Dr. Baker has mentioned briefly in the past that in the back of his mind, an enzyme will be the solution to the increasing CO2 levels as well. Once you successfully develop such a catalyst, you probably incorporate it into a living cell such as a bateria where nature will automatically replicate it as long as a sufficient food supply remains available. The ultimate outcome might be that you sprinkle some bacteria here and there in the ocean and they absorb carbon, release (or "desorb") oxygen and then die and fall to the bottom much like organisms in nature already do.

It's probably too early to find much reading material on the subject, but please post links to any additional material on current research findings and projects that are being investigated in this groundbreaking field.

Here's an interesting commentary on the issues surrounding ocean fertilization. Its oriented more towards current schemes for increasing phytoplankton production by adding iron (a limiting nutrient in many cases) but much of it seems applicable to any similar kind of ocean fertilization. Bottom line: there are a lot of problems with this kind of scheme.

http://www.realclimate.org/index.php/archives/2007/05/thin-soup-and-a-thin-story/

---

i think ideally the carbon from CO2 should be stored as something useful like diamond (useful for industry) or carbon fibres for use in all sorts (construction could be very different with cheap carbon-based frames), and, with an environmentally friendly energy input it could be used to create fuel... Then you don't have the problems of acidification of the sea...

how hard can it be? :D

---

i think ideally the carbon from CO2 should be stored as something useful like diamond (useful for industry) or carbon fibres for use in all sorts (construction could be very different with cheap carbon-based frames), and, with an environmentally friendly energy input it could be used to create fuel... Then you don't have the problems of acidification of the sea...

how hard can it be? :D

Useful though these products undoubtedly are, additional usage isn't going to make much of a dent in CO2 levels; not when we're currently releasing about 24 Gigatonnes per year of the stuff. It's necessary to sequester carbon dioxide on a much bigger scale.

---

i think ideally the carbon from CO2 should be stored as something useful like diamond (useful for industry) or carbon fibres for use in all sorts (construction could be very different with cheap carbon-based frames), and, with an environmentally friendly energy input it could be used to create fuel... Then you don't have the problems of acidification of the sea...

how hard can it be? :D

Useful though these products undoubtedly are, additional usage isn't going to make much of a dent in CO2 levels; not when we're currently releasing about 24 Gigatonnes per year of the stuff. It's necessary to sequester carbon dioxide on a much bigger scale.

sequestering CO2 on any scale can still mean it's products are useful ... ;)

---

Thank you for the Nature News link, Dr. Baker. Good photo :)

I was wondering if you had any news on potential CPU optimizations to Rosetta (using SSE/SSE2 instructions, for example). I know this has been discussed in the past, but no updates in a while.

I ask because I figure that while it's probably a painful feature to add, once it's there, everything from that point on is faster -- it would probably be like adding tens of thousands of new CPUs to the project, depending on the % of speeding up -- and that would lead to better predictions and the ability to work on more proteins.

If Rosetta programmers are too busy, maybe a call for an external programmer is possible. I'm sure someone out there would love to add this to his/her portfolio, maybe even for free (the open source community is not short on talented programmers).

Keep up the good work!

---

The article here: http://www.sdsc.edu/News%20Items/PR110106_casp.html (it's not the one in nature (here), but is linked to by one of the comments at the bottom of that article)

says:

“The post-production data analysis is as important as generating the models themselves,” said Vatsan Raman of Baker's team. “Similarly, the experience gained on the SDSC machines was invaluable when it came to running the code on over 40,000 processors. Being able to run on the full machine at IBM was a real breakthrough for the Rosetta team, allowing us to run more complex calculations than we had attempted before, in a fraction of the time.”

Does that mean it would be beneficial to run multi-threaded tasks in parallel on say on a quad-core machine at some point in the near future? Or would the memory requirements be too great for typical home machines? Or would this result in reduced productivity due to the required interaction of the cores vs current one-model-per-core setup?

---

The article here: http://www.sdsc.edu/News%20Items/PR110106_casp.html (it's not the one in nature (here), but is linked to by one of the comments at the bottom of that article)

says:

“The post-production data analysis is as important as generating the models themselves,” said Vatsan Raman of Baker's team. “Similarly, the experience gained on the SDSC machines was invaluable when it came to running the code on over 40,000 processors. Being able to run on the full machine at IBM was a real breakthrough for the Rosetta team, allowing us to run more complex calculations than we had attempted before, in a fraction of the time.”

Does that mean it would be beneficial to run multi-threaded tasks in parallel on say on a quad-core machine at some point in the near future? Or would the memory requirements be too great for typical home machines? Or would this result in reduced productivity due to the required interaction of the cores vs current one-model-per-core setup?

my guess is that it would be possible to run a wu on 2 cpu's with a memory of =/> 1024. or in the future to add your pc to a network of boinc. so we are 1 big supercomputer, but that requiers high speed internet. glassfibre preferred ;)

Actually, the Blue Gene computer's memory per CPU is quite limited. The Rosetta team had to work to reduce their memory footprint to be able to use it.

If you will note, the article mentioned how great it was to have 40,000 processors available... but on R@h there are 150,000! The difference is that our 150,000 aren't all connected all the time, and they don't report their results immediately.

...but it would be closer to that, if the project used the BOINC trickle-up support for reporting partial completion of a task. This would seem very straightforward to implement by trickling at the end of every model, or every hour, whichever takes longer. It would also mean the project could save download bandwidth because a client machine could cruch the same task for much longer, and in fact the project could decide when ending that task would be desireable by sending out a kill messge during a trickle up.

Just an idea.
==============

I was wondering if the models done on Blue Gene were done to help determine which approach to use and/or which chain breaks should be sent out on Rosetta@home, or if Blue Gene actually did all of the work on one of the protein targets??

---

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/

so were the blue gene CPUs all working on the same model or each on a seperate model?

---

so were the blue gene CPUs all working on the same model or each on a seperate model?

all toghether working at 1 model. or in clusters of lets say 1000 cpu's at a model, just what was needed.

Any chance we could get a rough estimate of when the game is released, if all goes according to plan? I'm really looking forward too it. I just hope it's not too addictive, so i fail all my classes. Then again, I guess I'd learn a lot about protein structure, which is more than you can say for super mario world...

---

Any chance we could get a rough estimate of when the game is released, if all goes according to plan? I'm really looking forward too it. I just hope it's not too addictive, so i fail all my classes. Then again, I guess I'd learn a lot about protein structure, which is more than you can say for super mario world...

the current plan is to go public in January, with a continuous trial starting in a few weeks for people willing to help us work out the bugs

---

Any chance we could get a rough estimate of when the game is released, if all goes according to plan? I'm really looking forward too it. I just hope it's not too addictive, so i fail all my classes. Then again, I guess I'd learn a lot about protein structure, which is more than you can say for super mario world...

the current plan is to go public in January, with a continuous trial starting in a few weeks for people willing to help us work out the bugs

excellent, where do i sign up for bug testing?

---

Yes! We have been working hard on computational design of such an HIV inhibitor, and already have a candidate that look quite promising in silico. We should have experimental results on this designed protein in a couple of months.

I don't know if you've seen this:
http://parts.mit.edu/igem07/index.php/Ljubljana

---

among the new things to come in the relatively near future include virtual drug screening by large scale docking of small molecules onto protein targets. accurately docking even a single small molecue onto a protien target is fairly computationally intensive, so if we want to dock 100 or 1000 possible drug candidates onto a protein target large scale computing will definitely be critical.

would the results from the find-a-drug project be useful as a starting point for the virtual screening? The project was virtual screening of billions of molecules against thousands of protein targets...

---

would the results from the find-a-drug project be useful as a starting point for the virtual screening? The project was virtual screening of billions of molecules against thousands of protein targets...

i guess it is.. at least a something to compare results with so you know if your algorithm is right.

among the new things to come in the relatively near future include virtual drug screening by large scale docking of small molecules onto protein targets. accurately docking even a single small molecue onto a protien target is fairly computationally intensive, so if we want to dock 100 or 1000 possible drug candidates onto a protein target large scale computing will definitely be critical.

would the results from the find-a-drug project be useful as a starting point for the virtual screening? The project was virtual screening of billions of molecules against thousands of protein targets...

yes, they could be. we will start by studying successively more difficult drug docking problems where the solution is known experimentally, improving the docking methodology along the way, and then move to problems where there is no known solution.

---

It's great to hear that new cleaner code is coming. A big "thanks" to all the coders working on the project!

With all the excellent work going on at Baker labs and all the cycles that Rosetta volunteers contribute, I'm sure we'll accomplish something really big at some point or other. Exciting times!

---

Exciting times!

Indeed - in fact, the two announcements that there's a "new improved" version coming along, and that virtual drug screening is also on the cards, are a couple of good bits of news. Looking forward to sticking around and helping out...

And as dcdc said, if there was any way that the results from Find-a-Drug could be used - even just as a comparison between the FaD approach and the Rosetta approach, then that would be another great use of the FaD results. I don't think the FaD results are public, or were ever intended to be, but maybe with some sort of non-disclosure agreement, the results could be used for comparison and something useful might be learned.

---

Alver Valley Software Ltd - Contributing ALL our spare computing power to BOINC, 24x365.

ok this seems the place to post then,

ill be willing to volenteer for testing the game

luukNOlageschSPaar@hotAMmail.com

remove NOSPAM

hint include NOSPAM in your emailaddress, to avoid getting spammed ;) if a programm grabs the adress this way it won't know what to do... :)