Shake Accuracy does not affect Shake

Case number:845813-991092
Topic:Game: Tools
Opened by:Susume
Status:Closed
Type:Bug
Opened on:Monday, November 14, 2011 - 00:23
Last modified:Tuesday, July 24, 2012 - 20:16

The shake accuracy slider is not affecting the output of the shake tool. I tested shake accuracy at 0 and 4 against the same protein state, multiple times, across multiple proteins. Shake produced the same results regardless of the accuracy setting.

In each case puzzle was reset, then every movable sidechain was moved to the first position in its list *after* the position it held at reset. This was done to leave significant room for improvement in sidechain positions. Shake was run 1, 2, 3, 4 and 5 times (total of 15 iterations per test) and the resulting scores recorded. This was done 3 times at shake accuracy 0 and 3 times at shake accuracy 4. Repeated test against three different proteins. There is sometimes a small amount of variability (a few thousandths of a point) in the shake routine at any one setting; however, taken to four (in some cases ten) decimal places, there was no significant difference between scores at accuracy 0 and at accuracy 4. Clash importance was at 1.0 and Wiggle accuracy at 3.0 for all tests.

(Mon, 11/14/2011 - 00:23  |  4 comments)


Joined: 05/09/2008
Groups: None

thanks for reporting this bug...

spmm's picture
User offline. Last seen 9 hours 42 min ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

any progress?

Joined: 05/09/2008
Groups: None
Status: Open » Open

I believe this bug has been taken care of and will be released in the next slider update (which includes even more sliders!)

Joined: 06/17/2010
Status: Open » Closed

Sliders removed. Closing for now. They may sneak back in .... someday....

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