[NewChapter] Hand building notes, a bug and questions

Case number:699969-996661
Topic:General
Opened by:auntdeen
Status:Open
Type:Question
Opened on:Thursday, January 16, 2014 - 00:39
Last modified:Sunday, January 26, 2014 - 22:06

I've just opened the Spider Toxin both in Main and in newchapter. I ran a blue fuse v1.1 in both.

In main, the locked segment is central, but in newchapter it is on the terminal end - it looks as if 3-4 segments are locked, where we are used to seeing one locked.

I did a hand rebuild of 2 segments on 47 - 48 on both. The Main client finds more rebuild poses than newchapter, and more varying ones (this could be purposeful).

***However, this is a bug - restoring very best didn't restore very best in newchapter. The pose with the highest score did not restore in the 2 segment rebuilds.

On a rebuild of 3 segments, 46-48, newchapter did restore the highest for very best. Main did twice as many rebuild poses in the same timeframe as newchapter. There also was much less variance in the newchapter rebuild graph than the Main client - has the library been truncated for this new client?

I then did a rebuild on segments 39-44. Main found 23 poses in the time it took for newchapter to find 2.

I'm finding it a little harder to visually distinguish which areas to work on in newchapter. The Blue Fuse v1.1 caused the protein in newchapter to gain many red areas, while the protein in Main is what we are used to, with varied reds and greens. We are used to visual clues based on color of segments - this could make it a little more difficult to distinguish areas for handwork.

Running a rav3n_pl Push found points much more quickly in the Main client.

My impression is that newchapter presentations are stiffer than Main. Which makes me ask - are we now simply refining?

(Thu, 01/16/2014 - 00:39  |  1 comment)


Joined: 11/10/2007
Groups: Window Group

Recent updates with fixes should address these issues with rebuild and the terminal being frozen.

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