[phenixbb] Fwd: Re: help with refinement in Phenix

hosur at barc.gov.in hosur at barc.gov.in
Mon Jun 16 23:07:53 PDT 2008


Hi PHENIX developers,
I am forwarding herewith my correspondence with Dr. Pavel Afonine regarding a
reinement issue in PHENIX.
I would appreciate if other developers can help me with the solution before
Pavel comes back on June 25th. Thanks.
The requirement is to be able to refine a model in which a piece of electron
density is interpreted as partial occupancy protein residue and a set of water
molecules.
Hosur M.V.

----- Forwarded message from Pavel Afonine <pafonine at lbl.gov> -----
    Date: Sat, 14 Jun 2008 21:57:36 -0700
    From: Pavel Afonine <pafonine at lbl.gov>
Reply-To: Pavel Afonine <pafonine at lbl.gov>
 Subject: Re: help with refinement in Phenix
      To: hosur at barc.gov.in

Hi Hosur,

I'm out of town and not regular at email, so the delay with the answer,, 
sorry for this. Actually, it is the best to address your questions to 
phenixbb (http://www.phenix-online.org/mailman/listinfo/phenixbb) so all 
the PHENIX developers get it and you don't hang up with the delay if a 
particular person is temporarily absent. I will be back in office on 
June 25.

Regarding your question:

> In my case, the set of 4 water molecules has to have the same occupancy, and
> this occupancy with the occupancy of a side chain must make up unity. That is
> the sum of two occupancies must be unity.
>   

This particular feature is currently not available in phenix.refine. I 
will add this option once I'm back in office.

> If I understand correctly, if i identify each of the four water molecules by a
> different altloc label, then the sum of 5 occupancies will be adjusted to one,
> with the possibility of each water occupancy converging to different value.
>   

Yes, this is expected behavior.

> Further, the final occupancy would be close to 0.25 of the required value. The
> complication, in my opinion, is because the set of four water molecules is not
> like set of atoms that make one residue which can be assigned a single altloc
> label. I hope I have been able to convey what I mean. Let me have the suitable
> PDB, which I can try out.
>   

Yes, this is very clear for me and like I said above I will add this 
feature to PHENIX as soon as I get back to office.

All the best,
Pavel.

----- End forwarded message -----




-------------------------------------------------




More information about the phenixbb mailing list