-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hello Andrew--
>
>
> When determining Da and Rhom, is it acceptable to lump all the (scaled)
> data into one file and determine a single Da and Rhom for the whole lot?
>
> The convention in the literature seems to be to determine Da and Rhom
> for one data set and the estimate Da for all other data sets by scaling
> the calculated Da by the appropriate gyro_mag/bond length ratios. The
> current Xplor_NIH convention seems to be to scale all data to NH and
> then choose a single Da and rhom for all data sets that is close to the
> Da and rhom values for the individual scaled data sets. I'm a little
> leery of these using either of these conventions because I have few data
> points (21 CH and 10 NH), and if I run the scaled individual data sets
> through SVD I get fairly different values for da and rhom between the
> two data sets (14.04/.093 for scaled CH and 8.86/.057 for NH). {For
> comparison, I ran the individual data sets in eginput/gb1_rdc through
> SVD and got very similar Da and rhom values} My thought was then to lump
> the scaled data together since, in general, more data points = more
> accurate results. Is this a fair assumption? Are there better ways I
> could go about this?
>
I'm not certain what you mean by your question. Whatever method you use,
all observed rdcs should be used to determine the Da and Rh. If you have
pre-calculated structures, eginput/protG/calcTensor.py does this using
SVD. If you have no structures, use calcDaRh and use that as a starting
point for eginput/gb1_rdc/anneal.py, followed by refine.py. anneal.py
fixes Da/Rh until final minimization. refine.py allows Da/Rh to float
throughout. In all cases, all available data should be used (from all
experiments).
best regards--
Charles
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8+ <http://mailcrypt.sourceforge.net/>
iD8DBQFHaBSmPK2zrJwS/lYRAoeOAJ9+I4NAFjYRc4Z/2poxobrdAuqv8ACfTQMN
PkJM6m38Zwfq3lu99NUg9Ag=
=09bj
-----END PGP SIGNATURE-----
