Score is the attribute that we're all trying to maximize. Foldit uses the same algorithm as Rosetta for scoring. Information about the scoring algorithm can be found at [The Baker Lab ]
Essentially the score is a sum of the scores of each segment, plus an arbitrary 8000.
The score for each segment is a combination of scoring for these parts:
- Clashing ( The infrequency of 2 electrons trying to occupy the same energy state )
- Packing ( How well-hidden the backbone is, compared to its environment )
- Hiding (How well-hidden Hydrophilic / Hydrophobic sidechain are, appropriate to designated behavior)
- Bonding ( Whether the sidechain has achieved its optimal bonding angle )
- Backbone ( Correct peptide linkages, allowing the sidechain to achieve its lowest energy )
- Sidechain ( Correct sidechain / sidechain snap position )
- Reference ( Reference energy for each amino acid; Based on AA, See also get segment score part )
- Disulfides ( Correct disulfide couplings )
The reference score is a fixed number based on the Amino Acid type. The other score is for minor terms not included in any of the other scores.
You can find the scores for each of these, without fractions, by hovering the mouse over a segment and pressing the tab key. This will bring up a dialog showing the segment score. Scores with a value of zero are not shown.
Points = Max(1, RoundUp( ( 1 - (Rank - 1)/(NumPlayers - 1) )^5 * X))
(before july 2012, the exponent was 7 and for a while to 6 )
This formula gives the number of points that each player will earn at the end of a puzzle, based on his/her/their rank when the puzzle closes. These rank points are distinct from and not directly related to Score points.
see of bottom of page Punkte in german wiki for a open office sheet to calculate
Clashing is related to atom distance. If atoms are too close, their electron clouds would be repelling each other, while nothing bad happens if atoms are far apart (they just don't react). Note that this applies to both the backbone and side chains-- so, if you see a segment with a bad clashing score but can't figure out why, consider showing the side chain and/or switching to a full-atom viewing mode to see what's clashing.
How to improve: Shake and/or wiggle, and rebuild if necessary.
Packing refers to the extent that atoms are surrounded by other atoms. This is related to van der Waals interactions (see http://en.wikipedia.org/wiki/Van_der_Waals_force ).
How to improve: Manually flip large side chains or use bands to bring pieces closer together. Again, rebuild if necessary.
Hiding is related to the hydrophobicity of the side chain. In general, orange side chains need to point inward, while blue side chains need to point outward. The general trend is roughly consistent with the size of the side chain, so you've better bury a tryptophan and leave an arginine on the outside (as a general rule-- there are exceptions).
How to improve: Manual side chain tweaks, bands, or rebuilds.
Backbone: Refers to the configuration of backbone atoms. It is directed affected by dihedral angles (see http://en.wikipedia.org/wiki/Dihedral_angle ).
How to improve: Rebuild, in a full-atom viewing mode if higher precision is desired. Note that you may have to use bands to guide the direction of the rebuild and/or set the segment to loops to increase the range of choices for the rebuild tool.
Side chain: The score is related to the statistical distribution of side chain configurations based on the backbone configuration. A more negative score means that the configuration is rarer in nature, and thus less likely to be correct.
How to improve: Shake and manual side chain tweaks. Rebuilding the backbone can also be considered, as sometimes you simply can't fix the side chain due to an incorrect backbone configuration.
Reference: Apparently, it's the "reference energy" of each amino acid. This is perhaps the least intuitive part of the score function; its main use is for design puzzles as a "normalization" factor, to correct for the bias for larger side chains (which are not necessarily more stable in nature)
How to improve: You can't.