I know this is silly but I really like to watch the WU do its thing and rosetta not having any graphics kinda upsets me. Are there any others who would like to see the graphics too?

---

I believe there is a screensaver in the works. But I have no idea when it might be released.

Kathryn

---

Kathryn :o)
The BOINC FAQ Service
The Unofficial BOINC Wiki
The Trac System
More BOINC information than you can shake a stick of RAM at.

A really neat screen saver is almost ready that will allow you to watch the protein as it is folding up. We are exctied that collectively looking at the screen saver and comparing results may lead to new insights into protein folding and structure prediction.

When the screensaver is ready, it will be very easy to explain what "abrelax" means (the question on the other thread). The current "abrelax" batch of runs have two stages. the first is the "ab intiio" stage, in which the protein starts out as a completely straight featureless chain, and rapidly searches through a very large number of conformations. you will see the protein flailing around wildly during this stage. the second, more time consuming "relax" stage, a much more detailed representation is used, and you will see the energy decreasing dramatically but relatively little chainge in the protein backbone. "abrelax" is our shorthand for "ab intio" followed by "relax".

---

Oh, now this sounds interesting ... :)

Maybe we can attract more people with something like this ...

---

I can't wait myself. Since I use the BOINC screensaver I generally set projects without graphics to backup status. You can see this looking at the RAC in my sig. Projects with graphics tend to have 50 to 70 RAC, projects without tend to have 1 or 2 RAC.

---

BOINC WIKI

BOINCing since 2002/12/8

A really neat screen saver is almost ready that will allow you to watch the protein as it is folding up. We are exctied that collectively looking at the screen saver and comparing results may lead to new insights into protein folding and structure prediction.

When the screensaver is ready, it will be very easy to explain what "abrelax" means (the question on the other thread). The current "abrelax" batch of runs have two stages. the first is the "ab intiio" stage, in which the protein starts out as a completely straight featureless chain, and rapidly searches through a very large number of conformations. you will see the protein flailing around wildly during this stage. the second, more time consuming "relax" stage, a much more detailed representation is used, and you will see the energy decreasing dramatically but relatively little chainge in the protein backbone. "abrelax" is our shorthand for "ab intio" followed by "relax".

Thanks for explaining that, I can't wait for the new stuff. It sounds cool

---

Hi, I'm a researcher in David Baker's lab and one of the people working on the screen saver graphics. We're very excited to have BOINC users watch our simulations in action. We believe you may even be able to provide us with feedback about the behavior of our algorithms. To that end, we'd like to make the graphics show as clearly as possible what is happening during the simulation. And we'd like to make it possible for you do discuss with us, and each other, features that you observe.

Some aspects of protein folding are quite amenable to human intuition. On some of the other message boards there have been discussions of the important features of proteins. From a visual perspective we can think of proteins as molecular chains that fold into specific shapes. A very common way of viewing proteins is as a "backbone trace." This view strips away most of the atoms, and simply shows the path the chain takes through space.

This reduced view allows one to see the overall "fold" of a protein, without worrying about the details of where every atom is. We propose that this would be the first kind of representation we would use on the Rosetta screen saver. This view would be colored so that you could see and talk about specific regions of the fold.

What you would actually see on your screen would be the repeated search (using a different random seed) from an extended chain to a compact folded structure. This will take a few minutes on a typical computer. This structure may or may not be close to the right fold, and may or may not have a good energy. We are trying to understand what features of search allow us to sometimes get the right fold.

We are interested in details you notice during this search process. At first this might be things like "The red and yellow part always lock in first, while the blue part flops around." Or, "The structures with the lowest energies always have the green part fold last."

I anticipate that we can evolve the screen saver, particularly as people become more experienced looking at it, to provide more technical information. As David says, it will also be easier to discuss the nature of our search algorithm once you see it in action.

I will try to post a screen shot

---

Just to get you started thinking about what you might see, here is an example of what a protein might look like on the rosetta screen saver.



It shows some common features of protein folds. The green region is helical, like a slinky, and has the technical name "alpha-helix". Usually we just say "helix". The other main features are extended "(beta) strands". Strands often associate into "sheets." The orange part is a "hairpin" bringing the yellow and red strands together. It is hard to say from this static image, but it looks like the blue strand is trying to sneak in and associate with the yellow strand.

On the screen saver we could also show you what the correct fold for this sequence looks like, so you could compare them.

---

Just to get you started thinking about what you might see, here is an example of what a protein might look like on the rosetta screen saver.

Cool! Would be nice if it could also, perhaps in a corner, show the RMSD and Energy figures, as the calculations are happening. Where we're working on a known structure, it could also show the "target" figures so we can see how it's moving along.

Something like this perhaps:

         Current:  Target:
RMSD:       1.57      1.06
Energy:  -195.75   -214.51

---

*** Join BOINC@Australia today ***

Oh! Oh! (hand waving in the air) ...

Rather than just talking about it, a way to save an image ...

I can use one of several tools, but not everyone may have a screen capture tool (yes windows has the key thing, but how many know how to use it?) ...

I mean, if I see some "behavior" and see something that is "common", I would take before/after shots to try to give a feel for what I "saw" ...

Just a trivial addition ... as we used to say (when I did coding), "SMOP" ...

Small Matter of Programming ...

---

And I'll chime in for the 10% of the male population with some form of red-green color blindness. . . If there is something useful that can come out of reporting things based on the color of a section of the protein, can you use colors that won't make it hard for us to report?

For example, try to not have browns or purples. Otherwise, you might get reports back that don't make sense.

Not a huge deal, it's obviously a small group of people, but it may save you from hunting down the 'green spiral that is getting close to the blue zig-zag' when neither colors are in the picture.

Cheers,
Ethan

---

All three of these last posts raise very good points that we have already been considering.

1) The energy and RMSD will be on the screen. At least as numbers, and perhaps a graphically, as a function of time.

2) When known, the structure of the experimental "right" answer will also be shown.

3) Eventually, there will be some sort of capture mode. Not just of images, but probably the coordinates of the actual protein structure. It's not completely clear yet how this will work, because some parts of the folding simulations go by very quickly.

4) 10% of potentially thousands of users will be a significant number. I've been reading up on color scales that might work for all viewers. If you have any other suggestions, I'd like to hear them. I'm hoping to design in some more cues besides color.

---

Jack,

The suggestion of a graph is super. I always liked the SETI@Home screen saver (when I ran it) in that it gave me different looks at the work in progress. So, that kind of display can be very powerful. Not sure how to recomemnd its look ... but, the current state, with long term displays ... killer ...

Also not sure how to do this ... but, you can also look at the CPDN advanced visualization tools. I have YET to use them (time, always no time) but I like the idea of being able to make a log and then look at it later ...

Sort of like playing a movie... with that, I would likely collect the data and then look at several at the same time ... just ANOTHER thought ...

With this capture, you can then, ask us to save the data (as does CPDN) for later return ...

---

It would be nice for Rosetta to have graphics, but I probably wouldnt use them as screensavers tend to process workunits slower than leaving it alone...

Spectre

---

It would be nice for Rosetta to have graphics, but I probably wouldnt use them as screensavers tend to process workunits slower than leaving it alone...

Spectre

Indeed, from a purely production standpoint having zero graphics is the best. And my understanding is that boinc will allow you run without graphics to maximize your rate of credit accretion.

We will also use a boinc feature that allows us to limit the graphics to no more than a specified percentage of the cpu.

Our hope is that the graphics are interesting enough to encourage more people sign up.
Also, it might be educational for people to see what their computers are doing. And finally, we really do hope that you will share with us your insights into what you see happening.

Each structure we produce (the points on the Energy vs. RMSD plot) results from a trajectory of tens of thousands of moves. We are currently not equiped to analyze each of these trajectories in detail. It is a very large quantity of data. It seems quite possible that interested people watching animations of these trajectories may provide useful feedback for us.

---

I myself always leave the graphics off but I know alot of people that love to see them. Having these new cool graphics will definitely help the project!

---

Yes, a screensaver or a graphic will be wellcome, but again, as mentioned by others earlier, not at the expense of developing the "real" project. I don't have any screensavers myself, as I want to crunch so effectively as possible, so my screen goes into black when I don't do anything on my computer. But some kind of graphic can attract a lot of people, who like to watch, what's going on in their WU.

---

[b]"I'm trying to maintain a shred of dignity in this world." - Me[/b]

Here's something that could please both sides of the argument: add a pretty screensaver, but also add an option for it to stop after X minutes (at which point a user could set the 'standby mode' of their monitors to kick in). I might be pleased watching my proteins right after I'm done working, or coming back from a break, but there's no point for my CPU to be doing 3d animation while I'm asleep...

---

Team CFVault.com
http://www.cfvault.com

I would vote for a graphics button so I could pop into the graphics then close them when I had my 'fix'. I am in the camp that wouldn't use a screensaver unless it's the only way to see the graphics, then turn it off to maximize crunch time. At any rate, congrats to Dr. Baker and team for the very satisfactory project.

---

"Be all that you can be...considering." Harold Green

I would vote for a graphics button so I could pop into the graphics then close them when I had my 'fix'. I am in the camp that wouldn't use a screensaver unless it's the only way to see the graphics, then turn it off to maximize crunch time.

I agree, this would be great. I just about never view graphics because I want to concentrate my resources on doing the science. But if there was an easy way to turn them on and off "at the touch of a button", this would suit me just fine! :)

---

Regards,
Bob P.