On Thu, Sep 15, 2011 at 11:17 AM, Dale Tronrud <det...@uoxray.uoregon.edu>wrote:
> The practical problem with this definition is that the break > point lies directly in the middle of the most likely value for > the chi angle. A lot of residues have chi angles near 90 deg and > small shifts in the positions of the atoms will necessitate relabeling. > The next cycle of refinement the labels may have to be restored. If > you do a bit of model building you may be required to relabel too. > > You can either stick to the convention, swap labels all the > time and rewrite all your analysis tools to check for label swapping, > or just wait until you're ready to deposit and fix them at the last > minute. The latter is what all refinement programs I've seen do. > Oh, I see. Phenix definitely doesn't do anything about this - I suppose we could add something to adjust the atom names before writing out the PDB file, if it helps prevent downstream errors. However, I suspect the pop-up in Coot is going to annoy a lot of people, since there are many ways for the atoms to be reversed. -Nat
