[phenixbb] NCS

[Pavel Afonine]

Hi Jeff,

> This is a great point. There should probably be NCS output in the PDB 
> header.
>

yes, it's definitely good to have, but can we really say more than 
something like:

REMARK 3 NCS RESTRAINTS: TYPE TORSION

?

> > 2. What is difference between NCS type, global and torsion angle, 
> and when they are preferred over the other?
>
> Global NCS uses Cartesian-based restraints, which force the matching 
> chains to be identical.
>

Just a minor correction: what you described are NCS *constraints*, and 
we don't have them in phenix.refine yet. NCS restraints force the 
coordinates and B-factors (residual B-factors if TLS is used) of NCS 
related copies to be *similar* (but not necessarily identical). The 
degree of similarity is determined by a parameter, and this parameter 
cannot be set such that NCS restraints become NCS constraints.

All the best,
Pavel