Hi Tim,
just had finally a chance to check this again in the latest release 0.8.1 just to make sure this was not an artifact of using 0.7.2 for a long time. Adding terminal residues into a structure which contains explicit hydrogens does not add residues with hydrogens. Mutating an existing residue gives a slightly different picture, the CA and N hydrogens were retained, while the other side-chain hydrogens were lost. So at least in this version the issue still persists and caution is warranted.
As to your comment of explicit vs. implicit hydrogens, I believe this has been discussed exhaustively on the ccp4 BB a while back. No ultimate conclusion was reached. I personally would side with explicit hydrogens. Pavel, I believe, made a good argument that the methodology for adding hydrogens can be better reproduced when they are written out explicitly vs. implicit treatment in the application. As you said: A matter if choice or preference....
Cheers,
Carsten
-----Original Message-----
From: Tim Gruene [mailto:[log in to unmask]]
Sent: Wednesday, May 27, 2015 6:05 PM
To: Schubert, Carsten [JRDUS]
Cc: [log in to unmask]
Subject: Re: Dragged refinement fails with riding hydrogens
Hi Carsten,
I don't remember such problems in Coot when I was working with explicit hydrogen atoms with neutron data, as long as the atom names were matching the cif-files.
If the model is not refined against neutron data, there are, in my opinion, good reasons to follow the refmac policy and not write the hydrogen atoms to the output PDB, so there is not so much reason for Coot to change. I guess, though, this topic is as personal as whether one should remove side chains when you don't see density.
Best,
Tim
On Wed, May 27, 2015 at 04:53:58PM +0000, Schubert, Carsten [JRDUS] wrote:
> Hi Tim,
>
> that does not work for phenix, not sure about Buster/TNT; refmac behavior is different. In phenix, which uses a riding model as well, but defines explicit hydrogens, one has to be very conscientious to add hydrogens back after mutating or adding residues. In part this is a phenix issue as well, but all in all I do have to side with Oliver on this one. Coot is not very well adapted to a handle explicit hydrogens and could benefit from improvement in that regard.
>
> Cheers,
>
> Carsten
>
> -----Original Message-----
> From: Mailing list for users of COOT Crystallographic Software
> [mailto:[log in to unmask]] On Behalf Of Tim Gruene
> Sent: Wednesday, May 27, 2015 12:16 PM
> To: [log in to unmask]
> Subject: Re: [COOT] Dragged refinement fails with riding hydrogens
>
> Hi Oliver,
>
> unless these are neutron data, you could simply remove the hydrogen atoms from the model before model building. They would be regenerated on the next round of refinement. At least this is the philosophy of some refinement programs.
>
> Cheers,
> Tim
>
>
> On Wed, May 27, 2015 at 04:16:38PM +0100, Oliver Clarke wrote:
> > Hi all, dragged refinement doesn't seem to work for structures with riding hydrogens - see linked screen recording showing dragged refinement of the same structure with and without hydrogens.
> >
> > https://www.dropbox.com/s/15xs8ylhrx98ljk/dragged_refinement_H_bug.m
> > ov
> > ?dl=0
> >
> > Would it be possible to alter this behaviour at some point such that hydrogens remain firmly attached during dragging, to remedy this?
> >
> > In the past I have got around this by just doing real space corrections in the absence of any hydrogens and adding them back in before every cycle of reciprocal space refinement, but this does not seem like the optimal way to be doing things.
> >
> > Cheers,
> > Oliver.
> >
>
> --
> --
> Dr Tim Gruene
> Institut fuer anorganische Chemie
> Tammannstr. 4
> D-37077 Goettingen
> phone: +49 (0)551 39 22149
>
> GPG Key ID = A46BEE1A
>
--
--
Dr Tim Gruene
Institut fuer anorganische Chemie
Tammannstr. 4
D-37077 Goettingen
phone: +49 (0)551 39 22149
GPG Key ID = A46BEE1A
|