The difference between a rotamer with the possibility of 0% and an allowed rotamer is sometimes not discernable by eye. I always try autofit rotamer in Coot on this type of sidechain, and I often end up with something that, by eye, fits the density as well as the original conformation and is judged perfectly fine by MolProbity. If the automatic fit doesn’t work, it’s easy to hit undo and go back to the previous conformation.
Of course, sometimes, after refinement, the sidechain returns to the original low-probability conformation. Refinement always gets the last word over manual buidling. Sue > On Feb 14, 2015, at 5:44 AM, Robert Immormino <immorm...@gmail.com> wrote: > > Dialing Pretty, > > It is certainly possible to observe real conformations of amino acids, > especially Arg, that are not quite common enough to have been identified as > rotamers. To this point, note that defined rotamers accounted for only 82% of > Arg conformations in the highly validated data set used to generate the > MolProbity rotamer database. > http://kinemage.biochem.duke.edu/databases/rotamer.php#soPRL > > If your Arg conformation fits your electron density, and does not have any > internal or external steric clashes the conformation is likely real. As > examples like yours are added to the database more rotamers may be defined! > > Best, > -bob > > On Fri, Feb 13, 2015 at 1:12 AM, Dialing Pretty > <000003f1d08ed29c-dmarc-requ...@jiscmail.ac.uk> wrote: > Dear All, > > I have a set of data, in which there are very nice electron density for > several Arg residues. However MolProbity analysis demonstrated that the > percentage of the Arg rotamer in my data is 0%. Although there is the > possibility that the Arg rotamer in my data is 0%, I still wonder is any > possibility that to correct the Arg rotamer in my data so that it would > belong to the scope of commonly observed Arg rotamer. > > I am looking forward to getting a reply from you on it. > > > Dialing > > >
