After CASP is over, what's rosetta going to do? Is it going to be doing its on protein stuff... not the already determined structures from CASP?

After CASP is over, what's rosetta going to do? Is it going to be doing its on protein stuff... not the already determined structures from CASP?

We have a long list we are excited to get to when CASP is over. first on the list are improvements to the basic method suggested by our observations on the CASP targets. these will be tested on the CASP proteins whose structure has been solved and the proteins in our previous test set. it was really useful to look at rosetta models for the larger casp proteins--there were some clear aspects to improve.
Second, we want to get the HIV vaccine design calculations running ASAP; with the Gates foundation funding we have funds to support testing of quite a number of designs as possible vaccines and we are eager to get going with this! Third, there are other modeling projects in the group, such as predicting the structures of the amyloid fibers implicated in many human diseases we would like to start running as well.

---

Prior to CASP, Rosetta was working on improvements to the modelling ability of the Rosetta client, and a number of things that dramatically reduced the bandwidth used by the client.

After CASP is over, we'll be helping out with improving the client again.

We've been creating 100,000 decoys/models for some/all of the CASP targets; and one of the ideas that we didn't get to test out was creating 10,000 decoys/models, and then with the data from the first 10k decoys/models, create a new target that includes data from the first 10k models so that the second run of 10k models will be relatively fast, and produce decoys/models as accurate as we're getting now with 100,000 decoys/models. (i.e. produce 20k models instead of 100k..) They'll try new approaches - some which will be improvements, some that won't. They'll see certain types of proteins or types of protein structures the Rosetta client doesn't work as well with - and test out new ideas on how to create accurate predictions for those types of proteins/protein structures.

The more this client is improved, the better it'll be at helping find a cure or treatment for the ailments it'll be targetted with in the future.

Casp is a competition to see how well the various scientists involved in protein structure prediction are at predicting proteins of an unknown shape. (The structure of the proteins are known, but unpublished until after the Casp competition.) When Casp is over, we'll go back to working on proteins with published structures.

---

Second, we want to get the HIV vaccine design calculations running ASAP; with the Gates foundation funding we have funds to support testing of quite a number of designs as possible vaccines and we are eager to get going with this!

Just a possibly dumb question, but figure I'm due. How will this HIV work differ from WCG faah? Are they very similiar in goal & method?

---

Prior to CASP, Rosetta was working on improvements to the modelling ability of the Rosetta client, and a number of things that dramatically reduced the bandwidth used by the client.

After CASP is over, we'll be helping out with improving the client again.

We've been creating 100,000 decoys/models for some/all of the CASP targets; and one of the ideas that we didn't get to test out was creating 10,000 decoys/models, and then with the data from the first 10k decoys/models, create a new target that includes data from the first 10k models so that the second run of 10k models will be relatively fast, and produce decoys/models as accurate as we're getting now with 100,000 decoys/models. (i.e. produce 20k models instead of 100k..) They'll try new approaches - some which will be improvements, some that won't. They'll see certain types of proteins or types of protein structures the Rosetta client doesn't work as well with - and test out new ideas on how to create accurate predictions for those types of proteins/protein structures.

The more this client is improved, the better it'll be at helping find a cure or treatment for the ailments it'll be targetted with in the future.

Casp is a competition to see how well the various scientists involved in protein structure prediction are at predicting proteins of an unknown shape. (The structure of the proteins are known, but unpublished until after the Casp competition.) When Casp is over, we'll go back to working on proteins with published structures.

well said! CASP definitely exposed areas where Rosetta needs improvement, particularly for larger proteins, and we are eager to address these issues.

---

Second, we want to get the HIV vaccine design calculations running ASAP; with the Gates foundation funding we have funds to support testing of quite a number of designs as possible vaccines and we are eager to get going with this!

Just a possibly dumb question, but figure I'm due. How will this HIV work differ from WCG faah? Are they very similiar in goal & method?

You can read about our approach in the press releases on the Gates grant, and in the "medical applications" section on the home page.

---

Hello !

While we're speaking over "what happens after CASP7" : when is CASP7 due to stop ?
Because I'm now actually only crunching for rosetta@home (part of the CASP7 effort !) and I'd like to know when I shall be able to crunch for other projects again ?

Thank you.
Duanra

when is CASP7 due to stop ?

I've heard that the final deadlines are on Monday, August 7.

Here is the target list with the expiration dates up to Aug. 7. but they have one more list here having Aug. 11. as the latest target deadline. I don't know if that means anything for the Rosetta submissions.

I suppose that if the deadline of the WUs won't be shortened, today the last CASP7 WUs will be delivered. Am I right that all WUs with a deadline beyond 08/07/06 won't count for CASP7 anymore? In any case I have to start resuming some other projects tomorrow morning to be sure that the suspended WUs are finished before their deadline. So my question: Is CASP7 finished when the last WUs with <report deadline: 08/07/06> are delivered or is CASP7 finished at 08/07/06 regardless od the deadline of the WUs?

---

Love, Michi

Any CASP WU you submit after the 7th can't be used in the CASP challenge anymore.

They'll still send the CASP WUs at least until the 6th and maybe a few days after that while they prepare new ones. Even results found after the deadline can be used to refine Rosetta. Keep in mind that CASP is merely to test how the different methods work and doesn't actually find anything new. I'd imagine doing well is important for receiving funding (which would make it very important) and whether we crunch an already released protein or a CASP target doesn't make a difference in testing/refining Rosetta.

So basically, the massive cpu power is still needed afterwards, just that the WU name doesn't display CASP anymore - the work is still the same though and any publication they can make equals a bigger chance for more money ;)

Although after the 7th I'll return to crunching SIMAP partially as well - seems like it'd be an efficient way to figure out the purpose of so many proteins in relatively short time. That being said, protein folding is probably still the holy grail in the field. And of course QMC@home tries to refine the very algorithm Rosetta uses... I rest my case that BOINC just has too many projects worth contributing processing power to and that the world really needs more supercomputers ;)

Maybe we need to find a way to raise $9m and buy a supercomputer like the new japanese one someone posted about... 1TB of processing power, what an amazing RAC that'd deliver ;)

---

...Am I right that all WUs with a deadline beyond 08/07/06 won't count for CASP7 anymore?

In the past, they've released WUs with the normal 1 week deadline, in HOPES of getting results back much sooner. They did this to squeeze in a few more models prior to submitting their prediction to CASP. Most participants' results are returned in 2 or 3 days. They didn't want to have inconsistent deadlines. So, on average, this was a way to bring more models crunched to specific challenging proteins. Any results, even after the CASP deadlines are still valuable, just not directly to the prediction submitted to CASP.

...so I wouldn't be surprised to see this happen again. And if you can report the results back earlier then the deadline, that will be the most helpful for the CASP work.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

I think that the CASP 7 WU's completed before the competition deadline will count even if the WU deadline is past that date.

I may be wrong, but I may be right.

---

I think that the CASP 7 WU's completed before the competition deadline will count even if the WU deadline is past that date.

...I KNEW there was an easier way to say it! Thanks.

Crunch and report back all that you can thru Aug. 7th.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

I think that's my psychological profile:

General insanity coupled with spirts of brilliance.

Threw in some Billy Joel for good measure.

---

Maybe we need to find a way to raise $9m and buy a supercomputer like the new japanese one someone posted about... 1TB of processing power, what an amazing RAC that'd deliver ;)

Bruce Allen over at Einstein@Home has some 640 Opterons at his disposal...

His RAC is currently 670,745.10!

---