1137: Two-component Abeta binder design

Closed since over 10 years ago

Summary

• Created: September 12, 2015
• Expires: September 27, 2015 at 23:00 UTC
• Max points: 100

Description

Note: The expiration date for this puzzle has been extended one week, to 09/27/2015 - 23:00.

This is another puzzle in which we ask you to design a protein that can bind the Abeta peptide, the aggregation of which causes Alzheimer's disease. The hope is that by designing something to bind it up, we'll be able to block its toxicity and halt Alzheimer's disease progression. This time, we're giving you eight different starting points for design, each with a different pattern of disulfide bonds, and each with two distinct halves (which we might ultimately mix-and-match). We'd like nice, well-folded designs with well-packed hydrophobic cores, lots of hydrogen bonds networks on the surface, and no voids. (You can turn on the display of voids in the core in the display options to help you eliminate these.)



The loops in the starting conformation are so-so, so you can keep these or redesign them. The Remix tool is available for your use in this puzzle. Regarding the helices and sheets, we'd like you to keep the overall topology of the starting structure, but there are no constraints, so you can move the backbone quite a bit in this puzzle.

Top groups

• 1. Gargleblasters 100 pts. 10,958
• 2. Beta Folders 84 pts. 10,938
• 3. Anthropic Dreams 70 pts. 10,937
• 4. Contenders 57 pts. 10,865
• 5. Go Science 47 pts. 10,807
• 6. Void Crushers 38 pts. 10,803
• 7. L'Alliance Francophone 30 pts. 10,714
• 8. Deleted group pts. 10,707
• 9. FoldIt@Netherlands 19 pts. 10,537
• 10. HMT heritage 15 pts. 10,511

Top soloists

• 1. MurloW 100 pts. 10,937
• 2. LociOiling 99 pts. 10,886
• 3. Blipperman 98 pts. 10,877
• 4. Deleted player pts. 10,874
• 5. actiasluna 95 pts. 10,868
• 6. Bletchley Park 94 pts. 10,851
• 7. BitSpawn 93 pts. 10,843
• 8. johnmitch 91 pts. 10,842
• 9. Vredeman 90 pts. 10,823
• 10. spvincent 89 pts. 10,821

Comments

[Susume]

Structure 8 is the only one where all of the disulfides have a leapfrog pattern - wouldn't the peptides be more stable if they were all like this?

The beta turn at 92-94 in all the structures has a very unlikely chirality, according to the Koga & Koga protein design paper. That suggests the odds are against it folding in that direction. Is there a reason you wanted the sheet strands in that particular position? Swapping the strands would allow the turn to stay short while fixing the chirality, but that would mess up the given disulfides. The only way to fix the chirality without moving the sheets is to lengthen the loop to 4 or 5 residues (counting 91 as part of the sheet, not the loop), which presumably will make it floppier (give it more ways to misfold).

[gitwut]

If possible, could you please extend this puzzle by several days? I've worked on five of the structures to get the fragment scores cleaned up and roughly mutated, but am not seeing any stand out scores among them yet.

Since I can't run all eight at once, it is unlikely that I'll have time to give them all a fair shot (only 5 days left).

[bkoep]

Sounds good. We'd definitely like players to have a fair shot at each of the eight starting structures, so the expiration date for this puzzle has been extended one week.

[Skippysk8s]

I would not have opened a puzzle closing when I can not work it. In the future, it would be better to have posted this as separate puzzles since it is so large and let each have it's own deadline. While I may be one of the few disappointed in the extension, I feel like you've now wasted my time

[v_mulligan]

Those are good observations. Regarding the first, the leapfrog disulfide pattern is probably more important in small peptides that get almost all of their stability from disulfides than in larger proteins that have nice hydrophobic cores. Many different patterns are observed in larger proteins. We've tried to select designs that have disulfides between parts of the protein that are far apart in linear sequence, whether or not they make a leapfrog pattern. (In smaller peptides, the only way to have all of the disulfides form between parts that are far apart in linear sequence is with a leapfrog pattern. In larger proteins, there are other patterns that can satisfy the linear distance criterion.)

Regarding the second, you've got a good eye – I might post another puzzle with the strands swapped. For now, lengthening the loop is fair game.

[gitwut]

Hi Skippy. I always try to put myself in other people's shoes but I'm not sure I understand your reasoning. If you started the puzzle knowing that you could only work up until the original deadline, how does extending the deadline cause any problem at all? Whatever work you planned to do in the original time frame can still be done–you'll just finish earlier than the rest of us. How has that wasted your time? Working more than one or even all eight structures is not a requirement.<p>
Apologies if I've totally missed something here–it wouldn't be the first time.

[inkycatz]

We're going to take some steps here to improve communication around puzzle extensions to alleviate frustration that may surround this sort of thing. Thanks for chiming in, even if you feel it may be an unpopular opinion!

[Skippysk8s]

Gitwut – I worked this puzzle suboptimally to meet the deadline, as I expect many others did. I have a small machine, so can really only run one client of this size at a time.
As a rule, I don't open anything where I will miss a lot of time to work a puzzle before close. Given my small machine, I don't feel like I have any chance to get a competitive solution for either science or points in that situation.
I know my time problem is unique regarding this deadline. But I still think it would have been better to post this as several puzzles. I think many of us would have done a better job from the start rather than scrambling. This felt like CASP – just sort of do your best and move on as there are too many to do a total work up of the puzzle. I didn't enjoy that experience and will probably take a pass on it next time around
So do solve it for me. And maybe share some points ;)
Skippy

[gitwut]

Thanks for the reply Skippy. I do remember your situation from previous feedbacks, chat, etc., though I don't keep up lately as I used to. I think you're a very good folder and I'm sure many at the top will rue the day you finally get a better computer. :D

There are many puzzles that I have to put in sub-optimal effort as well.

I'm very methodical in my approach to puzzles in general, but there are some, due to the protein size or complexity (filters, mutate) that force me to change strategies because there is not enough time to run the same pattern/number of scripts that I would on simpler puzzles. Even so, I am sometimes surprised by having better scoring solutions that I originally expected. For example, contact maps rarely give me more than minor clues so I work them like regular denovos. EDs are even stingier because I can't see squat in the "cloud". Yet in both cases, I usually score fairly well.

If you enjoy points/ranks as a secondary aspect of the game, remember that you won't get any points for puzzles you don't start. A puzzle you start and do the best you can on will probably yield some points (some > none).

I enjoy CASP, but I can see where being limited to one or two clients would be a real handicap with so many puzzles running simultaneously. I'm gungho when they start and relieved when they end.

I'm not smart enough to know if I am contributing to science or not. If I am, hopefully someone will tell me. Otherwise I'll never know.

I think I'm missing the competitive genes altogether. I do the best I can in the time given, hope for good results, and most importantly, don't worry too much about it.