Yes anything useful coming out of all this processing - is it possible to also provide a full ranking system of the top 1000 users ?) just to see who is in front i guess the big unis etc would take that honour.

Yes. I also would like to know how much of the subject project has been completed.

1. How long would it take to complete the total project? Any estimation?
2. How many % of the project has been completed, so far?
3. What can we do to help to expedite the progress?

As of today (mid feb 2005) we are 22% of the way through.
We are now in the process of post-processing that data and
getting the resulting protein annotations into our database
that will be free to the public. From those 22k runs we will likely
generate 11k fold predictions with confidence greater than 90%.
I won't know for sure the total number of good predictions for human
until the project is over, including resending failed runs and other
touch up type efforts at the end.

Also work units that will in effect be internal controls are in the process
of being completed. This work will help us to refign the confidence function
that we use to determine good from bad predictions and will make the
technology ultimately more useful.

Rich Bonneau
ISB

beekeepr,Waxing sarky, are we?? Well, let's try asking Rich some specific questions about our immediate progress and about the program that we see running in the Application View of the Grid Agent.

First off, let's start by stipulating that the precise number of proteins to be run is still not determined, but for the sake of discussion is taken to be 100K. Also, there will be some touchup units at the end, so the project will run full steam ahead until it is 99% complete, and then will slow sharply. By this time we can expect to mostly be running another project or projects, but some HPF units will be occasionally doled out. Perhaps, for PR purposes, we will call the project complete and give the touchup some other name. In essence this was done with the Human Genome Project. Everybody hates a PR black eye, but real scientific research seldom meets the standards of Madison Avenue, which is why all those people in white lab coats in the ads are actors.

Specific questions about project progress. You have received results for about 22K proteins back from our 2 grids, GRID.ORG and the World Community Grid.

1) How many have you sent from ISB?
2) How many have you got batched up ready to send at ISB?
3) Are you still running a computer program over the Human Genome to prepare more proteins to send or is it all done?
4) Are you using some specific database to prepare only a specific subset of the human proteins for us to crunch on?
5) If so, is there any question we could ask about which subset that would not be impossibly long to answer?
6) Could you drop a note when a batch of results is received back at the ISB, so we can mark another step toward project completion?
7) How do you batch up proteins to send to us – is there some set number of proteins per batch?
8) What sorts of batches of results come back? I assume that all the results for a protein will be kept together. But is an entire batch of proteins that you sent returned as a group? Several batches at a time? Are individual proteins, regardless of how they were received, just queued up as they are completed and returned? How many proteins are returned at a time?
9) Please note that I am not asking for any specific breakdown by grid. That is a public relations decision that should be decided jointly by Paul and Bill. But with that granted, I would personally prefer information about progress of this sort correctly attributed to each grid.

Added: [ Happy camper here] In accordance with Rick Alther's post below, note the terminology change. Read 'fold prediction' wherever I use the term 'result' below. 'Result' should only be used to describe a completed work unit.]

Then there are specific questions about Rosetta. The Application View shows progress jumping up several tenths of a percent when a result is completed and written to the file on the hard drive. If it jumps up 0.2%, I assume that the work unit requires about 500 results. If it jumps up 0.7%, I assume that the work unit only requires about 140 results and so on.

1) Are there a standard number of results required per protein?
2) Is this variation in individual work units just an attempt to split the work into short chunks?
3) Or, does the number of results required per protein vary with the nature of the protein?
4) For that matter, what is a result? I have been assuming that it represents an attempted fold, with some parameters varied.
5) Is each result returned with an attached Rosetta value?

There are more questions that could be asked about Rosetta, but I may already be wandering far off the path in the wilderness.

I do not ask for a long essay attempting to answer everything at once. You need to sleep and get a lot of work done while keeping up paperwork and attending meetings. Still, I think the questions presented here are of general interest.

Lawrence

( p.s: beekeeper, see the difference between specific questions and generalized discontent? )

Hey Joe,

I share your need for feedback, but I must say there has been at least SOME feedback on the progress, but you still have to ask for it. The UD software is still very much in development, and aside from other stuff, a linux client is supposed to be in the works, so these people probably got their hands full.. Also in relatively short notice, I expect to hear about the freely accessible database with the results that WCG produced.

As a last thing I`d say it`s nice to see your conscience patted on the back relatively often, but it`s not beneficial to anything. What matters is that you keep crunching, and that eventually people can be cured with what we do here. I inquired also for enough reason to keep crunching, and I received enough information to decide for myself that this project is indeed a good cause to invest free cycles in, and that I (Joe public) get something back out of this. WCG, Grid.org and UD are still in the process of establishing a concept for free contribution of resources to good causes, protecting the contributers rights and preventing commercial interest from running off with the good stuff. I hope they are sincere in their endeavers. I believe in the long term, this model can change the way the medical industry is built around their multi billion dollar research investments which they need to see repayed. I hope projects like this can be to the medical industry what the free software movement (linux et al) has been to the general software industry.

So there is more to this than just results. I hope this can make you feel better participating ? :)

Cheers