fix rebuild of helixes; enforce design puzzle goals via filters

Case number:699969-997148
Topic:General
Opened by:LociOiling
Status:Closed
Type:Suggestion
Opened on:Friday, February 28, 2014 - 20:26
Last modified:Tuesday, July 31, 2018 - 20:37

Based on other feedbacks and chats with players, it seems like the rebuild tool is deliberately biased against creating helixes at this point.

The possible reason for this bias is the results for design puzzles. Designs consisting of nothing but helixes tend to score highly in design puzzles. The scientific value of a new batch of large helixes is questionable, however.

The downside of this approach is that it's extremely difficult to form helixes in non-design puzzles such as 849b. For this player, the large helix predicted for residues 53-65 doesn't want to form using rebuild. Many rebuilds return useless knots that look like they were formed at CI=0 even when CI=1. Manually banding the segments produces similar results. A long series of manual steps produces a poorly scoring approximation of a helix.

If indeed there is some kind of grudge against helixes, it's making the game less enjoyable and less productive for many players.

The suggestion is to eliminate the bias, and restore the rebuild function to a more neutral stance vis-a-vis helixes.

In design puzzles, it should be a Simple Matter of Programming (SMOP) to create a "Novelty" filter that would lower the score for Honkin' Big Helixes (HBH). The filter might consider factors like the percentage of residues in a helix and the total number of helixes, the length of loop sections between helixes, etc.

This change would help ease some of the lingering pain results from the New Chapter changes.

(Fri, 02/28/2014 - 20:26  |  4 comments)


brow42's picture
User offline. Last seen 4 weeks 2 days ago. Offline
Joined: 09/19/2011
Groups: None

Actually, David Baker has gone on record saying he loves helix bundles. I think rebuild is just broken.

LociOiling's picture
User offline. Last seen 3 hours 58 sec ago. Offline
Joined: 12/27/2012
Groups: Beta Folders

Yes, well puzzle 851 dropped after this suggestion, and it's rather helixly.

Agree that something in rebuild is broken, the question is whether it's broken by design.

LociOiling's picture
User offline. Last seen 3 hours 58 sec ago. Offline
Joined: 12/27/2012
Groups: Beta Folders

See also http://fold.it/portal/node/997092 for a related discussion.

LociOiling's picture
User offline. Last seen 3 hours 58 sec ago. Offline
Joined: 12/27/2012
Groups: Beta Folders
Status: Open » Closed

Old, closing.

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