Hello Leonidas-- > Indeed I dont use scale_toNH since my data are prescaled. So > apparently i do have to put them as raw (without prescaling) and let > XPLOR scale everything (both values AND weight of energy terms). Do i > understand correct? >
If you prescale the RDCs, you should scale the energy terms, as in eginput/gb1_rdc/refine.py in the Xplor-NIH distribution. If you instead use scale_toNH, then the relative energy scale factors can be 1. But as Gabriel remarked, do watch for sign problems. > The violations that I get are mainly ANGL violations not from RDCs. > If the violations are small it *may* be no problem, but you might try disabling any Cartesian minimization (i.e. always work in torsion-angle space) to see how much worse the fit is. best regards-- Charles _______________________________________________ Xplor-nih mailing list [email protected] https://dcb.cit.nih.gov/mailman/listinfo/xplor-nih
