241 replies to this topic

[DominO]

Since there was no answer when I proposed http://folding.stanford.edu/, I want to know how many are interested.

The thing is pretty much like SETI, but insteed of analysing datas from space, it calculate protein foldings, there are billions and billions of billions of possibilities that have to be calculated.

All what you need to do is to install and run the program in the background or/and screen saver.

This was as answer to Azat critic about SETI. http://folding.stanford.edu/ is practical in the research of many diseases, as well as the understanding of life.

If you oppose to it, I want to know why, I just want to know how many people are interested and if it worth building a hyeforum group to be in the first 100.


EDITED: If you agree don't install it now, first the Hyeforum group must be submited, I'll do every technicality, I just want to know if there is enough people wanting to participate.

Edited by Fadix, 04 October 2004 - 01:28 PM.

[vava]

Interesting idea Domino!

I found this info on the site - as we discussed in the SETI thread, it's a collective computing initiative. I wonder how effective collective computing is? Sip?

What does Folding@Home do?
Folding@Home is a distributed computing project which studies protein folding, misfolding, aggregation, and related diseases. We use novel computational methods and large scale distributed computing, to simulate timescales thousands to millions of times longer than previously achieved. This has allowed us to simulate folding for the first time, and to now direct our approach to examine folding related disease.

How can you help?
You can help our project by downloading and running our client software. Our algorithms are designed such that for every computer that joins the project, we get a commensurate increase in simulation speed.

[DominO]

Two more people voting yes! and I'm going to found the group.

BTW, I count for 2, because I will be runing 2 computer on it.

So we have right now 5 computer in our disposal.

I have 2800+ Athlon and 2000+ Athlon that will be runing it.

So, how many other offer some processor juice for the sake of humanity.

Eve. where are you dear.

Edited by Fadix, 05 October 2004 - 09:15 AM.

[DominO]

OK! I have founded the team regardless of the results.

The team web-site is hyeforum.com, and I placed its logo as the HyeForum one.



All what you need to do is go here.
http://folding.stanf...u/download.html

And download the Graphical client of the system you have(Windows, MAC, Linux etc...)

And then, install it, for the name, choose your Hyeforum alias.

What is IMPORTANT, on the user option, when the program first open, after that you have set the User Name(the same as yours in Hyeforum, so we can see how much you did ) DON'T FORGET the Team Number.

Our team Number is: 39937

Without the team number, the computations done with your system won't be added in our team.

[vava]

Thanks Domino - will install this evening.

[DominO]

Don't be surprised if you get one WU that would take over 24 hours to be compleated... in one of my computer I got one taking 30 minutes and another one, something that would take 1 day 18 hours. This is probably why the positions are done on a scoring system based on processed informations rather than WUs.

EDITED: Sound that WU are cut pretty much the same way, I have made a mistake when taking a look at the other computer.

Edited by Fadix, 05 October 2004 - 04:27 PM.

[vava]

Thanks Domino - will install this evening.

So what's a WU?

[DominO]

So what's a WU?

It's one pack of information to be processed, they do not require all the same time to be processed, this is why the positions are not given based on that.

Eachtime the program download before being processed, it is one WU that you have downloaded, once the WU is finished it send the information processed and recieve a new WU.

Edited by Fadix, 05 October 2004 - 03:37 PM.

[DominO]

Will be cool if people could try to find as much Armenians as possible to join. We could have an Armenian team on the top 100 of an international team competition...

Everyone will know about Hyeforum then.

Edited by Fadix, 05 October 2004 - 04:10 PM.

[DominO]

Ahmmmm. ahmmmmmm... anyone trying it?

[Azat]

Okay Fadio. I installed it.. What are we folding anyways. Laundry?

[DominO]

Is it a real question or a joke?

[Azat]

Its a joke my friend. I know we are not folding laundry, but we are making origami. right?

[DominO]

MosJan, can you suspend Azat for 3 days please?

[Twilight Bark]

Well, they may not be folding laundry, but they are not folding sufficiently realistic models of proteins either. It is unclear to me whether their force fields are reliable enough to almost always lead to the right geometry for all proteins. And the part where they claim that not having explicit water molecules near the protein or some critical sites (which are probably unknown before folding) is no big deal is downright scary. The project seems like a scheme to supply a particular research group with free computational resources rather than a true "grand-challenge" where one is certain to reach the desired end by applying enough brute force.

I dare the computer wizards of HyeForum to come up with a generic grid-computing platform that is to be shared by Armenians worldwide. Something like "Massively Armenian Computing", where a poor graduate student or a professor would have an "unfair" advantage by virtue of being Armenian.

[DominO]

Well, they may not be folding laundry, but they are not folding sufficiently realistic models of proteins either.  It is unclear to me whether their force fields are reliable enough to almost always lead to the right geometry for all proteins.  And the part where they claim that not having explicit water molecules near the protein or some critical sites (which are probably unknown before folding) is no big deal is downright scary.  The project seems like a scheme to supply a particular research group with free computational resources rather than a true "grand-challenge" where one is certain to reach the desired end by applying enough brute force.

I dare the computer wizards   of HyeForum to come up with a generic grid-computing platform that is to be shared by Armenians worldwide.  Something like "Massively Armenian Computing", where a poor graduate student or a professor would have an "unfair" advantage by virtue of being Armenian.

TB, of course in such projects, the risk always exist that you may end up doing some computations for a project other than the one you are supposed to help.

On the other hand, the group results were made available, and from everything I have seen it seems to be very transparent.

"It is unclear to me whether their force fields are reliable enough to almost always lead to the right geometry for all proteins."

This issue was raised in an article published in Nature (Nov. 2002.) Comparing experimentation and the simulation, and the simulation predicted the experimentation with a pretty good prediction. Oh and the simulation used was folding@home.com. Two years later, they report major changes and advancements in the software used, so I guess it is precise.

[DominO]

I have found the article I was talking about in this page.

http://www.ifrance.c...ome/papers.html

There are some others too there that I have not read. So can't comment on everything on that page.

Edited by Fadix, 10 October 2004 - 09:20 PM.

[Twilight Bark]

TB, of course in such projects, the risk always exist that you may end up doing some computations for a project other than the one you are supposed to help.

On the other hand, the group results were made available, and from everything I have seen it seems to be very transparent.

"It is unclear to me whether their force fields are reliable enough to almost always lead to the right geometry for all proteins."

This issue was raised in an article published in Nature (Nov. 2002.) Comparing experimentation and the simulation, and the simulation predicted the experimentation with a pretty good prediction. Oh and the simulation used was folding@home.com. Two years later, they report major changes and advancements in the software used, so I guess it is precise.

I wasn't referring to any deception by the researchers. Such dishonesty is rare enough not to worry about it until there is reason to. What I am trying to say is that it's not like a true experiment in which you are querying nature using brute force. It is a computational "experiment" that exquisitely depends on the details of the model. One can easily fool oneself about the reliability by getting back what is wired into the model to begin with. Clearly, they would be using force fields that favor beta-sheets or alpha-helices for the appropriate aminoacids, by way of their parameterizations. This would give the correct result for small proteins with simple folds. However, it could easily fail miserably for the more challenging cases.

There is no doubt that such computational experiments will one day become foolproof, reliable, and feasible. However, I suspect that, the current effort is more helpful to the computational researchers in understanding and improving their methodologies than it is to pharmaceutical companies trying to find a cure for some disease. Still a valid and useful goal, but perhaps not a widely understood one.

[DominO]

I wasn't referring to any deception by the researchers.  Such dishonesty is rare enough not to worry about it until there is reason to.  What I am trying to say is that it's not like a true experiment in which you are querying nature using brute force.  It is a computational "experiment" that exquisitely depends on the details of the model.  One can easily fool oneself about the reliability by getting back what is wired into the model to begin with.  Clearly, they would be using force fields that favor beta-sheets or alpha-helices for the appropriate aminoacids, by way of their parameterizations.  This would give the correct result for small proteins with simple folds.  However, it could easily fail miserably for the more challenging cases.

There is no doubt that such computational experiments will one day become foolproof, reliable, and feasible.  However, I suspect that, the current effort is more helpful to the computational researchers in understanding and improving their methodologies than it is to pharmaceutical companies trying to find a cure for some disease.  Still a valid and useful goal, but perhaps not a widely understood one.

Really complex protein simulation is pretty new for them, and by default the software disable them. You have to enable the over 5 mb file to enable this option. I have enabled them for the next folding units. Those ones are on trial and errors to be more precise, but what is as default(not the new generation) are not very complex ones.

[DominO]

Hyeforum is 7529nt position (there are 33709 teams). I've compleated 2 WU, Siamanto and Azat 1.

Where are the others?

Edited by Fadix, 13 October 2004 - 03:42 PM.