wyckoff position, as listed in the International Tables. If the occupancy factor is convoluted with the
multiplicity_of_the_special_position ------------------------------------ multiplicity_of_the_general_position
quotient, the result is not an occupancy factor anymore since two distinct concepts are mixed.
But ultimately, what should be in the PDB occupancy column is a matter of convention.
Prompted by your message I looked at:
http://www.wwpdb.org/documentation/format32/sect9.html
This leaves me uncertain. Then:
http://mmcif.pdb.org/dictionaries/mmcif_std.dic/Items/_atom_site.occupancy.h...
Which says:
The fraction of the atom type present at this site. The sum of the occupancies of all the atom types at this site may not significantly exceed 1.0 unless it is a dummy site.
There is no mention of including the symmetry correction.
Is anyone aware of other resources that define what to do about the symmetry correction factor for atoms on special positions?
Ralf _______________________________________________ phenixbb mailing list [email protected] http://www.phenix-online.org/mailman/listinfo/phenixbb
participants (4)
-
Dale Tronrud -
Edward A. Berry -
Maia Cherney -
Ralf W. Grosse-Kunstleve