Hi, I have to say that I am definitely not an expert in this subject. When I have this issue I just use Molmol to add the protons, which I would assume is based on ideal geometry. For any simulation I would guess that the proton will move around in and out of the peptide plane. So, if you can assume that this motion is accurately simulated which, if done correctly, these fast motions should be then maybe the average position would be the best for model-free analysis. Be careful though because this is a whole can of worms you could be opening here! Model-free analysis (as well as many other theories in the field) assumes in its theory that you have a defined bond vector orientation but the analysis implies motion which in turn means the vector moves again in turn changing the results of the analysis. Note the circular loop of dependence and the lack of theory addressing this - this really is not an issue for the faint hearted! Regards, Edward On 10/24/07, Sebastien Morin <sebastien.morin.1@xxxxxxxxx> wrote:
Hi Ed Thanks for the information about the relative PDB orientation and Euler angles... For adding hydrogens to a PDB, I wonder what approach is best so that it can be used (in relax) for non isotropic relaxation analysis. In other words, what approach would be best to get good NH vector orientations ? I have a PDB at 1.95 A resolution and without any hydrogen. I tried two different approaches to get the hydrogens in a good orientation... 1. Hydrogens were added inside CHARMM and then a short Langevin minimization (200 ps) was done. 2. Hydrogens were added inside CHARMM (as for approach 1) and then a molecular dynamics simulation was performed (by another PhD student) with TIP3P water and 3 snapshots were taken at time 10, 12 and 14 ns (after 2 ns equilibration)... I could take snapshots up to 40 ns, but they should be equivalent to others... My questions : What approach do you think is the best to get good NH orientations ? Would it be relevant to take different snapshots from MD and use them in separate relax runs ? Would you suggest another approach ? Cheers Sébastien :) Edward d'Auvergne wrote: Hi, The only differences in the analysis should be due to different bond orientations. The PDB reference frame doesn't matter - this will only change the values Euler angles of the diffusion tensor and nothing else. So if the same structure exists in two different PDB frames, then the difference in the optimised Euler angles will be solely due to the rotation between the two structures. If you were to change the PDB frame in any of your files, the final chi-squared value should be identical to a non-rotated PDB. The differences you do see are because the relative XH bond orientations in the different structures are different. Regards, Edward On 8/30/07, Sebastien Morin <sebastien.morin.1@xxxxxxxxx> wrote: Hi, I am doing different tests with the full_analysis.py script, data at three magnetic fields and different PDB structures (taken from MD simulations). I realized that those different structures give rise to different optimization times and also to different chi-squared values for the different diffusion tensors (I have only tried the prolate diffusion tensor). I wonder if the structures need to be processed before giving as inputs for relax. Especially, do the structures need to be centered relative to their inertia tensor (like with the pdbinertia program that one should use prior to model-free analysis using ModelFree). Thanks ! Sébastien :) -- ______________________________________ _______________________________________________ | | || Sebastien Morin || ||| Etudiant au PhD en biochimie ||| |||| Laboratoire de resonance magnetique nucleaire |||| ||||| Dr Stephane Gagne ||||| |||| CREFSIP (Universite Laval, Quebec, CANADA) |||| ||| 1-418-656-2131 #4530 ||| || || |_______________________________________________| ______________________________________ _______________________________________________ relax (http://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 -- ______________________________________ _______________________________________________ | | || Sebastien Morin || ||| Etudiant au PhD en biochimie ||| |||| Laboratoire de resonance magnetique nucleaire |||| ||||| Dr Stephane Gagne ||||| |||| CREFSIP (Universite Laval, Quebec, CANADA) |||| ||| 1-418-656-2131 #4530 ||| || || |_______________________________________________| ______________________________________