The graphics that run during rosetta@home are nifty, colorful and better than broadcast TV, but I have no idea what I am viewing. Is there an explanation somewhere, for the layman, of all the information being presented?

Hi Clif,

Here are two analogies that may help:

If you mix water with oil and then let it sit there, what happens? The oil and water separate.

Now imagine you have a string, with different beads. Some beads are "water-like" and some are "oil-like"... you throw the string into a bucket of water, what happens?

These properties or rules are the "score function". what Rosetta is trying to do is move the string of amino acids around, trying to maximize the score (putting hydrophilic (water loving) amino acids outside and hydrophobic residues inside). This is "folding".

Simulating the folding process (MD) is very time consuming (this is what folding@home does), Rosetta@home on the other hand takes a short cut by copying fragments from experimentally solved structures. Though any two proteins can be very different, if you chop it up into fragments of 3 to 9, these fragments are very similar.

What you see on the screen saver is rosetta trying to insert fragments into random locations, keeping those that improve the score... and flipping a coin on those that make the score worse.

Take a look at this page https://boinc.bakerlab.org/rosetta/rah_about.php, especially the videos at the very end.

=]

-krypton

Thanks for your quick response - I'll read over the information on the page you linked. I'm still working through the intro exercises on Fold-It, but I understand the oil/water example you gave.

With regard to the graphics display, I should be more specific...

- What is RMSD?

- low/high energy...which is better?

- what is the scatter plot showing?

- what constitutes completion of a run...is it a set time per run or is it waiting until it finds an acceptable configuration or energy level?

Thanks for your quick response - I'll read over the information on the page you linked. I'm still working through the intro exercises on Fold-It, but I understand the oil/water example you gave.

With regard to the graphics display, I should be more specific...

- What is RMSD?

- low/high energy...which is better?

- what is the scatter plot showing?

- what constitutes completion of a run...is it a set time per run or is it waiting until it finds an acceptable configuration or energy level?

RMSD is the root mean square deviation, put extremely simply its a measure of accuracy.

You want low energy's, you want to find the lowest energy state for a certain strain of amino acids (native structure = folded structure)

If I remember correctly, scatter plot is showing energies of the predicted states.

Basically when you run the program it tries and finds the best amount of suitable folded states in a certain amount of time, and one WU isn't enough time to find the native state therefore it needs more compute time. Pretty sure its simply time dependent, and simply finds the max number of lowest energy structures in an allotted time. i e, the longer you run the most structures you predict and the best chance you have of finding native.

Thanks for your quick response - I'll read over the information on the page you linked. I'm still working through the intro exercises on Fold-It, but I understand the oil/water example you gave.

With regard to the graphics display, I should be more specific...

- What is RMSD?

- low/high energy...which is better?

- what is the scatter plot showing?

- what constitutes completion of a run...is it a set time per run or is it waiting until it finds an acceptable configuration or energy level?

RMSD is the root mean square deviation, put extremely simply its a measure of accuracy.

You want low energy's, you want to find the lowest energy state for a certain strain of amino acids (native structure = folded structure)

If I remember correctly, scatter plot is showing energies of the predicted states.

Basically when you run the program it tries and finds the best amount of suitable folded states in a certain amount of time, and one WU isn't enough time to find the native state therefore it needs more compute time. Pretty sure its simply time dependent, and simply finds the max number of lowest energy structures in an allotted time. i e, the longer you run the most structures you predict and the best chance you have of finding native.

Thanks, Carl.