Will rosetta ever be out of work, because it has tested all the proteins that it can test?

And is there an estimation of how long before it will be done if there ever will?

Roland, the current work on Rosetta is focused primarily on understanding how the amino acids that make up a protein interact, and function. The ultimate goal is that the program will be able to predict accurately, based solely on the amino acid sequence of a given protein, what shape it takes. And once you can do that, you can learn more and predict what shape it will take when bound to a second protein, and whether a second protein will in fact bind to it.

This ultimate goal is not one that will be achieved in a flash of inspiration. It is more evolutionary with greater and greater accuracy as the research continues. And with larger and larger proteins being predicted.

Once you have that, then there are literally 100s of thousands of proteins that would be useful to know the shapes of, and to understand the potential interactions between. At Rosetta@home's current rate of work processed, my best guess, assuming the processing time used by today's program will eventually provide the desired higher accuracy described above, is that you have enough computing power to solve roughly one protein per day of work. So, even once the program is "completed", you still have at least 100,000 days of work (that's 274 YEARS!) ahead. Hopefully computing speeds continue to increase over that time :) On the other hand, it is likely that the processing time taken to solve the protein with higher accuracy will require MORE processing then is used today.

Once you know the shape of 100,000 proteins, now you'd like to study the interactions between all of them.

In short, there's no end in sight. The only light in this tunnel is the one shining forward.

I didn't mean to sound hopeless. If we can accurately solve just one protein and predict a fraction of it's interactions, we could cure HIV. And it is entirely likely that once we better understand the interactions between the amino acids, that accurate predictions could be made with significantly less computing time then it takes today. It is also likely that more computers would be devoted to the task, once research is largely completed and you are in to doing production runs on thousands of proteins. So, the basic assumptions I made will ultimately prove to be flawed. But, from here, there's no telling in which direction :)

---

Rosetta Moderator: Mod.Sense

Thank you, great reply.

Thank you, great reply.

He got a million of them ... :)

I will point out to your edification that there are nearly a dozen BOINC projects doing various studies in one form or another on these issues. That is how important and useful solving as the man said, even one of them...

I am getting muzzy headed, but WCG I think has 3 or 4 sub projects using some form of folding technology, IBERCIVIS has I think one sub-project, Docking, ProtienPredictor (though they still seem to be dead), Proteins, and on and on ...

CLosely related are Cels (old and new), Spinhenge ...

Where is the 'new' cels@home Paul? I crunched the 'old' I presume, but that went away last year.

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

Where is the 'new' cels@home Paul? I crunched the 'old' I presume, but that went away last year.

Here

Though they have not had work for some time and have been off the air for a couple weeks.

I think I only about 6K before they went back into their hidie-hole

Okay, I tried but nothing. Ah well.

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

Yeah, well they came up, ran for awhile and dropped below the waves again ... lots of projects do that ... the "low cost" seduction ignores the "hidden" costs of standing up a BOINC project. Like pesky participants, managing the servers, software bugs, etc. ... all that junk ...

Yeah, well they came up, ran for awhile and dropped below the waves again ... lots of projects do that ... the "low cost" seduction ignores the "hidden" costs of standing up a BOINC project. Like pesky participants, managing the servers, software bugs, etc. ... all that junk ...

I was thinking Dr. A and group ought to design a basic Boinc setup, computer, software the whole nine yards and then publish that somewhere so someone could start their own Boinc project. Kind of like a plug and play setup. Yes I know the Science, it is all about the Science and IT is very flexible, but maybe the holdup with some newer projects, and the demise of some older ones, is the lack of a stock setup.

---

mikey, they have. Check the details on the BOINC server VM here. This has been available for some time, but is still not the entire story. Most projects still have to customize an assimilator, validator, often the scheduler, etc. etc. etc. And they have to keep their servers running, their network, their science.

All of these unwanted complexities is why I am planning to offer a BOINC server as a service. But I question if projects will be able to pay someone rather then endure the higher costs they encounter doing it themselves month by month.

---

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/

Yeah, well they came up, ran for awhile and dropped below the waves again ... lots of projects do that ... the "low cost" seduction ignores the "hidden" costs of standing up a BOINC project. Like pesky participants, managing the servers, software bugs, etc. ... all that junk ...

I was thinking Dr. A and group ought to design a basic Boinc setup, computer, software the whole nine yards and then publish that somewhere so someone could start their own Boinc project. Kind of like a plug and play setup. Yes I know the Science, it is all about the Science and IT is very flexible, but maybe the holdup with some newer projects, and the demise of some older ones, is the lack of a stock setup.

It is more of a story of the lack of a system that can be easily setup and configured. The saddest part is that the trial and tevails of new project are not captured and used to refine the processes and documentation. There is no proactive intervention as it were. If you watch the mailing lists you tend to see pretty much the same questions arise from new projects all the time.

We see the same attitude towards documentation of the BOINC Manager. It is superficially done and leaves too much to the imagination of the participant. In that many people are barely able to get the computer turned on and working it is a lot to expect them to figure out much of this nonsense.

Well, I have beat on this topic over the years to no avail ... there is no interest in making BOINC easier.

mikey, they have. Check the details on the BOINC server VM here. This has been available for some time, but is still not the entire story. Most projects still have to customize an assimilator, validator, often the scheduler, etc. etc. etc. And they have to keep their servers running, their network, their science.

All of these unwanted complexities is why I am planning to offer a BOINC server as a service. But I question if projects will be able to pay someone rather then endure the higher costs they encounter doing it themselves month by month.

ALRIGHT!!!! Now if you can also help with getting project xyz off the ground then people could be beating a path to your door! Costs are always issues but if you could find a way to fix the costs, kind of like a contract, then you might just find people/places willing to spend the money. I am sure there are many groups with ideas floating around that think they could use Boinc but have no idea how to do it, or even for sure if it would help. I think those kind of people are your target audience.

---

DC has not come of age yet. It's mainly the preserve of computer "geeks" (I use the term in the nicest possible way)

The majority of people don't know about DC and more importantly don't want to know. I have spent many hours telling people how they could help scientific research on a myriad of fronts, also helping themselves and the rest of mankind but unfortunately APATHY and the good old I'M ALRIGHT JACK attitude is alive and well.

Then there's the scientific community itself. Most scientists wouldn't trust their research to be done by Joe Public infact the majority consider DC and BOINC as a joke and its participants as ignorant philistines.

Few project scientists that DO run DC don't see why they should spend anytime at all giving to their contributors any information or updates about the work they're doing on their computers afterall they wouldn't understand, would they?

EDIT to add; It struck me while reading my post that some of you may wonder why I contribute to DC given everything I have written. The answer is I am NOT most people or the majority. I believe in DC and wish man would realise that collaboration is the key to surviving this century in any recognizable form.

---

DC has not come of age yet. It's mainly the preserve of computer "geeks" (I use the term in the nicest possible way)

I resemble that remark ...

DC has not come of age yet. It's mainly the preserve of computer "geeks" (I use the term in the nicest possible way)

The majority of people don't know about DC and more importantly don't want to know. I have spent many hours telling people how they could help scientific research on a myriad of fronts, also helping themselves and the rest of mankind but unfortunately APATHY and the good old I'M ALRIGHT JACK attitude is alive and well.

Then there's the scientific community itself. Most scientists wouldn't trust their research to be done by Joe Public infact the majority consider DC and BOINC as a joke and its participants as ignorant philistines.

Few project scientists that DO run DC don't see why they should spend anytime at all giving to their contributors any information or updates about the work they're doing on their computers afterall they wouldn't understand, would they?

EDIT to add; It struck me while reading my post that some of you may wonder why I contribute to DC given everything I have written. The answer is I am NOT most people or the majority. I believe in DC and wish man would realise that collaboration is the key to surviving this century in any recognizable form.

Okay so how do we change the attitudes of those that don't 'believe' in DC? I have long advocated that those that use Boinc should say so in their Scientific Papers that they write and espouse the benefits at the conferences. But how do we actually get them to do it? Is it the fact that Seti is so 'out there' and no one wants to be associated publicly with it?

---

Okay so how do we change the attitudes of those that don't 'believe' in DC? I have long advocated that those that use Boinc should say so in their Scientific Papers that they write and espouse the benefits at the conferences. But how do we actually get them to do it? Is it the fact that Seti is so 'out there' and no one wants to be associated publicly with it?

It is the old and guiding principle behind unions... strength in numbers. Yet all the projects act alone and work alone. In the early days for example Einstein's developers did a lot of work and much of that was rolled back into the server code. But, they also have a very robust architecture with their systems and I am not at all sure that those changes and enhancements are in the base system. It certainly is not in the documentation. So, that advancement is not available. The same for whatever WCG has done behind their closed doors ...

It was pointed out on the mailing list yesterday that we are at 293K active participants, not good numbers by anyone's accounting. I mean I am doing my part with adding two i7 920s with two more GTX260 cards while retiring two old Xeons and a dual core (net gain 6 CPUs, higher speeds and 2 GPUs), but, this is the brutal math ... too few participants ... it has always been true ... and almost looks like it will always be true ... because the very people that we need to do some work are just not interested ...

Okay so how do we change the attitudes of those that don't 'believe' in DC? I have long advocated that those that use Boinc should say so in their Scientific Papers that they write and espouse the benefits at the conferences. But how do we actually get them to do it? Is it the fact that Seti is so 'out there' and no one wants to be associated publicly with it?

It is the old and guiding principle behind unions... strength in numbers. Yet all the projects act alone and work alone. In the early days for example Einstein's developers did a lot of work and much of that was rolled back into the server code. But, they also have a very robust architecture with their systems and I am not at all sure that those changes and enhancements are in the base system. It certainly is not in the documentation. So, that advancement is not available. The same for whatever WCG has done behind their closed doors ...

It was pointed out on the mailing list yesterday that we are at 293K active participants, not good numbers by anyone's accounting. I mean I am doing my part with adding two i7 920s with two more GTX260 cards while retiring two old Xeons and a dual core (net gain 6 CPUs, higher speeds and 2 GPUs), but, this is the brutal math ... too few participants ... it has always been true ... and almost looks like it will always be true ... because the very people that we need to do some work are just not interested ...

I still think part of the problem is Seti and the way the other projects don't try to set themselves apart from it. I mean I talk to people at work, and in my personal life too, and they are all very interested when I tell them about Rosetta, Malaria, Folding, etc but when they ask how I got started and I say Seti, their eyes glaze over and they think 'what a geek'! Seti is just a project that you either love or hate and too many people dislike it, or don't know enough about it. Me I have moved on from Seti, but that is me and a problem with how I see management running things, not because of a problem with the concept. Star Trek and Star Wars both came out when I was already grown and I am old enough to think the idea could be true, if only we could find them.

---

Is there away to know how many proteins I have contributed to Rosetta? Is it possible to show on the front page how many proteins have been completed? I know it's possible to view from work units that you are processing but I find the instructions to hard to follow. Is there a video showing us how to view the proteins?
Thanks in advance

---

Have a crunching good day!!

Speedy, those instructions are for viewing the specific models your machine came up with. i.e. the various possible shapes that you found being possibilities for the protein studied by a specific task.

You might be interested to see the list of your results. Click the "Results" link on the homepage under the "returning participants" section. It shows a list of the various types of work you have done, and how your predictions compared with others that worked on the same proteins.

As transient says, there is no finite list of proteins being studied, nor of techniques being used. In fact many of the proteins being examined have already had their structures determined by other means. The purpose of the tasks is to see if the same (or similar) structure could have been predicted on a computer rather then using the expensive and time consuming alternatives. Since the result is often close but not identical, the scientists rework the approach and send more tasks to see if they can come closer to the result that was found in the lab.

See also my prior post in the thread here.

---

Rosetta Moderator: Mod.Sense