A serious question merits a serious answer. If you read up on the Rosetta program in PUBLICATIONS at http://bakerlab.org/ at the University of Washington where the program was developed, you will see that there is never a definite indication that a protein structural prediction made by Rosetta is correct. There are a lot of tweaks possible that can produce various predictions and likelihoods but they all need verification by other means.

Consider the implications of this. [Really Big Grin] For some reason I am thinking about a bumper sticker about a gun owner's flexible fingers from which his gun can be pried. The original press release made on 16 November 2004 by the Institute for Systems Biology estimated about a year to run the project, without even knowing how many computers would be running Rosetta. [Snicker]

I am sure that the Advisory Board is on top of this, but I do not personally expect any definite ‘per cent completed’ announcements to be forthcoming from the ISB. [Arch Lift of Eyebrow] Rather I expect that at some point in time other projects will be added and the server will control what percentage of our computers still run Rosetta. Eventually, I expect that the completion of Phase 1 [Muted Chortle] of the Human Proteome Folding Project will be announced. Perhaps from time to time there will be some supplementary runs accepted as projects.

All of this will be a matter for discussion between the Advisory Board and the Institute for Systems Biology. My personal guess is that full and frank discussion on these matters will be conducted more in private than in public.

But all this is just my personal opinion. I have no more information than anyone else. My guesses are based on my past experiences in other fields.

Lawrence

Whoops! My apologies if you thought I was reacting to your post. I was just thinking in different terms. The real importance of this project in human terms will be the use made of our results by medical researchers, but these are several steps away from us. We are part of a chain. We help the ISB. The ISB produces structural predictions of proteins with various degrees of verification. Then in other labs around the world medical researchers will see if this information helps them produce an application, such as better diagnostic tests or better treatments.

I was just thinking in terms of the immediate technical answers we are working on, rather than in terms of our ultimate goal.

You answered it quite well, Graham. I did not realize how my version might seem.

Lawrence
Added: A good analogy would be with a screwmaker helping to build a steam engine. I get carried away discussing the various possible pitches I could use in cutting the screw thread and take the overall steam engine for granted.

Hi, Graham. [Grin] your memory is eliding events there. We spoke together on the Memory Allocation thread at http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=236 where I was being acerbic. This was just after your admirable post http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=1246 and before the possibly better http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=1271 post.

Part of the cause of my ire was that I am perfectly happy to let hardware handle virtual memory. Of course, if the working set of memory pages remains so large that we limit the number of computers that can run it, then it is useful to strip out unused subroutines and their associated data structures. However, the important thing to concentrate on is the algorithm. My experience has been that statistical programs tend to put scientists at a loss. If you are lucky there is one scientist in a project who can speak with a mathematician, so you have to find him and use him to translate. So the really important thing is to get the ISB people experimenting on the method of selecting test snips of proteins that Rosetta uses. Using the best set for each type of protein will give the quickest, most accurate results. This sort of methodological improvement is a far more productive use of their time than asking them if they want a variant of Rosetta that does not have function Zeta in it. [Insert my standard rant on how mathematics is generally taught as a non-constructive subject so students have a hard time applying theorems to real problems.]

Graham, tis is a really good link at http://www.forcasp.org/modules.php?name=Papers&file=article&sid=2178

You are digging up some very good information.

Thanks,
Lawrence