3 replies [Last post]
phi16's picture
User offline. Last seen 1 day 19 hours ago. Offline
Joined: 12/23/2008

Voids seems to be colateral damage, a consequence of the geometry of backbone and sidechains that just happens along the way to a finely folded protein. Yes, it is good information. Perhaps, even useful to the more adroit folders out there. But it seems to be what's left over when the story is finished; not the story itself. It's like trying to draw a birdcage by drawing the spaces between the bars. The spaces are the voids; the bars are the backbone and sidechains.

What folders want as feedback is what are the forces acting on the various sidechains and where do they come from. A void is just a left over piece of space that occurs as a result of the interactions of the geometry of the backbone and magnetic forces of sidechain atoms.

axcho's picture
User offline. Last seen 9 years 38 weeks ago. Offline
Joined: 12/18/2007
Groups: None
Ah, I see what you were

Ah, I see what you were talking about now. Yes. This makes sense.

I had an idea that may be similar to this. I wonder if it is the same sort of thing that you are hoping for. I was thinking that it would be very helpful if after you made a move (such as a pull) not only would there be the little text appearing saying how much your score changed and "Good Clash Clearing!" or whatever, but also, the most strongly affected segments of the protein would be highlighted in some way indicating whether their score was improved or worsened, in which of the four areas of packing, clashing, bonding, and solvation (hydrophobics). Like if two sidechains started to clash because of your move, a sphere would show up around each of them like force fields, and you could see the fields intersecting so you know they're too close. There would be a different visualization for each of the four areas, so you'd know which was affected.

However, this still isn't a depiction of forces. Hmm. Do you think you could give an example of what your suggestion might look like?

I'm not sure if this is even possible with the way Foldit evaluates proteins. Foldit doesn't score based on forces - the way it scores is by adding up a bunch of terms like attraction (less voids = better attraction score), repulsion (less clashes = better repulsion score), solvation (hydrophobics in, hydrophobics out), bonding and statistically derived measures of backbone and sidechain naturalness. The forces are never explicitly considered or defined.

But then again, if they help people fold better, adding in a depiction of forces would be great. I'm just not sure how it would know what the forces are. Do you have any ideas?

axcho's picture
User offline. Last seen 9 years 38 weeks ago. Offline
Joined: 12/18/2007
Groups: None
Actually, maybe it is possible...

I just talked to Seth about this. He says that the Foldit scoring function actually does keep track of score derivatives, not just magnitudes. So we could find the direction that each residue wants to go according each area (packing, clashing, bonding, solvation) and display that somehow.

Is that more like what you're looking for?

Joined: 12/14/2008
Sounds interesting

I never really understood how the score for a single piece is calculated. If you use note mode and look at the score, if you change something, a single score more often than not goes the other way as you have thought it would.
How do you think it should look? I'm, not sure if we talk about the same.


Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, Boehringer Ingelheim, RosettaCommons