CASP11

Message boards : Rosetta@home Science : CASP11

To post messages, you must log in.

AuthorMessage
Profile David E K
Volunteer moderator
Project administrator
Project developer
Project scientist

Send message
Joined: 1 Jul 05
Posts: 1018
Credit: 4,334,829
RAC: 0
Message 77808 - Posted: 6 Jan 2015, 19:26:37 UTC

Here is a synopsis of how we did in CASP11 with the help of your contributions, which were absolutely critical for our success. There's also a bunch of great info posted by Sergey in this thread. The meeting was held in early December and we were invited to give 4 talks. The slides for these talks will soon be available from the CASP11 website. In CASP11 we focused on our automated methods, the contact assisted and refinement categories, and interesting targets for which we thought manual intervention would significantly improve model quality.

We did particularly well when considering the best model out of the 5 submissions for each target.

For the fully automated category, our Robetta server (BAKER-ROSETTASERVER) which uses Rosetta@home was ranked 1st for both the TBM (template based modeling) and FM (free modeling) categories with significant gaps compared to the second ranked servers.

Assessors ranking of best models for server only TBM

Assessors ranking of best models for server only FM

For the "all groups" category, our human group ranked 1st in the FM category (this category consists of very difficult targets due to the absence of structural homologs). One of the main highlights of CASP11 was our T0806 prediction from this category, an exceptional example of significant progress made since last CASP. Also worthy to note since this category includes human predictors is that our server ranked 3rd.

Assessors ranking of best models for 'all groups' FM targets

As Sergey describes in this thread, we also did well in the contact assisted categories, ranking 1st in Ts (w/ sparse simulated NMR data) and Tx (w/ experimental cross-link data) categories for the first and best models.

Assessors ranking of first model for Ts

Assessors ranking of first model for Tx

Hahnbeom will likely post additional information here describing his exceptional results for the refinement category. For this category we ranked 1st when considering the best models with a significant gap compared to the next ranked group.

Assessors ranking of best models for refinement

At the meeting it was made abundantly clear through the excitement displayed by the assessors that there was significant progress made in the field compared to last CASP, and this is mainly attributed to the use of evolutionary sequence information to accurately predict contacts used to guide modeling. Sergey describes this here and his spectacular T0806 prediction in this cool video.

We of course could not have done any of this without R@h and all your contributions. Congratulations and thank you all for participating in R@h and CASP!

ID: 77808 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Polian
Avatar

Send message
Joined: 21 Sep 05
Posts: 152
Credit: 10,141,266
RAC: 0
Message 77809 - Posted: 7 Jan 2015, 3:13:41 UTC

David and Sergey,

Thanks for the updates and the links!

While I may not understand all the statistics here, I do see a large list of competitors and I know what #1 means :)

I'm glad to see that processing all these workunits has led to advances in the field.
ID: 77809 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile proxima

Send message
Joined: 9 Dec 05
Posts: 44
Credit: 4,148,186
RAC: 0
Message 77815 - Posted: 9 Jan 2015, 14:46:16 UTC

This is excellent news - congratulations to everyone (teams members, and volunteers alike).

Nice video, too - good luck with getting to CASP12, Sergey :-)
Alver Valley Software Ltd - Contributing ALL our spare computing power to BOINC, 24x365.
ID: 77815 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Rosetta@home Science : CASP11



©2024 University of Washington
https://www.bakerlab.org