Residue IE Score: Monitors that all PHE, TYR, and TRP residues are scoring well.
Core Existence: Ensures that at least 30% of the residues are buried in the core of your design.
Secondary Structure: Checks that no more than 50% of residues are in helices; penalties are incurred if more than 50% of residues form helices.
Secondary Structure Design: Penalizes all CYS residues. Penalizes GLY, ALA residues in sheets; penalizes GLY, ALA, SER, THR in helices.
Residue Count: Penalizes extra residues inserted beyond the starting 65. Players may use up to 100 residues in total.
Ideal Loops: Penalizes loops that do not conform to preferred ABEGO patterns. Use Auto Structures to see which residues count as "loops."
I'd have been curious to see the result when loading a former puzzle
My assumption on Wiggle behavior would be that enabling or disabling "Residue Count" should have no bearing on Wiggle behavior. The assumption being that Counting the number of residues *should* be innocuous. But that is not observed behavior.
After achieving a somewhat stable structure and then Mutating, if RC is Enabled and 100% CI, Wiggle does not really do much, if anything. Disabling and 100% CI does, and then re-enabling improves the score.
Not sure if this was intended behavior. There are ways around this, of course. But it requires babysitting and not all of us are retired :)
how many iterations did u run wiggle for? I noticed that nothing happens for ages - then it does. But I still have this wretched
'game dynamic' - so most stuff is screwed anyway.
Heya Frood - I typically run everything for a minimum of 20 iterations.
23 points each may be a bit high for a penalty on added segments. Of 42 alpha-beta monomer design puzzles since the rama map came out, the average base points per residue (calculated as (score-8810)/residuecount) for the top 10 solos on each puzzle is 20.63. The number of solo solutions to those 42 puzzles having base points per residue >= 23 is zero.
It's more appropriate to look at the marginal value (see https://en.wikipedia.org/wiki/Marginal_value) instead of average points per residue. The marginal value is defined as how many points on average are gained from adding 1 segment to your solution. That value might be closer to 23.
This can be explained by the fact that the first and last segments generally score less than the rest of the segments.
We arrived at this value using a simple linear regression of scores from recent puzzles; this indeed yields the marginal value, rather than the residue average.
The fit is a little noisy, to be sure (and if we cared to, we could probably be more sophisticated with our approach), but the trend was surprisingly linear. In addition to tokens's observation about terminal residues, I think that a surface residue (especially in a loop) will also tend to score more poorly than a well-packed residue in the core. For this reason, I would have expected scores to increase quadratically with residue number, as the volume of a sphere increases with respect to surface area. However, this effect seems to be negligible, at least within the size range of a typical Foldit design.
Not a huge issue, but I just felt it worth mentioning that the Puzzle Info wording seemed rather misleading.
- "Variable Length Monomer Design", which generally on these they state in the puzzle name their base length. In this instance I feel it should've been stated as 65. Since it wasn't, it shifted my thinking on what I read in the description...
- Which stated that the residue count was tweaked, and now lengths beyond 65, at a max of 100, would score equally. As such, based on being a "Variable Length" puzzle, AND lacking a specific number in the puzzle name, indicated that you may run up to 100 segments without receiving a penalty from the Residue Count filter. Which is why...
- It was a bummer to see the Residue Count filter saying:
Residue count: lost points (-805)100 residues found. (35 over limit of 65)
To me, everything seemed to indicate that "Hey, you're able to make a protein with lengths up to 100 and have just as fair of a scoring chance against those with a length of 65!" However, the reality is that sure you can have a length of 100 as you can in any other design puzzle, but you're going to be penalized for it... same as every other design puzzle. :P
Just bummed me out was all, as I had built a new computer at the same time that the last 100 segment design puzzle came out, and due to some teething issues with Windows 10 and AMD's brand new processor design, I was unable to play it. Which seeing that I could (so I thought) run up to 100 on 1361 made me happy, till I saw that huge point deduction. *tear*
Anyways, just wanted to share that!
I think your initial impressions were correct. The filter has been re-calibrated to level the playing field between large and small designs.
The main difficulty with variable length puzzles is that every residue contributes positively to the score. The simplest way to increase your score in a puzzle is to add a residue; bigger proteins score better in Foldit.
If you compare the top-scoring designs from a 65 residue puzzle and those from a 100 residue puzzle, you'll see that the scores are much greater for the latter (by about 800 points). That is not because the designs are better; the designs simply have more residues.
The new Residue Count filter is tempered so that the penalty is only enough to offset this effect (about 23 points per residue). Please, try not to let the penalty discourage you! Use as many residues as you like!