Blueprint: Wiggling at ci 1.0 reduces the score

Case number:954892-2003007
Topic:Developer Preview
Opened by:tokens
Status:Open
Type:Bug
Opened on:Thursday, October 27, 2016 - 09:11
Last modified:Thursday, November 3, 2016 - 19:09

After some mutating at shaking at 0.1 ci, I tried wiggle at 1.0 ci. This reduced the score.

Before wiggling:
http://fold.it/portal/files/chatimg/irc_370468_1477558805.png

After wiggling:
http://fold.it/portal/files/chatimg/irc_370468_1477559186.png

After going to the blueprint window and removing the blueprints, wiggling at full ci worked normally again.

(Thu, 10/27/2016 - 09:11  |  8 comments)


tokens's picture
User offline. Last seen 15 weeks 4 days ago. Offline
Joined: 11/28/2011

Shared "Try wiggling at ci 1.0" with scientists.

tokens's picture
User offline. Last seen 15 weeks 4 days ago. Offline
Joined: 11/28/2011

Looking further at the shared solution it looks like the blueprints have been moved 1 to the right (or is it to the left?) compared to where I originally put them. This probably happened by me doing the following:

I removed 1 segment at the beginning of the protein and added one segment at the end. Then I decided that that didn't improve the protein so I undid the changes.

tokens's picture
User offline. Last seen 15 weeks 4 days ago. Offline
Joined: 11/28/2011

Ok. I'm able to reproduce the bug:

1. Insert a blueprint (and mutate loop segments to appropriate acids).
2. Wiggle.
3. Delete first segment. The blueprint moves one to the left as expected.
4. Undo. The blueprint doesn't move and is now misplaced.
5. Wiggle. The score goes down.

jflat06's picture
User offline. Last seen 20 hours 57 min ago. Offline
Joined: 09/29/2010
Groups: Window Group

This is actually intended behavior - sorry for not fully explaining the tool yet.

Inserting Building Blocks onto your pose will also attach constraints internally to those segments. These constraints keep that building block close to the shape that was inserted.

This is important because we're specifically interested in your final designs containing these shapes, and without constraints, designs will tend to drift away from the blocks that are inserted.

The building blocks included in the tool are highly stable, ideal fragments. So the closer your final design is to being composed entirely of building blocks, the more likely that your structure is to fold up in the lab.

tokens's picture
User offline. Last seen 15 weeks 4 days ago. Offline
Joined: 11/28/2011

Thanks for clearing that up. But shouldn't undoing the deletion of the first segment also move the building blocks back to their original place?

jflat06's picture
User offline. Last seen 20 hours 57 min ago. Offline
Joined: 09/29/2010
Groups: Window Group

It should - I'm not sure what's going on there. It might just be the UI failing to update on undo. Either way, looks like a bug.

Susume's picture
User offline. Last seen 23 hours 17 min ago. Offline
Joined: 10/02/2011

I notice that dragging a rectangle off of the blueprint leaves the ideal loop in place - does dragging the rectangle remove the constraints on that ideal loop? If not, how do we remove the constraints? I understand you want more building blocks, but sometimes I don't care if one ideal loop is replaced by another during a rebuild or remix script. I may not have chosen the best ideal loop for that spot, and another one might score better.

jflat06's picture
User offline. Last seen 20 hours 57 min ago. Offline
Joined: 09/29/2010
Groups: Window Group

Dragging the rectangle off will remove the constraints - however, it wont change the conformation at all, so you'll still be left with the shape.

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, RosettaCommons