On Wednesday 03 February 2010 10:04:36 Pavel Afonine wrote:
I always suggest to run these refinement jobs and see which
> one gives the best result:
>
> 1) TLS + individual isotropic ADP refinement (tls+individual_adp);
> 2) TLS + group individual ADP refinement (tls+group_adp);
I do not think that your (2) is a physically plausible model.
I would dis-recommend this option.
> 3) individual isotropic ADP refinement;
> 4) group individual ADP refinement (with one refinable B per residue);
> 5) group individual ADP refinement (with two refinable B per residue).
I also tend to agree with Ed that option (5) has little physical
plausibility in its favour. Better to stick with individual Biso
and tight B restraints.
cheers,
Ethan
> + on top of it you can add automatic weights optimization to remove the
> arbitrariness in what is "loose" and "tights" restraints.
>
> This will give the conclusive, rock solid answer about which ADP
> parameterization and refinement protocol is good for given model and
> data. An alternative is an endless speculation.
>
> In the future the proper parameterization will be chosen automatically
> based on a large array of data, model and map characteristics.
>
> > The existing
> > two-adp-groups-per-residue implementation (CNS and phenix) is, imho, an
> > example of *improper* parametrization.
>
> As you see, in phenix.refine you can combine any B-factor refinement
> strategy with any (group, individual iso, aniso, tls), and apply it to
> any selected part of your structure. So, I assume at this point of the
> software automation, it is up to a smart researcher to decide which
> refinement strategy to use. You cannot blame the software for giving you
> the freedom to do what you may want to do. For example, if you choose to
> refine two-B-factors per residue when you could safely refine individual
> B-factors, it will be the example of improper parameterization that you
> have chosen (and not the software did for you). In phenix.refine
> technically you can refine individual isotropic or anisotropic B-factors
> at any resolution - the program will not crash, and it will be up to the
> user whether he/she will enjoy the results. Like I said, in the future
> the model parameterization will be done automatically.
>
> I guess I'm going off this discussion - otherwise phenix.refine will get
> less new options in the future if I keep writing -:)
>
> All the best!
> Pavel.
>
>
--
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle 98195-7742
