Some movable sidechains get stuck

Case number:845799-936229
Topic:Biochem
Opened by:dejerpha
Status:Closed
Type:Bug
Opened on:Saturday, January 31, 2009 - 17:11
Last modified:Tuesday, June 2, 2009 - 01:52

Sometimes, a sidechain that was perfectily movable a moment before gets "stuck".

It happened to me with most recently with 32:Valine in puzzle 124.

Scenario:

- move sidechain manually
- wiggle
- try to manually move it back to its previous position: impossible.

(Sat, 01/31/2009 - 17:11  |  6 comments)


xiando's picture
User offline. Last seen 6 years 50 weeks ago. Offline
Joined: 10/19/2008
Groups: Oma Gawd

I have seen the same. I attribute it to some ratcheting effect but have wondered about its real cause as well and would also like an explanation.

admin's picture
User offline. Last seen 24 weeks 3 days ago. Offline
Joined: 11/10/2007
Groups: vi users

Assigning to our friendly biochemist.

Joined: 11/10/2007
Groups: Window Group

Each sidechain has a number of degrees of freedom (DOFs) that determine its position. Pulling on a sidechain selects these DOFs from a discrete set called rotamers. Other operations, such as wiggle, may change these DOFs to ones that are not in the rotamer set used by the sidechain pull.

Currently, the rotamer sets are fixed for a puzzle, and you may end up able to pull a sidechain out of a position, but not back into it, in a single pull. In the development version, the rotamer sets are updated to include the current rotamer, which should reduce the occurrence of this problem.

Joined: 05/10/2008

Hello,

This problem is still occuring. Can you take a look at it again?

In the 152 puzzle, my current solution has at least 2 sidechains that cannot be adjusted manually to their "guide" position. Which is rather strange, as the guide is supposed to show an actual solution.

My advice: don't be too rigorous concerning DOFs. Sometimes one adjusts something just to create clashes.

Joined: 12/14/2008

I experienced it also e.g. in the last (150) or the one before Rosetta puzzle.

It's probably because the arc for pointing the mouse to to get the sidechain move is shrinking from the wiggle?
I had one case where I couldn't get the sidechain back into the original position, but it flipped to a very different with nearly just clicking on it. It switched between the 2 and I got it to work only after undo the move-wiggle and do it again and then in the third try of moving the sidechain back.

Judecca's picture
User offline. Last seen 1 year 39 weeks ago. Offline
Joined: 05/09/2008
Groups: None

sounds like it has to do with which library is used to determine the rotamers you can pick from... it appears the game is using the "backbone-dependent" library (bbdep02.May.sortlib). so after certain sidechain changes and a wiggle, the backbone is no longer in the same angles for it to be considered a favorable position to move back. It would be kinda interesting to see how differently we would fold with the "backbone-independent" libraries loaded in there..... could also throw us completely off track :D

see http://dunbrack.fccc.edu/bbdep/

[Certain rotameric states are higher in energy than others because of steric interactions with neighboring atoms. Electrostatic interactions including hydrogen bonds also affect side-chain energies. These interactions can be "backbone-independent", that is, not depending on the conformation of the local backbone of the residue or "backbone-dependent", that is, depending on the local backbone conformation as determined by the backbone dihedrals phi and psi.

The backbone-dependent rotamer library provides average dihedral angles, standard deviations in dihedral angles, and and probabilities of protein side-chain conformations as a function of the backbone dihedrals phi and psi. The May 2002 version of the library is now available. It was built from data contained in 850 chains from the PDB of resolution 1.7 Angstroms or better. We also use the quality criteria described by Lovell et al. to eliminate side-chains from the data set with questionable coordinates.]

Sitemap

Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, Boehringer Ingelheim, RosettaCommons