-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello Andrew--

>     I am using RDCs to refine an nmr structure and I want to make sure 
> I'm setting things up correctly.  I have 1 bond CH and NH RDCs from the 
> same sample in the same alignment medium with the same measurement error.
> 
> 1)RDCs for methyl groups were multiplied by a factor of -3 to compensate 
> for the effects of rotation, and the angle defined along the C-C bond.  
> (A bit of confusion here, some sources say multiply by -3, and some say 
> divide.  Also, if I'm defining the methyl CH RDC using the C-C bond, do 
> I need to rescale the data relative to CC, and then again to NH?)

start with eginput/gb1_rdc/anneal.py. The comments in the script walk
you through this. Easiest is to leave the rdcs unscaled and uncomment
the scale_toHN line.

>  
> 2)I scaled (by hand) the CH RDCs to NH by multiplying them by a factor 
> of about -1/2, using values for gyromagnetic ratios and bond lengths 
> from rdcPotTools.py, and the fact that the gyromagnetic ratio is 
> negative for Nitrogen and positive for Carbon.

- From the comment at the top of python/rdcPotTools.py:

  protein residual dipolars are usually normalized relative to
  measurements of N-HN - this is achieved by calling the function
  scale_toNH. This normalization involves the gamma_A * gamma_B /
  rAB^3 prefactor (see the docs for <m rdcPot>). 15N has a negative
  gyromagnetic ration, but it was deemed simpler to treat as positive,
  so that if there is an experiment which doesn't include 15N, the
  dipolar coupling sign must be flipped. If you would rather not use
  this convention, call the function correctGyromagneticSigns() at the
  beginning of your script, and then the correct signs will be used
  for the Da prefactors.


> 
> 3)I used gridsearch.inp to find two sets of Da and rhom values, one set 
> for the scaled CH data, and one for the NH data.  (If I have to rescale 
> the methyl CH RDCs to CC, do I need a third set of Da and rhom
> values?)

If you have structures you should use
eginput/protG/calcTensor.py

This is faster and more accurate than a gridsearch, and allows you to
easily incorporate all of your rdcs.

> 
> 4)Force constants for CH were multiplied by about 2 relative to NH (the 
> inverse of the scaling factor used in step 2)
> 

Again, see gb1_rdc/anneal.py.

best regards--
Charles
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8+ <http://mailcrypt.sourceforge.net/>

iD8DBQFHUggUPK2zrJwS/lYRAs6+AJ0RsnqnP9iW/XL8XD3fRiCuZ6bG0gCfbIVB
toVyf6sCC1NuMhKXtyMuskQ=
=73o9
-----END PGP SIGNATURE-----

Reply via email to