Has the wiggle function changed recently ?

Case number:699969-2009378
Opened by:Steven Pletsch
Opened on:Sunday, April 5, 2020 - 22:40
Last modified:Monday, April 13, 2020 - 18:51

I have noticed that the wiggle is taking many more cycles to complete, and it often gets stuck very easily.

It is making many scripts ineffective because of the longer wiggle times.

It is also making it more difficult to minimize proteins that have loop issues. The wiggle tends to get stuck, and one must stop, do a local wiggle on segments with issues, and then resume the wiggle all.

(Sun, 04/05/2020 - 22:40  |  5 comments)

bkoep's picture
User offline. Last seen 4 hours 40 min ago. Offline
Joined: 11/15/2012
Groups: Foldit Staff

Are you using a devprev Foldit client?

We haven't updated the main Foldit client since November, so nothing should have changed recently. We did just release a devprev update in March. We don't expect it to change Wiggle behavior, but if devprev is behaving differently, then we'll want to take a look a closer look at the March update.

Joined: 05/26/2008
Groups: Hold My Beer

It's the main client I believe (20191029-32048a6a12-win_x86), working with other puzzles it seems normal, but i have been having the issues with both of the ORF puzzles, much more noticeable on this newer one, maybe i just have a bad segment in there that is causing it. When wiggle is on Medium it is behaving much like it was when it is on Auto, where it keeps getting stuck easily, and loops don't seem to want to wiggle out properly.

I know it is difficult to explain a "feeling" about how it is working, but a single backbone walk on a fairly refined puzzle yielded ~100 pts with many segments contributing to the increase. Also reworking areas requires switching between local wiggles and wiggle all multiple times to get the area to minimize fully.

bkoep's picture
User offline. Last seen 4 hours 40 min ago. Offline
Joined: 11/15/2012
Groups: Foldit Staff

Can you share your solution with scientists? I can at least take a closer look at the solution to look for things that might trip up Wiggle.

Joined: 05/26/2008
Groups: Hold My Beer

It is shared, titled wiggle issues.

Hopefully it is something simple on my end. Thank you for your time :)

Joined: 09/29/2016
Groups: Gargleblasters

I have a few thoughts on what could be causing this (all of which I feel have a high likelihood of occurring at the same time and making it harder to determine which is the primary cause)...

#1 Potentially just due to sidechain interaction which is preventing the backbone from moving. If doing Wiggle-Backbone doesn't produce the slow-ticking point increase, but Wiggle-Sidechains does... I'd say that lends credibility to this as the cause. A 'Bounce-Wiggle' will probably solve it. (while wiggling, bounce the CI slider down for a split second, then back up; if it doesn't compact in that brief low-CI moment, then let it sit at the lower CI for a second or two)

#2 Also could come down to the Filters, as the game seems to want to keep those conditions and make the protein stiff. In those instances, I do a Bounce-Wiggle , which generally "frees" it up. I've always attributed it to

It's not always that simple, though. If the 'bind' has a lot of pent-up energy (in other words, is actually not entirely stable), a bounce will release that which in turn can lead to it no longer passing and therefore changing your design; in extreme cases, sends loops flying off into space.

#3 Simply a user "oops", which is caused by never removing the Blueprint Building Blocks, which are restraining your protein from wiggling further! Not that it doesn't want to, but those Ideal Loop block pieces are helping it keep its shape. :P



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, RosettaCommons