Frozen backbone can move

Case number:845813-1998899
Topic:Game: Tools
Opened by:Bletchley Park
Opened on:Sunday, December 14, 2014 - 22:54
Last modified:Sunday, December 21, 2014 - 09:53

I find that a fully frozen backbone can move when wiggling all. I would have thought it was really frozen. If this is not a bug, please explain why the frozen backbone can still move.

(Sun, 12/14/2014 - 22:54  |  6 comments)

jeff101's picture
User offline. Last seen 18 hours 28 min ago. Offline
Joined: 04/20/2012
Groups: Go Science

I think I saw this when I first made DoBandsObeyHookesLaw1 ( I wanted to freeze the protein and shuttle it back and forth. If the conformation doesn't change, the score should stay the same. Yet I did see the score change slightly.

Joined: 05/19/2009
Groups: Contenders

I'm glad you noticed it too Jeff, I can see the protein backbone change shape while frozen, that seems incorrect to me.

spmm's picture
User offline. Last seen 7 weeks 6 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

Agree it is easy to see the 'frozen' bits move, most apparent in the ends of sheets, but also makes a mess of helixes in a particularly unpleasant fashion :(
It has always been like that as far as I can tell.

jflat06's picture
User offline. Last seen 1 day 11 hours ago. Offline
Joined: 09/29/2010
Groups: Window Group
Type: Bug » Suggestion

Not really a bug, but could be a feature -

The general case is that you have a piece of backbone with X number of segments frozen (whether it be none/some/all). The correct way to handle this is to allow the rigid body degrees of freedom to vary. If they aren't allowed to vary, you see weird artifacts where one residue in the chain seems 'stuck' in place, with everything else hinging around that.

That paricular problem isn't an issue when ALL segments of a chain are frozen, so you could imagine handling that case differently. But it isn't inconsistent with the general behavior. There may be cases where people actually WANT the whole frozen chain to move, so there's a bit of a design decision there as well.

My personal favored approach on this would be to keep things as-is, but add a new 'pin' tool. This would let you pin 1 residue in any particular chain, so that it was completely locked in space (you can only do this with a single residue per chain). That would let you completely lock a frozen chain in space, and would also give you more control over where wiggle places the "root" of the chain (right now, it usually just places it in the middle).

xiando's picture
User offline. Last seen 7 years 27 weeks ago. Offline
Joined: 10/19/2008
Groups: Oma Gawd

Hasn't it always been the case? I don't ever remember a time when a frozen section literally stayed frozen in position during a global wiggle (do you still use that term?), from May of 2008 until I stopped playing daily, and it has behaved the same way each time I've returned, although I'll admit I'm only here now and then these days.

Joined: 05/19/2009
Groups: Contenders

To the best of my 5+ year knowledge, a frozen backbone never moved, not even rotating within itself. I'll observe further, thanks.


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