Hello! I am Donuts554. Here I am going to share some of the strategies I use when playing Foldit.
- 1 Introduction
- 2 Amino-Alternation Strategy
- 3 Hydrophobic-Selective Strategy
- 4 A Note on Determining Secondary Structures
- 5 Recipe-Revolved Strategy
To me, overall, when I protein fold, I try to take into consideration each and every amino acid so that they all have their own true say in the protein's tertiary structure. Currently, I think and wonder that just generalizing parts of the structure of the protein into secondary structures and then aligning the secondary structures is unrealistic. It is because I don't think Amino Acid residues can instantly curl up into an organized secondary structure, including loops. However, I do use the Idealize SS tool in some of my strategies. Currently, I think that the string of amino acids gradually curls up into shape, overcoming sorts of energy barriers along the way.
This means I also follow the general rule, orange-colored hydrophobic amino acids are inside the protein, and blue-colored hydrophilic amino acids are outside the protein. Of course this is prevalent in most proteins, with a hydrophobic inside with a hydrophilic crust, except for membrane proteins, which have to be stuck in the membrane. In order for those membrane proteins to be stuck in the membrane, you have to have exposed hydrophobic Amino Acids sticking out. However, in the Foldit client as of 07/04/2020, exposed hydrophobic amino acid residues contribute to a low score in the protein. This is one of the imperfections in the Foldit scoring system, that have to be worked around.
My Amino-Alternation strategy is where the secondary structures are determined by the alternation pattern of the hydrophobic and hydrophilic amino acids from designated sections of the amino acid sequence, and then making the secondary structures themselves through the Idealize SS tool, and then aligned so that the hydrophobic sides of the secondary structures are on the inside, and so that the hydrophilic sides of the secondary structures are on the outside.
The reason I follow this current strategy for now is that I think that this proteins form secondary structures according to the alternation pattern of hydrophobic and hydrophilic amino acids on the amino acid sequence.
Determining the Secondary, by Examining the Primary
When deciding on what parts of the amino acid sequence are the "secondary structures", the alternation pattern for at least 3 consecutive residues of hydrophilic amino acids to hydrophobic amino acids for different sections of the primary structure is determined. The alternation pattern can be 1:1, like an alternating pattern of orange and blue in the "Hydro" view option, or 2:1 or 2:2, like an alternating pattern of 2 blue and one/two orange residues.
Then if the alternation pattern is 1:1, I determine the SS to be a sheet, because there is about an even amount of the hydrophobic amino acids on the inside, and the amount of the hydrophilic amino acids to the outside of the sheet.
If the alternation pattern is 2:1 or 2:2, the secondary structure is determined to be a helix, because each full cycle of the helix is about 3-4 residues long, and so that there should be about an even amount of hydrophobic amino acids on the inside, and the amount of the hydrophilic amino acids to the outside of the helix.
If the alternation pattern for at least 3 consecutive residues is not 1:1 or 2:1 or 2:2, then the secondary structure is determined to be a loop.
After determining the secondary structures of all of the distinguished sections that are outlined in Structure Mode, I then use the Idealize SS tool to form the respective secondary structures on those designated respective sections.
Forming the Tertiary, by Aligning the Secondary
To form the tertiary structure, I start with the closest secondary structure that is not loop towards the N-terminus, and then band the hydrophobic sides of each consecutive secondary structure that is not a loop together. After that, I use the Wiggle tool to bring those secondary structures together. I repeat this process for each pair of secondary structures, pair by pair.
After all of the secondary structures are pulled together after using Wiggle to bring together the last pair of secondary structures that are not loop, I then remove all of the bands and then I use the Wiggle tool on the Auto Wiggle Power setting. I do this to have the secondary structures in the protein aligned more "naturally" through the Wiggle tool's preference.
If the protein doesn't look that good with not much green and a bit too much orange, then basically the process is repeated, but this time the known secondary structures are worked with first, and then modified.
It seems that the helical structures fold in more unstable regions to fold into a more stable structure with more angled turns and folds, and the sheets are found in more stable regions with less angled turns and folds, but stuck to each other to satisfy their backbone hydrogen bonds.
When I Use this Strategy
So far as of now, I would use this strategy in the CASP-14 competition and in Prediction puzzles where the sequence can not be mutated but the structure can be changed.
The Hydrophobic-Selective Strategy is where I band each consecutive pair of hydrophobic amino acids, from the N-terminus to the C-terminus, and wiggle each time.
The reason I sometimes follow this strategy is because it seems to model how proteins fold, one by one, due to the presence of hydrophobic amino acids along the amino acid sequence.
Once I have banded all of the consecutive pairs of amino acids together, and I have wiggled them all, then I disable the bands, and then wiggle, to relieve some unideal loops or stress present in the segments. After I wiggle, then I shake, and wiggle again, but this time on "Medium" wiggle power, to make the loops more ideal.
I then save this solution for myself in the puzzle for reference, and afterwards, I remove all of the disabled bands, and I determine the secondary structure.
Determining the Secondary, by Angle Analysis
To determine the secondary structure, I examine and eyeball each of the peptide bond angles for each and every segment.
- If I see more distinct curves in the protein, then I assign those residues as a helix
- If I see more straightened angles in the protein, then I assign those residues as a sheet
- If the peptide bond angle is neither more of a distinct curve nor more of a straightened angle, then I just leave the residue as a loop
After assigning all of my residues with their own respective secondary structures in Structure mode, I then go into Pull mode and idealize all of the secondary structures that are present in the protein through the "Idealize SS" tool.
I then start to form the tertiary structure from these secondary structures. The strategy I use to do this is what is mentioned in the section, Forming the Tertiary, by Aligning the Secondary in my Amino-Alternation Strategy.
A Note on Determining Secondary Structures
In the Hydrophobic-Selective strategy, an alternate way of determining the secondary structures of the protein is by the 2D tool, at http://mw.concord.org/nextgen/#interactives/biology/intermolecular-attractions-bio/protein-folding-bio .
The reason I sometimes use this technique to determine the secondary structures is because I think the physics that is simulated in the 2D tool, including but not limited to the Brownian motion of the water, the attractive forces of polar &/or charged sidechains, and the hydrophobic forces of the orange hydrophobic amino acids, are realistic enough to determine the shape of small sections about 10 residues long of the protein accurately.
To determine the secondary structure of small parts of the protein using this method, if the protein in the 2D tool is moving by itself, I first press the pause button at the bottom of the 2D tool to stop the amino acid chain from moving.
Then, I change all of the residues to Lysine, so that the amino acid chain is more spread out due to Lysine's hydrophobicity. Then, on one of the endpoints, I change the residues on that endpoint so that the names of the amino acids at that endpoint are the names of the amino acids in the small part of the protein that I am going to determine the secondary structure of.
After assigning the names of the amino acids to the amino acid chain, I then press the play button at the bottom of the screen to make the protein move.
If the amino acid chain doesn't move much after a while, then I pause the chain, and then I determine the secondary structure based on the bond angles of the amino acids in the amino acid chain in the 2D tool that correspond to the small part of the protein. To do this, I use the same criteria which is mentioned in the section, Determining the Secondary, by Angle Analysis, in my Hydrophobic-Selective strategy.
My Recipe-Revolved Strategy is where Remix and Idealizing recipes are used to optimize the protein's shape.
The reason I use this strategy fairly often in Foldit puzzles as of now is because I think this strategy is fairly an easy and efficient way to optimize your protein automatically in a more quick and thorough manner. After I have done this strategy, I then save this solution, and I share my solution with the scientists if my solution seems interesting enough to me. This strategy is mainly an endgame one, as I usually then reset the puzzle and work on a different design after I have finished my strategy, so I can work on as many different varied designs I can in the given amount of time for each Foldit puzzle.
After I have hand-folded the tertiary structure and manually optimized it, I then run the recipe, Tvdl DRemixW 3.1.2., on the protein to idealize more of the loops, and to put the protein into a more stable shape. After 2 or 3 cycles, I then cancel the recipe, and I use the Wiggle action on the "Auto" and then "Medium" Wiggle Power to stabilize the protein. After the Wiggle action goes through ten iterations, I cancel it.
If the protein still looks fairly orange, especially at the loop sections, I then run the recipe, Rav3n_pl GAB BiS v2.0.2, to remove any more stress in the protein. After one to two, or possibly three generations, I then stop the recipe.
After I have removed as much of the stress I could, I then run the recipe, Quickfix 3.6, to idealize more of the loops and to stabilize the protein more. After the recipe finishes going through one or possibly two cycles, I cancel it.