Lowest RMSD

[alpha]

http://www.btinternet.com/~matt_mills/progress.jpg

http://www.btinternet.com/~matt_mills/generation_stats.jpg

http://www.btinternet.com/~matt_mills/energy_graph.jpg

...and yet...

http://www.btinternet.com/~matt_mills/rmsd.jpg

Now, I haven't done DF for a while, so.. is there something I am missing or am I just being stupid, or what? According to dfGUI, I should have the best score so far.

Yes, I have uploaded the generations with the lower scores :)

[willy1]

I believe that 'Energy' as shown by dfGUI is not the same thing as 'RMSD'.

There are a couple of threads in the main forum about this. RMSD is calculated by the server after you return your results.

willy1

[magnav0x]

Willy1 is correct, dfGUI doesn't accurately depict your RMSD. I've had dfGUI show best energies that were below 0.

[alpha]

Okay, so is there a relationship between energy and RMSD?

Also, is this a recent change?

[Hagar]

there is no direct relation between energy and RMSD
RMSD is calculated on the server after you uploaded your results
energy is calculated by the DF Client (NOT dfGUI)

dfGUI gets that information from the file progress.txt in your distridfold directory.

It has been like this ever since Phase II of the project started.

[Digital Parasite]

Originally posted by magnav0x
Willy1 is correct, dfGUI doesn't accurately depict your RMSD. I've had dfGUI show best energies that were below 0.

Not only does dfGUI not accurately depict your RMSD, it doesn't depict it at all. ;)

That why is says "Energy" and not RMSD in dfGUI. As Hagar pointed out, there is no real relationship between energy and RMSD. Before in phase I when we were crunching known proteins, the client would display RMSD since it had something to compare against. Since we are now trying to fold "blind" there is no way to see what the RMSD is before hand (and it gets calculated at the server end). If we were doing the CASP challenge there would be no known RMSD so all we would have is the energy values.

At the beginning of Phase II, the client displayed Energy values, and then for a few proteins Howard switched the client to display RMSD, but for a while now has switched back to Energy again.

If you are interested, you can read more about the RMSD and Energy here:
http://www.blueprint.org/proteinfolding/distributedfolding/trades_science.html

Jeff.