Big changes on Rosetta and Foldit

Message boards : Rosetta@home Science : Big changes on Rosetta and Foldit

To post messages, you must log in.

AuthorMessage
Profile [VENETO] boboviz

Send message
Joined: 1 Dec 05
Posts: 892
Credit: 3,347,076
RAC: 3,113
Message 88455 - Posted: 11 Mar 2018, 10:11:49 UTC

There is big changes on Rosetta code and these involve also Foldit

But perhaps the biggest change that comes with the update is improvements to scoring. There's been a *lot* of work put into the Rosetta scoring function recently, and just about every portion of scoring has been re-evaluated and re-optimized.

This improvement isn't limited only to protein structure prediction. The new scoring function shows discrimination improvements in a wide range of protein prediction problems, including protein-protein and protein-small-molecule interaction predictions.


From pubblication:
The resulting next-generation Rosetta energy function (opt-nov15) outperforms the previous energy function (talaris2014)13 on a wide range of structure prediction tests independent of the training set data. In contrast to opt-nov15, talaris2014 had been optimized solely relying on similar set of structural data we incorporate in the study without the use of small molecule data


I hope this changes is integrated in the new 4.x app (or will be implemented soon) and we give better simulations
ID: 88455 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 976
Credit: 20,844,333
RAC: 14,020
Message 88465 - Posted: 13 Mar 2018, 0:43:00 UTC - in response to Message 88455.  

What does "scoring" mean in this context?

It can't mean task credits for users as that would be largely inconsequential.
ID: 88465 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile [VENETO] boboviz

Send message
Joined: 1 Dec 05
Posts: 892
Credit: 3,347,076
RAC: 3,113
Message 88467 - Posted: 13 Mar 2018, 7:30:23 UTC - in response to Message 88465.  
Last modified: 13 Mar 2018, 7:47:10 UTC

What does "scoring" mean in this context?
It can't mean task credits for users as that would be largely inconsequential.


Uh, you are not a newbie here.
I think you know what kind of simulations R@H does.
They use the "minum energy" for their simulations. (and, at the end, they compared these simulations with in-vitro proteins).
So the "scoring" is reported to this energy
ID: 88467 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
rjs5

Send message
Joined: 22 Nov 10
Posts: 250
Credit: 8,037,564
RAC: 0
Message 88498 - Posted: 20 Mar 2018, 6:51:42 UTC - in response to Message 88455.  

There is big changes on Rosetta code and these involve also Foldit

But perhaps the biggest change that comes with the update is improvements to scoring. There's been a *lot* of work put into the Rosetta scoring function recently, and just about every portion of scoring has been re-evaluated and re-optimized.

This improvement isn't limited only to protein structure prediction. The new scoring function shows discrimination improvements in a wide range of protein prediction problems, including protein-protein and protein-small-molecule interaction predictions.


From pubblication:
The resulting next-generation Rosetta energy function (opt-nov15) outperforms the previous energy function (talaris2014)13 on a wide range of structure prediction tests independent of the training set data. In contrast to opt-nov15, talaris2014 had been optimized solely relying on similar set of structural data we incorporate in the study without the use of small molecule data


I hope this changes is integrated in the new 4.x app (or will be implemented soon) and we give better simulations


I only have access to the 3.78 source code and not the 4.0 source using my Academic/non-profit license.

Disassembly shows the 3.78 is now a 32-bit only app. It only uses x87 floating point operations. I would guess that it is using double precision since it only uses the 80-bit x87 registers.
4.07 seems to be quite different and is a 64-bit app that relies heavily, if not exclusively on 4 parallel single precision numbers stored in the xmm registers.

IMO, it appears that version 4 has taken the first step toward some level of parallelism. The developers and probably staff have some idea of how much of a difference and how much problem the changes have introduced. There is no real easy way as a cruncher to tell how much difference it makes. The code crunches for so many hours and makes as many iterations as it can during that time.
ID: 88498 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 976
Credit: 20,844,333
RAC: 14,020
Message 88502 - Posted: 21 Mar 2018, 1:27:43 UTC - in response to Message 88467.  

What does "scoring" mean in this context?
It can't mean task credits for users as that would be largely inconsequential.

Uh, you are not a newbie here.
I think you know what kind of simulations R@H does.
They use the "minum energy" for their simulations. (and, at the end, they compared these simulations with in-vitro proteins).
So the "scoring" is reported to this energy

You would be surprised how little I know... lol
ID: 88502 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile [VENETO] boboviz

Send message
Joined: 1 Dec 05
Posts: 892
Credit: 3,347,076
RAC: 3,113
Message 88726 - Posted: 20 Apr 2018, 15:09:03 UTC - in response to Message 88498.  

I only have access to the 3.78 source code and not the 4.0 source using my Academic/non-profit license.


The RosettaCommons team released the 3.9 version of Rosetta platform.
I don't know if 4.0 R@H source code is this version or the older (3.8 Rosetta)
ID: 88726 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Rosetta@home Science : Big changes on Rosetta and Foldit



©2019 University of Washington
http://www.bakerlab.org