On 10/11/06, Chris MacRaild <c.a.macraild@xxxxxxxxxxx> wrote:
On Wed, 2006-10-11 at 18:57 +1000, Edward d'Auvergne wrote:
>
> 3.5.3 The 'spin' user function class
>
> This new user function class could contain functions such as:
>
> spin.add()
> spin.copy() # Copy the spin info (name and num) from another pipe.
> spin.delete()
> spin.display()
> spin.read()
> spin.sort()
> spin.write()
>
> These functions could be applied selectively using the 'mol_num',
> 'mol_name', 'res_num', or 'res_name' arguments.
One extension to the current model which might have value is the idea of
a spin group. This would describe a group of spins for which relaxation
has been measured which are believed to undergo the same relaxation
causing dynamic process. This idea is less relevant to classic
model-free analyses due to the timescales involved, but might be more
important for studying ns motions by extended MF or slower motions by
relaxation-dispersion type approaches. To make it clear what I mean,
imagine a single peptide bond where you have both NH and Ca relaxation
and where there is evidence that ns motions are involved (or
equivalently a nucleobase with several relaxation probes). You might
hypothesise that the motion involved is coordinated motion of a peptide
plane which is essentially rigid on this timescale. I seems to me that
relax could be well placed to formulate and test such an hypothesis.
Mathematically, it implies that the spectral densities for the
respective spins are related by Wigner rotation terms that relate the
relative orientation of the respective DD and CSA vectors. Of course all
of the requisite maths would need to be coded before this was of any
use, but the spin-group concept would be a first step.
Brilliant idea! I would like to extend the concept even further
though. What about the idea of linking two parameters of the same
type? This would be equivalent to sharing the same parameter. For
example in residues 3, 4, and 5 in a protein (or multiple spins in a
single RNA base) could all share the same ts parameter? Or the same
S2s parameter? This is something I have thought of before. Have a
look at the Tjandra et al., 1995 paper from Eur. J. Biochem - in that
paper all residues of calmodulin were optimised with a single, hence
global, ts value.
Given its lack of immediate applications in relax, it might be deemed a
complication too far, but its worth considering in the context of a big
re-write.
This idea is outside of the scope of the proposed data model change -
it's implementation should not effect on the data structures. The
linking concept may require additional data structures, but the core
molecule-residue-spin system as well as overhaul of the 'run' system
and renaming it to the 'pipe' system should not be affected. But
please don't slow down with the ideas! Someone could, at the end of
these discussions, collect all the ideas which are outside of the
scope of the proposal and write a very short summary of each idea
linking back to the mail archive post in which the idea was first
raised. This could (or should I say should) be sent to relax-devel to
be archived for future reference/implementation.
Edward
Re: Redesign of the relax data model: 3. Molecules, residues, and spins