Developer Chat

beta_helix Hi everyone! 13:59
dflear hello 13:59
Marktoo hi, beta 13:59
spvincent hi beta 13:59
beta_helix We're about to start our Scientist Chat with DrLemming about the new Design puzzle (644) 13:59
firejuggler well I like it so far 14:00
beta_helix This is the blogpost that DrLemming wrote up about this puzzle (if you want to read up on it): 14:00
beta_helix http://fold.it/portal/node/993778 14:00
beta_helix So... welcome everyone, and welcome DrLemming! 14:01
spvincent hi dr lemming 14:01
CFC hey there 14:01
CarlKenner Hello Dr. 14:01
dr_lemming Hi everyone! Glad you like it so far, firejuggler. It's the first time we've tried using a covalent hub like this, so it's... different. 14:01
beta_helix (For the next 30-45 minutes, if new players can ask Intro Puzzle questions in the Puzzle Chat, that would be greatly appreciated. Thanks!) 14:02
spvincent how many puzzle rounds for this problem do you envisage? 14:02
firejuggler the un-solvable clash is a bit upsetting, but thats not really important. 14:03
dr_lemming I'm expecting about 2-4 (so 1-3 after this one) but I can't be sure; it depends on what everyone comes up with. But we'll definitely have at least one version where we incentivize putting interesting amino acid combinations near the center of the catalyst 14:03
spvincent We don't have to worry about that for now though? 14:04
dr_lemming re. the un-solvable clash: we're working on developing a more general way to set up these covalent ligand + peptide chain puzzles, and future puzzles SHOULD have that fixed... 14:04
dr_lemming Nah, no worries for now, just design, have fun. To be honest, this first puzzle is a bit of a "super-Beta test" (gamma test?) In that we know it's not 100% perfect, but it seems to work reasonably well for what we want. 14:04
marie_s do you want new bonds between the ligand and the portein part 14:04
dr_lemming For this puzzle, don't worry too much about new bonds between ligand & protein - we'll introduce that (with details as to what we need) in a follow-up puzzle. 14:05
dr_lemming Though if you're interested in a little foreshadowing: 14:05
dr_lemming we'd like to place hydrogen-bond donors near the central metal of this catalyst. More hydrogen-bond donors near the metal = better hydrogen-producing catalyst. 14:06
phallicies so right know just try to get a lof of bonding and sidechains? 14:06
dr_lemming Right now just try to get the lowest energy possible. 14:07
firejuggler So you want the ligand to be as 'free' as possible? ie : not 'inside' a box? 14:07
BletchleyPark irc what is that clash issue you mentioned ? 14:07
dr_lemming There isn't a lot of surface on the ligand to pack against the peptide, so peptide-peptide interactions are going to matter more here. 14:07
dr_lemming Feel free to keep the ligand free OR packed in a box. Packed might be a little better, both for Foldit scoring and for what we want to do with the catalyst, but it may not matter a lot. 14:08
TimovdL The puzzle starts with sheets, is that what you are interested in? 14:08
firejuggler (AA 1 and the ligand Bletchey) 14:08
spvincent Do we want to cage both sides of the haem-thing or will one side do? 14:08
dr_lemming Essentially, we're trying to make something enzyme-like, with a binding pocket for a ligand... but the ligand is the tiniest possible ligand (just a hydrogen atom!) So the space around the catalyst can be about nil, and it should still work. 14:09
dr_lemming The sheets are just there because we generally start symmetry puzzles with extended chains, to reduce biasing your results. Feel free to design any interesting structure where the peptide chains pack against each other. 14:09
dr_lemming Oh, and just one side will probably do, for caging. In fact, with the current symmetry we're enforcing, it might be really tricky to cage both sides. 14:10
CFC You say 'binding site', but do you know where on the ligand? to the side?, in the centre? 14:10
dr_lemming A follow-up puzzle will involve a different kind of symmetry; there's two ways that a four-chain peptide can be symmetric when all chains are equivalent: a "spiral" type symmetry and a "two up, two down" symmetry. 14:11
dr_lemming The latter will be much easier for packing around both sides of the catalyst. 14:11
CarlKenner I don't understand a tenth of what people are saying here... is that a problem? 14:11
MooMooMan attach at the centre of the peptide? 14:11
RajivPSN hello 14:11
dr_lemming Re. CFC's question: the binding site is at the very middle of the ligand; the center of the bullseye. 14:11
Lazarus-uk and what amino acids are more likely to attract hydrogen into the right area, acidic or? 14:12
beta_helix @CarlKenner not at all! It just a chance to have specific questions answered if you have them. Feel free to ask any more general questions as well! 14:12
firejuggler (CarlKenner : not a problem , you are a new player. experience will clarify) 14:12
dr_lemming Ah, good question Lazarus; for this first puzzle, don't worry about it, but for the future puzzle: both acidic and basic residues are good so long as they're polar, but the best are probably Asp, His, and Glu. We'll be very clear about what we want in the puzzle description. 14:13
Lazarus-uk ah, ok, thanks 14:13
CarlKenner So am I supposed to be bending this thing so the sheets get closer to the target in the middle? 14:14
mimi2 so we will get a "two up two down" puzzle option later? 14:14
dr_lemming (These are the sidechains that can donate protons most effectively at neutral or slightly acidic pH, so we'll probably set up the next Foldit puzzle to give a bonus for having them near the central metal of the ligand.) 14:14
firejuggler forget the sheets they are just a placeholder 14:14
BletchleyParkirc why not a four up and four down ? 14:14
firejuggler cause 8 -mer might be problematic? 14:15
dr_lemming @ mimi2: yep! Two-up/Two-down symmetry (technically called D2 symmetry) is coming in a future iteration of this puzzle. 14:15
CarlKenner Then what am I supposed to do with this puzzle? Or do you mean the sheets shouldn't be sheets? 14:15
dr_lemming @firejuggler & BletchleyParkirc: yes. 14:15
firejuggler the sheet might be sheet, or helix, or loop 14:16
firejuggler on this puzzle anyway 14:16
dr_lemming Yes, as in four-up four-down wouldn't work since we don't have a way to attach this ligand to a peptide 8 times. (Even 4 attachments is a bit tricky synthetically. I'll be making these myself when we complete the designs, and 8 peptide attachments is intimidating! Would be very low yields of 8mer; lots of 7mer, 6mer, etc.) 14:16
BletchleyParkirc why would that be problematic ? I can see how that would work out quite well. Can we have a BETA puzzle for that ? 14:16
BletchleyParkirc ah, ok 14:16
dr_lemming In terms of design, 4-up 4-down would be just fine, though, so good intuition there... honestly I'd prefer something like that if I thought I could make it. 14:17
dflear have confidence 14:17
MooMooMan with an 8 mer would you be able to use a special synthetic AA in the middle of the peptide to do an attachment? 14:17
dr_lemming An 8mer is technically possible, but no matter how I run the reaction (solution-phase, on-resin solid phase, etc.) it would be really tough - and we think 4 or even 2 attachments should work pretty well for this. (We may try a 2-attachment puzzle later, actually. With longer chains.) 14:18
CFC At this initial stage, are you looking for ligand bonds to stabilise, or a stabile protein to interact with the ligand? 14:19
TimovdL Isnt a 4 up enough also to see what a 4up 4 down would be like? 14:19
Brow42IRc You say you want the best score, but if the best score is just a long helix or 3-strand sheets, then how is that usefule? 14:19
firejuggler (I could see a 8 mer by extendingthe ligand fractally) 14:19
dr_lemming Just stable protein is fine. This first puzzle is just about getting nice structures built; you can treat it like the previous 4-chain symmetry puzzle, except with the ends of the peptide chains anchored in place. 14:20
dr_lemming And yeah, 4up 4down (D4 symmetry) would be a lot like half of a 4up (C4 symmetry) 14:20
firejuggler IMAGE: http://fold.it/portal/files/chatimg/irc_62014_1350940871.png 14:21
CarlKenner I'm going to ask a stupid question here, since I haven't played for months... but why is one side blue and the others not. 14:21
firejuggler blue hydrophile, orange hydro phobe 14:21
dr_lemming @Brow42IRc:  if the long helix or 3-strand sheets are compatible with the ligand, then that's fine!  Even if they're fairly "boring" looking structures, the main goal in this first puzzle is to find structures that work with the N-termini anchored where they are. 14:22
TimovdL Good to hear, I have something like that 14:22
marie_s Carl this is a symettry,you workon one and the other move the same way 14:22
CarlKenner @firejuggler, that's not what I meant. 14:23
CarlKenner @marie_s, thanks. 14:23
MooMooMan IMAGE: http://fold.it/portal/files/chatimg/irc_184172_1350941025.png 14:23
dr_lemming Ooh, neat. :) 14:24
MooMooMan rather open cavity near the ligand. 14:24
phallicies same with mine with the concavity near the ligand 14:24
CharlieFortsConscience IMAGE: http://fold.it/portal/files/chatimg/irc_102459_1350941087.png 14:24
phallicies IMAGE: http://fold.it/portal/files/chatimg/irc_428265_1350941099.png 14:25
CharlieFortsConscience IMAGE: http://fold.it/portal/files/chatimg/irc_102459_1350941106.png 14:25
dr_lemming Yup.  Oh, that reminds me: this molecule in the center can technically be used for different types of catalysts.  It's really neat, because it changes to become a different catalyst depending on what the central metal atom is. 14:25
Brow42IRc is the protien supposed allow direct access of the solvent to the reaction point, or we're just going to hop hydrogens along the sidechains? 14:25
CarlKenner Nice MooMoo! I can see why my structure isn't winning. Pear shaped with no clear secondary structures doesn't seem the way to go 14:25
dr_lemming SO, if these designs don't work well as hydrogenase catalysts, we'll try putting a different metal in them and use them for something else - CO2 reduction maybe, or ... well, porphyrins can do a lot of cool chemistry. :) 14:25
firejuggler everyone : nice shapes you got there 14:26
phallicies mine just looks like a bullet :/ lots of bonding but relatively low sidechains and sidechain bonding 14:26
dflear Is there any chance for this molecule to turn into a dimer or tetramer, like hemoglobin from myoglobin? Or would that not have any benefits? 14:26
MooMooMan make art, not biochemistry. 14:27
dr_lemming Brow42IRC: ultimately, for the hydrogenase catalyst puzzle specifically, it would be ideal to have no direct solvent access, but hop protons along sidechains.  Not sure if we'll penalize non-buried active site or not; I'll make sure it's clear in the puzzle description when the follow-up puzzle is posted. 14:27
firejuggler IMAGE: http://fold.it/portal/files/chatimg/irc_62014_1350941288.png 14:28
MooMooMan dr lemming - are you looking for more sidechains closer to the centre? 14:28
dr_lemming @dflear: probably no advantage to turn this into a hemoglobin-type system; hemoglobin is a really nuanced, bizarre system that has to bind oxygen in a really fine-tuned way; most heme-containing proteins don't bother putting hemes next to each other unless they need to shuttle electrons between the hemes, or bind stuff like oxygen (like hemoglobin does). 14:29
dflear very well :) 14:29
dr_lemming @MooMooMan: right now, I'm looking for whatever gives the lowest energies. :)  ...BUT if you want to recycle your current design into the follow-up puzzle(s), it would help to have some  part of the peptide close enough to the central metal of the catalyst, so you can put sidechains near it. 14:30
MooMooMan ok - this thing seems to like a cavity near the centre. Have to wait and see what constraints you can devise. :) 14:31
dr_lemming @CarlKenner: yeah, for design puzzles in general, secondary structure is the way to go. :)  (Loops are fine too, as long as the design isn't ALL LOOP, which is craaaazy. ;)  ) 14:32
BletchleyParkirc When will the first designs be made in the lab ? 14:32
dr_lemming @MooMooMan: cavity near the center is probably just fine; sidechains can be pretty long, so a medium-sized cavity might actually be ideal. 14:32
MooMooMan understood. 14:32
Brow42IRc I have a general question about symmetry design puzzles 14:33
TimovanderLaan IMAGE: http://fold.it/portal/files/chatimg/irc_308854_1350941589.png 14:33
dr_lemming @BletchleyParkirc: ahhhh heh, the eternal question.  Some of the initial designs: probably a month or two.  Something that works: longer. :P 14:33
TimovdL This gives a large hole in the centre, maybe a second protein can be put in there? 14:33
BletchleyParkirc thank you 14:34
Brow42IRc It seems really hard for me to believe that you can synthesize these designes, and still have the monomers fold into the complex we designed. Do they really form the designed complex? 14:34
Brow42IRc Especially given how small they are. 14:34
dr_lemming @TimovdL - maybe, but it would have to be reeeally small; more likely another ligand would fit in there. 14:35
dr_lemming @Brow42IRc: For normal tetramers: yes, a 30 residue peptide chain would have trouble folding up into something stable (though there are some cases that work).  But we have an advantage here: all 4 peptide chains are covalently bound to each other. 14:36
dr_lemming Binding peptide chains covalently gives us a HUGE entropy bonus; something that we can't really simulate directly in Foldit.  But it's fairly intuitive: it's easier to imagine 4 chains packing against each other than 4 chains all finding each other in solution at just the right angles and then packing just right. 14:37
dr_lemming If your question is whether the synthesis itself works, then I can confirm that it does. :)  (I've been attaching multiple peptide chains covalently and symmetrically to ligands for the past year or so, and I have most of the kinks out - though I admit that four attachments is a bit tricky and the synthetic yield will be on the low side.) 14:38
CFC history of foldit and lowest energies... is this a helical thing? 14:39
firejuggler (helical or ethical?) 14:39
MooMooMan theres the voice of experience :) 14:39
beta_helix yes... Rosetta just loves helices :-( 14:39
beta_helix and not cool ones like beta helices ;-) 14:40
spmm would you explain covalently in foldit terms please ? 14:40
dr_lemming By the way, we'll have other covalently-linked peptide chain puzzles in the future besides this one; some of them will be dimers (2 peptides) or trimers (3) instead of this tetramer. 14:40
firejuggler best shape is often helix spmm 14:40
spmm tx 14:40
dr_lemming @spmm: covalent attachment: effectively, chemically bonded so that the chains are permanently stuck together, no matter what. 14:41
firejuggler helix are often valued with lower energy wich lead to helix=> better score 14:41
beta_helix in Foldit terms: think of the super strong constraints we put on your cutpoints :-) 14:42
Susume2 stuck together means they don't have to find each other, is that the benfit? 14:42
dr_lemming @firejuggler: mostly true, though surprisingly, the highest-scoring symmetry designs have mostly been beta sheets.  (Though that might not stay true with the residueIE filter in place; we'll see.) 14:42
CFC ca c'est nas pas une helix... 14:43
MooMooMan @Susume - think of trying to find 3 people and stay facing each other in a mosh pit. 14:44
dr_lemming @Susume2: yep!  If peptide chains are covalently attached to each other, like these peptides are attached to each other through the ligand, then they don't have to bumble around in solution until they hit each other; they're already attached 14:44
dr_lemming like conjoined twins, you could say... 14:44
spmm ah so we cant create covalent bonds they come supplied? 14:45
dr_lemming And just being physically, covalently attached doesn't guarantee that they'll fold together how you want, but it helps a LOT. 14:45
firejuggler four head, one body... imagine the mess 14:45
dr_lemming @spmm: yep!  Sorry for the confusion: the covalent bonds are the bonds inherent to the peptides; those don't change as you play the game.  (Minor exception: sometimes you can make and break disulfides, which are covalent.) 14:46
spmm thank you - sorry I misunderstood -- 14:47
beta_helix Any last questions for DrLemming? 14:47
firejuggler what did you eat for breakfeast? 14:48
phallicies >.< 14:48
MooMooMan is that really your name? 14:48
beta_helix you don't have to answer that!!! :-P 14:48
CFC where  will stage 2 help you get to? 14:48
dr_lemming muesli, and ... 14:48
dr_lemming stage 2 (the 2nd or 3rd puzzle we do, with the constraints) will help us get a working hydrogenase catalyst. :)  ...so we can produce H2 easily from water and sunlight.  (By the way, this porphyrin molecule absorbs light, so it's a really dark brownish color.) 14:49
dr_lemming We might not get there from the first design, of course - especially since there are some tweaks left to do on the catalyst itself.  But we'll be a lot closer once we can design a peptide scaffold around this catalyst, and feed protons to the middle. :) 14:50
CFC ...so it's a real 'Science project' - 'Lets see what this does' 14:51
dr_lemming (Which is, of course, where all you awesome people come in.) 14:51
Susume2 it takes energy for the synthesis, i assume - how likely is it to "pay for itself" once it's working? 14:51
dr_lemming @Susume2: sunlight pays for it. :) 14:52
Susume2 :-) 14:52
beta_helix but we live in Seattle... we don't get sunlight! ;-) 14:52
TimovdL And scaling up would be a problem once you have something 14:52
dflear come to AZ,we get plenty ;P 14:52
dr_lemming Also, we can bind this to an electrode, so if we can't tweak its light absorbance to just the right spot, we can always feed it electrons to test it.  (That's what we're currently doing.  So we can actually see this working already - but we haven't got it to harvest light efficiently enough yet.) 14:52
phallicies yup plenty 14:53
spmm do you filter the sunlight or is out of the box ok? 14:53
dr_lemming @TimovdL: scaling up probably wouldn't be so bad.  Peptide can be super-crazy-mass-produced, and this particular porphyrin is a bit pricey but not super-expensive, especially if we were to scale up the synthesis... 14:53
MooMooMan how efficient do you expect proton hopping to be? I'm thinking of losses to the environment and what causes the energetic advantage to pass the proton to the centre. 14:54
dr_lemming but we'll probably need to alter the catalyst slightly for the final optimal design, which could make the end product a bit expensive.  We'll see. 14:54
CFC Thanks Dr Lemmimg... We're all Citizen Scientists here. Tip of the hat, and a big thank yo 14:54
dflear yeah thank you Dr Lemming :) 14:55
Marktoo yes, thanks, Dr L 14:55
CarlKenner thank you 14:55
MooMooMan Thank you for the explanations. 14:55
firejuggler Thanks Dr Lemming, may we see you again soon 14:55
dr_lemming @MooMooMan: proton hopping is quite common in natural catalysts; that's what we're hoping to mimic.  I tends to be quite effective - partly because a single proton doesn't jump all the way through, it's more like a chain of protons that gets "pushed", so the protons were already in place but they all just moved one spot forward. 14:55
beta_helix Thanks to all of you for joining our Scientist Chat! I promise DrL will be back here soon :-) 14:56
dr_lemming Thanks everyone for the good questions! 14:56
marie_s thanks 14:56
meatexplosion Hey Beta, will the chat be posted?  I seem to have missed it 14:56
beta_helix Yes... we'll try to have the chat up later today (we have to jet to a 3pm meeting now!). Take care everyone!!! 14:57
firejuggler soon (C) (and thats not  a Blizzard (C) soon) 14:58
beta_helix @firejuggler hopefully not that type of soon, and sooner than the last time DrL chatted with us (though that is mostly CASP10's fault) 15:00
beta_helix http://fold.it/portal/node/992469#t15:18 15:00
meatexplosion Thanks Beta! 15:01
beta_helix Ok everyone, last thing I want to add is this Forum post about using your Foldit Avatar in a presentation (if you didn't read it yet) Thanks again! 15:01
beta_helix http://fold.it/portal/node/993785 15:01

Generated by irclog2html.py 2.11.0 by Marius Gedminas
- find it at mg.pov.lt!

Download links:
  Windows    OSX    Linux  
Windows
(7/8/10)
OSX
(10.12 or later)
Linux
(64-bit)

Are you new to Foldit? Click here.

Are you a student? Click here.

Are you an educator? Click here.
Search
Only search fold.it
Other Games: Mozak
Recommend Foldit
User login
Topics
Top New Users
Sitemap

Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, RosettaCommons