OK Howard..... Tell me true

[Insidious]

Are the results from the present protien being folded providing any benefit to your project (or mankind)?

or are we just marking time until the "phase II" client begins?

(no disrespect intended, I think it is a fair question)

-Sid

[IronBits]

Hey Howard... in that same vein of thinking
Does the Beta Client help your work in the science, or just the development of the client or both?

[Brian the Fist]

In the 'main' project, we have been working through a set of 17 proteins which we have decided to focus on from the very beginning (not counting the CASP targets). These are meant to represent a diverse set of folds and protein sizes. To date (don't have my list here) we've covered about a dozen or so of them.

While it may seem at first pointless to continue to crunch on these with the old algorithm, when we have this great new improved algorithm just about ready to go, it is certainly not a 'waste of time'. The phase I was always intended to set a baseline of sorts. It answers the question 'how far can you go with naive random sampling?' If we don't have a baseline to compare with, we don't know if the new algorithm has really improved things or not.

So obviously we want to get the new algorithm going as quickly as possible but we don't want to rush it either and so are doing useful work until it is ready to be unleashed.

In terms of the beta client, we are trying to tweak the method in order to get it to fold to the native state, from some starting random structure, using the RMSD as the scoring function (for now). While we are close, we are not there yet (because the helices keep disappearing). This should be working very soon and will be very cool in itself, as we will then have a way of rapidly generating multiple folding putative pathways (movies) for a protein. This should be interesting to the scientific community by itself (so we could write a paper on it).

Then when we swicth back to an unbiased 'energy function' instead of using RMSD we will be able to begin phase II and see how the new algorithm performs compared with the old.

In terms of helping science, everything we do is contributing at least in some small way to our final goal of solving the protein folding problem, and thus advancing modern medicine.

[Insidious]

Thanks Howard!

You have given my inspiration the boost it needed and explained very nicely, just what we are presently doing.

-Sid

[m0ti]

Thanks for giving all of us a better feel for what's going on with the project (what we've been doing, what's coming up...).

[Brian the Fist]

I suspect no one has noticed, but last week I also updated the web site (for the first time in a year or so ). Almost every page was rewritten at least partially to reflect changes, both in facts and in our 'vision' over the last year. Especially I have updated the About->Science section to reflect what we have accomplished so far, and what we intend to do next (since this was largely outdated beforehand). Its still far from perfect, but as I wear many hats I can only do so much at once . So if you have a minute feel free to check it out.

[m0ti]

Thank you!

[IronBits]

Check your bugs.html file "Known Bugs and Workarounds".
1st para... (these are the couuupt ones)
I'm learning more and more about this project!

[Drowis]

Any Idea when we are going to switch clients?
I have to keep the guys on team HardCoreWare updated since we are a fickle bunch after being burned by that Piece of crap of a prodject DDOL. most of us have been working hard over there but with the constant server failures and total lack of support over there we are all a bit jumpy when it comes to a DC prodject.

[RaginSteveK]

While we are close, we are not there yet (because the helices keep disappearing).

?? this is why the image of the reconstructed protein when compared to the native is missing helices[coils] and strips ??

how does RMSD work when the spatial arrangement is off
[ missing or lost helices, etc].. ?? does it matter ??

to this layman, it looks awefully weird to see all these twists and alignments in the genuine article and nothing in the regenerated one..

do the H-bonds or H-bonding segments need to be predicted, or do they [ should they ] happen automatically when the opportunity for H-bonding occurs ??