HI Xi Huang, Your question is quite good one. The bond vector orientation is quite important and if your input structure does not have the vectors in the average orientation, you will start to see artificial motions. I would recommend you to read my review: d'Auvergne E. J., Gooley P. R. (2007). Set theory formulation of the model-free problem and the diffusion seeded model-free paradigm. Mol. Biosyst., 3(7), 483-494. (http://dx.doi.org/10.1039/b702202f). This talks about the artificial motions that arise dependent on the diffusion tensor anisotropy and rhombicity - both the artificial nanosecond motions of the Schurr 94 paper and the artificial Rex from the Tjandra 96 paper. This review also points to all the relevant literature (note there is not much at all). One problem with current model-free theory is that a single, averaged bond vector orientation is assumed. Therefore it is quite important to have a good starting structure. As I have discussed in my second 2008 paper (http://dx.doi.org/10.1007/s10858-007-9213-3 I think, though maybe it was the 2007 paper), I would recommend comparing the dynamics you see from the final results to that of the local tm models. This is a good test of consistency (in a way complementary to Sebastien Morin's relaxation data consistency testing analysis in relax), but be aware that the local tm models are sometimes not very stable (the tm value can sometimes head off to weird values). I would also run the model-free analysis on a number of your structures and simply compare. If you use Gary Thompson's multi-processor code built into relax and have access to a multi-core, multi-cpu, or clustered systems, then you should be able to blast many, many structures through and compare the results. I hope this information helps. Regards, Edward P. S. If you manage to theoretically solve and eliminate this bond vector problem from the base model-free theory, that would be quite an achievement and one very impressive paper! On 8 October 2012 18:02, Xi Huang <huangxi987412@xxxxxxxxx> wrote:
Hi Edward, The protein I am working with has several conformations and the exchange rate is in around 100micro seconds timescale. I am wondering when I choose different pdb structure as an input file to "relax", how big the final result (S^2) will differ from each other? Is there any literature talking about the influence of choosing structure coordinates? Thanks for your help -- Xi Huang PhD Candidate, Division of Physical Chemistry Gail E. Fanucci Research Group Department of Chemistry University of Florida _______________________________________________ relax (http://www.nmr-relax.com) This is the relax-users mailing list relax-users@xxxxxxx To unsubscribe from this list, get a password reminder, or change your subscription options, visit the list information page at https://mail.gna.org/listinfo/relax-users