Hello Jian--
> I tried to back-calculate chemical shift anisotropy (CSA) from a
> model that's generated by XPLOR-NIH but my results seem to be off.
> In XPLOR-NIH model calculation the model is already aligned along
> the z-axis as in AssignFit by:
> vz=unitVec(oTensor.zAtom().pos() - oTensor.oAtom().pos())
> vy=unitVec(oTensor.yAtom().pos() - oTensor.oAtom().pos())
> vx=unitVec(oTensor.xAtom().pos() - oTensor.oAtom().pos())
> AtomSel("all").apply( Rotate(transpose(Mat3(vx[0],vy[0],vz[0],
> vx[1],vy[1],vz[1],
> vx[2],vy[2],vz[2]))) )
> sigma11 = 57.3, sigma22 = 81.2, and sigma33 = 227.8, sigma_iso =
> 122.1, so for the input to "setsigma" we used, sigma_xx = 64.8,
> sigma_zz = -105.7 and sigma_yy = 40.9. Beta = -17.
> In XPLOR-NIH the CSA is calculated as
> calcedShift = sum_{i,j} A_i sigma_j * cos(theta_{i,j})^2 / DaScale
> I calculated it following "Application of solid-state NMR restraint
> potentials in membrane protein modeling", and it's
> sigma = sigma11 * e1z * e1z + sigma22 * e2z * e2z + sigma33 * e3z * e3z
> I figured those two ways of calculating CSA should be equivalent,
> but what I got from the latter way is different from what's given in
> the XPLOR-NIH model viols file.
The definition of the CSA tensor is rather subtle. Please carefully read
http://nmr.cit.nih.gov/xplor-nih/doc/current/python/ref/csaPot.html
and this document:
http://nmr.cit.nih.gov/xplor-nih/doc/current/python/ref/csa-methods-marvin.pdf
Please get back to me if you are still having difficulties reproducing
your results.
best regards--
Charles
