Rubber bands seem weaker when wiggling; "bander" scripts not scoring like they used to

Case number:845813-997090
Topic:Game: Tools
Opened by:Susume
Opened on:Saturday, February 22, 2014 - 01:56
Last modified:Monday, February 24, 2014 - 19:30

Rubber bands seem weaker when wiggling than they used to be. It takes longer for the protein to start to move, and it doesn't move as much as it used to with bands. This is true in hand folding ("w" key) and scripts (structure.WiggleAll, do_global_wiggle_all).

Many of our "bander" scripts (the ones that place bands and then wiggle, hoping to move the protein) are not working well, and this may be part of the reason. These scripts are a mainstay; I believe most people use one or more of them on every puzzle. It seems like the bands are just not pulling as hard as they used to, so they are not moving the protein much.

(Sat, 02/22/2014 - 01:56  |  2 comments)

Joined: 09/21/2011
Groups: Void Crushers

That should not really be the reason why those scripts do not perform as we used to see.
Most bander scripts adjust the strength until a certain percentage loss is reached.
But the length of the stabelizing wiggle is probably not high enough to compensate for the lower reparing speed of wiggle.

brow42's picture
User offline. Last seen 1 year 37 weeks ago. Offline
Joined: 09/19/2011
Groups: None

I've found that wiggle(2) on Low hits stability very fast, often on the first loop. I've increased the minimum # of wiggles but it seems scripts just work better on Med and High.


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