Coronavirus Binder Metrics Sandbox

Closed since over 5 years ago

Summary

• Created: October 21, 2020
• Expires: October 29, 2020 at 23:00 UTC
• Max points: 100

Description

This non-competitive sandbox puzzle features a new kind of Objective that we're calling Metrics. Metrics are slow to compute, so they are calculated in the background while you're folding your protein. That means that sometimes your score will be greyed out while the metrics are computing, but if you wait just a second then they will finish and your score will regain color. See the blog for more info about Metrics in Foldit. This is a multi-start puzzle with binders for the Coronavirus spike. You can start the puzzle with (1) a fragment of the natural ACE2 receptor, (2) the designed binder LCB1, or (3) an extended chain so you can design your own. Cycle through the different starting structures by resetting the puzzle. The puzzle Objectives include three Metrics (DDG, SASA, SC), so that you can practice designing protein binders using these binder metrics. This puzzle is non-competitive and will award no points.

Top groups

• 1. Go Science 100 pts. 224,931
• 2. L'Alliance Francophone 71 pts. 223,107
• 3. Anthropic Dreams 49 pts. 221,395
• 4. Contenders 33 pts. 220,779
• 5. Void Crushers 22 pts. 218,603
• 6. Gargleblasters 14 pts. 217,448
• 7. incognito group 8 pts. 217,316
• 8. MAB_STEM 5 pts. 217,217
• 9. FoldIt@Netherlands 3 pts. 217,217
• 10. Hold My Beer 2 pts. 217,217

Top soloists

• 21. Bletchley Park 29 pts. 220,779
• 22. Vinara 27 pts. 219,399
• 23. Mohoernchen 25 pts. 218,987
• 24. spdenne 23 pts. 218,603
• 25. twoodsOKC 22 pts. 218,504
• 26. JEAUSSIE7 20 pts. 218,488
• 27. GuR0 19 pts. 217,918
• 28. Lapinlexandre63 17 pts. 217,483
• 29. Formula350 16 pts. 217,448
• 30. pauldunn 15 pts. 217,419

Comments

[BootsMcGraw]

I get it: metrics are important, if we want to replicate what happens in real life.

I can't tolerate having to wait several seconds to see how every move of one angstrom affects DDG, SASA, and SC. Can you not make it so that these are not calculated until such time as we want them calculated? I tried disabling them in the objectives window, but they re-enable themselves without my assent.

In the early game, I would think most of us aren't concerned about these advanced metrics, focusing on the core and BUNS and ideal loops. Why are these forced upon us from square one?

And BTW, scripts that have never before crashed on me crash with these metrics enabled.

[APPAAP]

Why is it that SASA contributes most of the score?

[milkshake]

BootsMcGraw, could you please elaborate on the metrics being enabled without your clicking the enabled checkbox? When do you observe this behavior? The enabled checkbox is intended to allow the user to disable calculation of the metric. Regular folding should also be possible while the metric is calculating, regardless of whether it is enabled or not, so if something is causing you to have to wait for the metrics to finish calculating, we would like a report on that too! Thank you!

[APPAAP]

To foldit team.
I think something go wrong..
For example if I have DDG<=0 (a negative or zero) and achieve SASA high big enough i.e 3000 A I will get 3000x100 =300000 points. Although my solution may it will be rubish because BUNS may be not be good and hydrogen bonding also too bad at the binding site.
I summurize what I noticed
Max points for SC 1 X 100 =100 points (restricted to 0.6 60points)
DDG must be negative say -40 Kcal/mol x 100 =4000 points
(if DDG>0 reduces points from score dramatically)
SASA a 5000 A x 100 = 500000 points achieved???.
Therefore multiplying x 100 DDG, SASA and SC is not a good idea.

Another points is that the metrics are good for evaluation of the real thing but if one wants to use automation like lua scripts the rate of combinations per second (operations i.e mutations etc.) decreases dramatically!! Thus we restricted to manual use!!.

[BootsMcGraw]

I'd like to retract my statement. I wasn't understanding what it was I was seeing.

It does appear that disabling and re-enabling the metric works properly.

[BootsMcGraw]

The game dev's just want to give us a taste of how the new metrics will affect game play. I do not expect them to use the current scoring functions for these metrics in the future.

SASA is important for designing working proteins, but I doubt that it's twenty times as important as the energy (regular game score) function and five thousand times as important as SC. DDG is important, but I don't know how important, relative to the other metrics or the energy function.

[APPAAP]

Thank you BootsMcGraw.
The above clarifies a little bit the situation.
Another point is what we mean by core existance. Is it distance between helix to helixes (hydrophobic site) or sheets (just to make sure that the protein has more probability of true folding) and/or distance between hydrophobic's binding site with the designed protein? I ask this because if you design a good SASA and SC may not have a good core existance especially when you have to cover a big binding site and you are limited by segments as far I understand it by now.

[bkoep]

Core existence is measured by the "number of neighbors" that a segment has. A segment buried in the core of a protein has many more neighbors than a segment on the protein surface. By counting the neighbors of every segment, we measure how many segments are in the protein core and how many are on the protein surface. Good folding typically requires something like 25% of segments in the core.

You're right that it is hard to design a protein with good Core Existence and good SASA. That's one of the reasons this problem is difficult, and why we think Foldit players can help!

[milkshake]

Got it, thanks, let me know if you run into any more weirdness! It's a new feature so there's bound to be bugs.