Back to Recipes Homepage
recipe picture
Recipe: DRW for FIlters TvdL 2.1.1
Created by brow42 79 379
4.363635
Your rating: None Average: 4.4 (11 votes)

Profile

Name: DRW for FIlters TvdL 2.1.1
ID: 45819
Created on: Sat, 03/23/2013 - 14:57
Updated on: Sat, 03/23/2013 - 21:57
Description:

Modified to turn filters off during wiggle and on during scoring.



Best For


Comments

brow42's picture
User offline. Last seen 6 days 16 hours ago. Offline
Joined: 09/19/2011
Groups: None
Speeds up DRW but may be less efficient

The basic idea here is that (maybe) filters are calculated after a micro-wiggle occurs; wiggle doesn't actually use the filter contribution to decide what to do. So the slow filters don't need to be on during wiggle. This version of DRW has been modified to turn slow filters off before wiggle, and turn them on after (and also before any scoring).

Wiggling with filters off messes up the recent best scoring. Sometimes wiggle will make you gain backbone points but lose more points from the filter...for example, by letting your protein spread out enough to lose the core. Restore recent best would normally restore to right before you lost the filter (not necessarily the best thing to do). With filters off, restore recent best restores the state at the end of the wiggle, with the filter points lost.

I cannot fix this. However, I turn the filters back on and score that conformation. If it is a low scoring conformation, DRW will reject it in the usual way.

Joined: 09/24/2012
Groups: Go Science
interesting idea

I will try your recipe asap

Joined: 02/08/2012
Groups: None
working on selected

if you select what you want worked on, the script instantly re-selects the entire puzzle. working on selected doesn't work.

Want to try?
Add to Cookbook!
To download recipes to your cookbook, you need to have the game client running.
Parent
Children

none

Authors
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