Posts by Mike Tyka

1) Message boards : Cafe Rosetta : Team Thread (Ads only, not for discussions) (Message 68487)
Posted 9 Nov 2010 by Mike Tyka
Post:
Hi teams!

Rosetta@HOME now has a twitter account for news and updates!

www.twitter.com/rosettaathome

Come and follow us - we'll be posting news & science updates there!

Mike
2) Message boards : Rosetta@home Science : CASP9 (Message 66076)
Posted 11 May 2010 by Mike Tyka
Post:
Indeed. For jobs that contain _casp9 in their name, we need the results back within about 18hrs. Everything that comes back after that will ikely not make it into the final analysis - the reason is that we need some time for post analysis here before we can send the final models (we have to choose 5!) to the CASP organisers. So, for those kind of WUs it would help us tremendously if you could set your runtimes to 12-16hrs, even if you only end up crunching a single decoy (sometimes we get fairly large structures from CASP).
These deadline restrictions are only for the server predictions - the human team have about 3 weeks per target, but they use a different naming scheme.

Mike

3) Message boards : Rosetta@home Science : CASP9 (Message 66053)
Posted 10 May 2010 by Mike Tyka
Post:
Hi all!

First of all thanks for all your computing effords!

In answer to some of the questions:

The jobs ending in "casp9" and containing t000_ are homology modelling jobs ones submitted by our automated new rosetta server. Sadly the way it's set up they do not contain the original casp target name but instead are numbered internally so you can't tell from the name what target they are. But you can probably guess from the length and the day they were submitted (the submit day should match up with the target release date).
These targets are operating under the Server deadline, not the human deadline, which means only 3 days from release time to deadline !

OTOH jobs that do not contain casp9 in the name are not necessarily not related to casp. THe human group will be submitting various jobs relating to casp but will use custom names.

Basically a very large fraction of the jobs you'll see on your machines will be relaated to casp in one way or another - i'd say at least 60-70%.


Mike

4) Message boards : Number crunching : minirosetta 2.11 (Message 66024)
Posted 7 May 2010 by Mike Tyka
Post:
Wrt to RMS "flatlines":

We are now running real live true blind targets, which means we do not know the answer (unlike during testing where we do) and so we cannot compute the RMSD since there is nothing to compare to (RMSD is the root mean square deviation from the experimentally determined fold found in nature).
Hence the RMSD flatlines!

Mike
5) Message boards : Number crunching : Rosetta Application Version Release Log (Message 65841)
Posted 26 Apr 2010 by Mike Tyka
Post:
Minirosetta 2.11 is out !
6) Message boards : Number crunching : minirosetta 2.11 (Message 65840)
Posted 26 Apr 2010 by Mike Tyka
Post:
More bug fixes and CASP related preparrations !

Please continue to let us know about problems, its very useful to have your feedback !

Cheers, Mike
7) Message boards : Rosetta@home Science : CASP9 (Message 65811)
Posted 25 Apr 2010 by Mike Tyka
Post:
Hey everyone !

As some of you may know CASP9, a community wide experiment in structure prediction, is starting in less then 10 days. As usual we are participating with both a human prediction group and a brand new automatic server, called RosettaServer. We will be utilizing BOINC heavily during the next 4 months, as you can see we already have nearly 4 million workunits lined up.
We have been very busy getting everything ready in time!

CASP 9
CASP is an international experiment to assess the state-of-the-art of the protein structure prediction field. Sequences, whose structures have been solved but which have not yet been published are sent out to participating teams and we have a limited amount of time to send back predictions. The whole thing is conducted in a double-blind fashion ensuring fair assessment and truly blind prediction.

We will be testing a number of new strategies during this period, these will be carried out by the human prediction group and crunched on BOINC.

In the automatic category we have modernized out old prediction server (Robetta, created more then 6 years ago) and replaced it's heart with Rosetta's most modern implementation. Among many scientific improvements, we have also, for the first time, connected the public structure prediction server to BOINC for both ab initio prediction and homology modeling.

For the community, by the community:
After CASP this server will go public and provide Rosetta's Abinitio and Homology modelling capabilities to the scientific community (and to the public). The computation for this will be conducted on BOINC meaning that you guys will be crunching protein structure prediction jobs for real scientific studies conducted by researchers all over the world!


Thanks again everyone for crunching, we wouldn't be able to do this stuff without you !

Excitedly yours,

Mike
8) Message boards : Number crunching : Rosetta Application Version Release Log (Message 65734)
Posted 12 Apr 2010 by Mike Tyka
Post:
Minirosetta 2.10 is out !
9) Message boards : Number crunching : Minirosetta 2.10 (Message 65733)
Posted 12 Apr 2010 by Mike Tyka
Post:
This is the first update for some time, we've added a lot of functionality, maybe Sarel can comment in more detail ?

Please post bug reports here !

10) Message boards : Number crunching : minirosetta 2.05 (Message 65113)
Posted 26 Jan 2010 by Mike Tyka
Post:
If it's the high energy of 4K you're worried about - that's not unusual when runs are submitted with constraints - looks all good to me ..

Mike
11) Questions and Answers : Windows : Frequent hung work units (Message 64955)
Posted 13 Jan 2010 by Mike Tyka
Post:
Hmm. We've not been able to reproduce this problem here unfortunately.
There is an update going out today (2.05) (it may have already gone out in fact) that fixed a different issue to do with checkpointing. Is that version still giving you these troubles ?

Mike
12) Questions and Answers : Windows : Frequent hung work units (Message 64933)
Posted 12 Jan 2010 by Mike Tyka
Post:
THose are my jobs. Seems like the display counter is not being updated during this protocol. THis is a cosmetic problem though - the jobs are running just fine underneath and we're getting lots of good data back! I'll put a bug fix in in the next version.


Cheers, Mike
13) Message boards : Number crunching : minirosetta 2.03 (Message 64530)
Posted 16 Dec 2009 by Mike Tyka
Post:
Is the stack overflow issue the reason why mix_score13_env_rlbd_1hz6__IGNORE_THE_RESTlr8_DECOY_16523_77_09 made 129 decoys from 129 attempts?


No, but 129 decoys is a good thing - isn't it ?
14) Message boards : Number crunching : minirosetta 2.03 (Message 64522)
Posted 15 Dec 2009 by Mike Tyka
Post:
Ok, it appears we had to many old application backlogged. It was indeed a server problem - it should be resolved now :) - sorry for the hick up.
15) Message boards : Number crunching : minirosetta 2.03 (Message 64516)
Posted 15 Dec 2009 by Mike Tyka
Post:
Hi! 9.07 here now :)

Not sure what the problem is - we're looking into it now. THis version
worked fine on RALPH, so it we suspect something went awry during the actual update.

16) Message boards : Number crunching : Rosetta Application Version Release Log (Message 64479)
Posted 15 Dec 2009 by Mike Tyka
Post:
Minirosetta 2.03 is out!
17) Message boards : Number crunching : minirosetta 2.03 (Message 64478)
Posted 15 Dec 2009 by Mike Tyka
Post:
THis version fixes a stackoverflow error that we didn't catch in 2.02.

Please post issues here, thanks !

18) Message boards : Rosetta@home Science : mega score 12 rldb work units (Message 64008)
Posted 11 Nov 2009 by Mike Tyka
Post:
By looking at my results page, I see tasks with names of the form:
mega_lrx_seq_score12_rlbd_xxxx_

...and I note that those tend to have 500,000 or more models run, and they also tend to have RSMDs over 1.5. In other words, we ran 5 or 10 times more models and still did not come up with an answer as close to the true structure as we would like to be, or as close as we get for other proteins.

I was curious what is behind these work units? What is being studied with the larger sample size?


Here we're intentionally sampling structures ranging from 0 to 20 RMS to learn what the energy landscape looks like. In order to get a better resolution we have to run massive calculation sizes.
19) Message boards : Number crunching : Minirosetta 1.97 (Message 63009)
Posted 22 Aug 2009 by Mike Tyka
Post:
1.95/1.96 Repackaged again as 1.97 including minirosetta.linuxgccrelese
20) Message boards : Number crunching : Rosetta Application Version Release Log (Message 62997)
Posted 21 Aug 2009 by Mike Tyka
Post:
1.95 is up :)


Next 20



©2020 University of Washington
https://www.bakerlab.org