On Jan 28, 2009, at 7:22 AM, Katherine Parra wrote:
Hello!
I'm trying to get the hydrogen bonds between some selected residues during a trajectory, using per frame commands from tools. But when I get the possible hydrogen bonds formed between the selected residues I also get the following error message in the Reply Log:
Skipping possible acceptor with bad geometry: #0:2108@O
Wrong number of grandchild atoms for phi/psi acceptor #0:2108@O
What could be happening?
Hi Katherine,
A more verbose version of the error message might be: atom #0:2108@O has been classified as a phi/psi acceptor as per Mills et al (full cite in FindHBond docs) but the number of atoms bonded to #0:2108@O's neighbor atoms ("grandchild atoms") is in conflict with that classification.
The most likely cause is that #0:2108@O's bonds are wrong or that its neighbor atoms' bonds are wrong. There is also some chance that Chimera has computed the wrong atom type for #0:2108@O or its neighbors. If you were working with a single structure instead of a trajectory then it might also be possible that :2108 is a standard residue with missing atoms (Chimera assigns standard residue atom types using templates instead of computing them) but trajectories typically don't have missing atoms.
To investigate, you should have a look at #0:2108@O's atom type and it and its neighbor's bonds. To see the atom type open the command lines (Favorites->Command Line) and type "sel #0:2108@o" and then use Actions->Label->IDTAM type. To see what the atom type means, look at this table:
Atom Types in Chimera . If the atom type seems wrong, you can change it using the "setattr" command (
e.g. to change it to O3, type: "setattr a idatmType O3 #0:2108@o"). If the atom type seems okay, look at the bonds and particularly the neighbor's bonds. If there are extra or missing bonds you can fix that using the bond/~bond command (type "help bond" in the command line for info).
You could send me the trajectory to look at if you feel think the problem is wrong atom typing or Chimera getting the bonding information wrong, and I would try to fix it.
--Eric