On Sat, Apr 20, 2013 at 11:17 PM, Ethan Merritt <[email protected]>wrote:

> This is far from the first time I've ranted on the subject.
> What is really needed is a "TLS group" field in the ATOM record so that
> each atom is unambiguously associated with a TLS group that
> describes its displacement.  No separate ANISOU needed or wanted.
>

I think migrating to mmCIF may finally solve this issue (among other
things, it may actually have a formal specification for TLS groups).

[aside: I know that phenix can refine individual aniso ADP components
> on top of the TLS, but IMNHSO this is not a correct thing to do].
>

I should check this with Pavel, but my understanding is that the program
wouldn't actually allow both TLS and anisotropic ADPs for the same atoms -
just TLS for one set of atoms, aniso for another.

That, I believe, is false.  There is no program that refines TLS
> groups without writing out the refined parameters in the output
> file.
>

Sorry, I was unclear - I meant that other (non-refinement) programs which
manipulate PDB files don't reliably preserve remark records.  (Including
many in Phenix, I'm afraid.)

-Nat

Reply via email to