Behavior: Clashing Importance - Suggested modification

Case number:845813-990897
Topic:Game: Tools
Opened by:phi16
Status:Open
Type:Suggestion
Opened on:Thursday, October 20, 2011 - 15:58
Last modified:Monday, July 23, 2012 - 11:47

Since the introduction of the Behavior: Clashing Importance tool players have found good use and great success gain control of interactive forces by varying the strength and invoking greater/lesser strength at various times during play. It is now one of the most commonly used features of recipes. In fact, some players have found using low Clashing Importance for most of the design/development phase of puzzles works quite well. I would like to suggest modifying the tool to give further control by separating the influence of various amino acids.

If you put in order the twenty amino acids from most hydrophilic to most hydrophobic, nearest to farthest, or, from heaviest to lightest, the Clashing Importance for each might be controlled on a console that looks like a graphic equalizer in a sound app. Twenty clashing importance slides from 1.00 --> 0.00 for each amino acid type. There should also be a Mode selection: Weight, Hydrophobic/philic, and Distance choices.

In a sense, we would be able to say "Let's see what happens when only the heavier hydrophobes influence the protein." while we set all to 0.00 and begin to slide the tryptophan or tyrosine to greater values. These sorts of experiments would lead to a greater understanding of the forces at play.

Types of experiments: Using the influence of only heavier amino acids, proximity influence, hydrophobics, steric hinderance.

phi16

(Thu, 10/20/2011 - 15:58  |  8 comments)


Joined: 06/17/2010

...or change it by selection and scripts :) Nice idea (+) :)

phi16's picture
User is online Online
Joined: 12/23/2008

Hey Art,
Thanks for the comment. Rav has written several scripts which make selections, perform actions and change CLashing Importance already. What is new here is suggesting that Clashing Importance be broken down into component parts by animo acid type, distance or hydropathy index.

Sidebar [You're new to Fold.it so I suggest you join a team, make friends with scriptors or download a half dozen recommended scripts and try them out. Scripts are not just for scriptors. Even non-coders are learning how to make simple changes to existing scripts without too much difficulty. ]

phi16

spmm's picture
User offline. Last seen 45 weeks 2 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

IMHO this is the sort of thing which would be good to try once we have the ability to change script options in an interface not have to dig through code to understand, find and change them, save the script twice to stop it from crashing and so on. This would make it simpler for non coders but coders can still do mods manually in the code.

Joined: 06/17/2010

A bump.
CI settings have been changed some time ago (during sliders update?) and current setting is not optimal IMO.
When I have score 10k and do a wiggle on CI=0.5 I expect to my score drop by half to 5k.
When I wiggle on CI =0.05 I expect that my score drop to 5% of starting score (or even in negative zone).
For now, it works that shake/wiggle on 0.2 making about 80% of 1.0 score. Also setting like 0.01 in most cases was very close to collapse (score -9k or more), now it can make even 10% score.

infjamc's picture
User offline. Last seen 2 years 49 weeks ago. Offline
Joined: 02/20/2009
Groups: Contenders

A slider for cut points is also overdue... currently, the protein would explode even at CI=0.01 if there's a huge gap across a cutpoint.

tokens's picture
User offline. Last seen 28 weeks 1 day ago. Offline
Joined: 11/28/2011

I think the developers are working on it. jflat has said that there are technical issues with creating a slider that only affects cutpoints and doesn't affect bands and the pull tool.

tokens's picture
User offline. Last seen 28 weeks 1 day ago. Offline
Joined: 11/28/2011

I think it would be even more interesting to see sliders for the other aspects comprising the total score. Especially a slider for turning down the importance of the bonding score coming from sidechains. In the beginning of a puzzle the bonding of sidechains might prematurely fix the backbone in a particular shape. I would like to in the beginning to ignore sidechains bonds while I'm reshaping the backbone, and wait with creating the sidechain bonds until later.

infjamc's picture
User offline. Last seen 2 years 49 weeks ago. Offline
Joined: 02/20/2009
Groups: Contenders

If I remember correctly, at least one developer has mentioned the idea of a centroid representation, where the side chain atoms are simplified to a sphere placed at its center of mass. That would be another possible approach.

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