1817: Coronavirus ORF6 Prediction

Closed since almost 6 years ago

Summary

• Created: March 26, 2020
• Expires: April 02, 2020 at 23:00 UTC
• Max points: 100

Description

Refold this coronavirus protein! This protein is encoded in the viral genome of SARS-CoV-2, in a region called ORF6, but the protein's structure is still unknown. Evidence suggests this protein inhibits the natural immune response, helping the virus survive and replicate. If we knew how this protein folds, we might be able to figure out exactly how it inhibits the immune system. The puzzle's starting structure shows SS predictions from PSIPRED, and hints which parts of the protein might fold into helices or sheets. Refold this protein to find high-scoring solutions, which will tell us how this protein is most likely to fold!



Sequence:

MFHLVDFQVTIAEILLIIMRTFKVSIWNLDYIINLIIKNLSKSLTENKYSQLDEEQPMEID

Top groups

• 1. Hold My Beer 100 pts. 9,477
• 2. Go Science 87 pts. 9,395
• 3. Gargleblasters 74 pts. 9,366
• 4. Anthropic Dreams 64 pts. 9,286
• 5. Contenders 54 pts. 9,276
• 6. Void Crushers 46 pts. 9,268
• 7. Beta Folders 38 pts. 9,160
• 8. BOINC@Poland 32 pts. 9,066
• 9. Marvin's bunch 27 pts. 9,062
• 10. SETI.Germany 22 pts. 8,965

Top evolvers

• 1. Deleted player pts. 9,477
• 2. Steven Pletsch 94 pts. 9,446
• 3. pente_player 87 pts. 9,395
• 4. TurtleByte 81 pts. 9,395
• 5. DodoBird 75 pts. 9,395
• 6. mirp 70 pts. 9,394
• 7. Bruno Kestemont 64 pts. 9,394
• 8. jeff101 60 pts. 9,394
• 9. edpalas 55 pts. 9,393
• 10. argyrw 51 pts. 9,393

Comments

[Steven Pletsch]

Can you post the psipred prediction? so we can see how it weights its choices please

[Serca]

Psipred prediction gives long helix just like puzzle start structure.
It seems that this protein is transmembrane, but when you wiggle it enough that helix breaks on two parts. So it is possible that both tails of the protein are on the same side of the cell.

[bkoep]

Conf: 9630646303999999999985389986999999999998102000242227657998789
Pred: CCCCHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHCCCCCCCCCCCCCCCCCCCCC
  AA: MFHLVDFQVTIAEILLIIMRTFKVSIWNLDYIINLIIKNLSKSLTENKYSQLDEEQPMEID
              10        20        30        40        50        60

[Steven Pletsch]

Thank You !

[jeff101]

In the above PSIPRED Predictions, what do the Conf numbers mean?
Is 0 the lowest confidence and 9 the highest?
Is 1 the lowest confidence and 0 (standing for 10) the highest?

[Serca]

I found a paper from 2007 that researches the functions of the ORF6 protein from a different SARS strain.
It states that its OFR6 is a membrane protein: "we demonstrate that SARS-COV ORF6 protein is localized to the endoplasmic reticulum (ER)/Golgi membrane in infected cells"

If the 1817 OFR6 protein is transmembrane, that means that exists is in non-aquatic enviroment, so hydrophilics should be buried and orange sidechains of this protein have to be exposed outside. In this case we just cannot get high Foldit score for best solutions, isn't it?

Paper:
Severe acute respiratory syndrome coronavirus ORF6 antagonizes STAT1 function by sequestering nuclear import factors on the rough endoplasmic reticulum/Golgi membrane.

ncbi.nlm.nih.gov

, ,

[jeff101]

Doesn't PSIPRED give for each segment the
probability that the segment is helix (H),
sheet (E), or loop (coil so C)? I think 
these 3 probabilities should total to 1
for each segment. Doesn't PSIPRED also 
give diagrams with 3 letters of different 
heights for each segment, so some segments 
have large H's but small E's and C's, while 
other segments have large C's but small H's 
and E's? Would someone please post a chart 
with the HEC probabilities or an image with 
the HEC letters given in different sizes?

Thanks!

[bkoep]

Cool! I hadn't seen this work!

Glancing at the paper just now, I don't see terribly strong evidence that ORF6 is itself embedded in the membrane. It does look like ORF6 associates with the ER, but that could also just mean that it binds another molecule at the ER. Also, just looking at the ORF6 sequence, I don't see anything that screams "transmembrane" (like a long stretch of hydrophobics).

That said, I have not studied this particular protein, so I can't claim any authority on this subject—there may be other evidence that ORF6 is membrane-embedded! If it is, you are correct that the default Foldit score is inappropriate for structure prediction. To predict the structure of a membrane-bound protein, we would want to adjust our score function to account for the non-aquatic environment.

[DoctorSockrates]

It might not be perfectly hydrophobic, but I did end up with a triple helix solution with two "membrane embedded" helices that I tried my hand at satisfying the polar atoms for. Could always be the case that it links up with another protein and offers quaternary structure to cover the rest.

[Steven Pletsch]

is psipred capable of predicting pi helices and 310 helices ? if so does it specify this in its output ? or is it simply ptedictin alpha/beta/loop