1
1
|
I am wondering whether anyone knows if there is an established "best practice" on how many restraint violations (distance and dihedral angle) are acceptable in a refined ensemble. Specifically, I currently have an ensemble average of 4.8 distance restraint violations greater than 0.2 Angstroms, with no violations greater than 0.4 Angstroms. For dihedral angles, I have an ensemble average of 1.5 violations greater than 5 degrees, with the greatest violation having a value of 7.9 degrees. Do these values seem reasonable? |
|
1
|
Hi, I guess that you make three classes of distance range based on crosspeak height/volumn and you work on folded protein ? A deviation of 0.2 angstrom is really small compare to the precision of this method. Distance range class definition is a bit "arbitrary" : Because of crosspeak height/volumn depends on average inter-proton distance and correlation times of dynamical process underlie cross-relaxation. the latest is not know and cause this method coarse (additionnally, the lack of internal reference of the ratio between distance and crosspeak height/volumn). For dihedral, is it theorical geomatric restraint angle or J-coupling based dihedral angle measurement that undergo violation ? Any way, this deviation is too small to be unreasonable. Interpretation of this two deviations is a bit different if they occur in structured region or flexible region because flexibility averages constraint that can appear to be unrealistic but in reallity this particular value don't means any thing... hope that help. YM |
