Developer Chat

@beta_helix Ok everyone... We're about to start the Scientist Chat with 3 scientists today! 13:28
phoenixsomeone whom? 13:29
@beta_helix You all know Dr_Lemming, but I want to welcome Javier & yrl 13:29
mottiger hi DrL, Javier and yrl 13:29
@yrl hi 13:30
phoenixsomeone hey. 13:30
@beta_helix They are Symmetry Specialists and want to talk to you all about your symmetry results... but first, I hope all our Foldit players on the East Coast and their families are all safe. 13:30
@Dr_Lemming Hi all.    Symmetry time! 13:30
mottiger have you already gone through the data? can you present some interesting structures? 13:31
@Dr_Lemming @ mottiger: Yes to the first, maybe to the second.  I can definitely describe the general themes... 13:32
mottiger or is the presenting not so good, because other groups could 'pick up' the ideas? 13:32
@Dr_Lemming The structures so far have been very... interesting.  We're not particularly worried about other groups picking up the ideas... but not because they aren't interesting!  They're very interesting. 13:32
@Dr_Lemming And we'd love to share some tips and guidelines with you today. 13:33
@Dr_Lemming ...though, for reference, there will be information up soon in a blog post and/or on the wiki.  (A "how to guide" for building symmetric stuff.  As we'll discuss here, it's a bit different than standard protein folding.) 13:34
mottiger tips and guidelines always welcome, but sometimes it is super hard to follow these... like coordinating to the metal with a certain group of AAs 13:34
spvincent I saw screenshots of very imaginative designs. Maybe too imaginative 13:34
mottiger i loved the cylinder made out of sheets 13:35
@Dr_Lemming Indeed; and that's why we're trying to make as much as possible based solely on the energies, so that there's as much in-game feedback as possible.  But we've found that we can't cover everything... 13:35
@beta_helix @spvincent: spot on... that is what we want to talk about 13:35
spvincent Perhaps you could give visual examples of the sort of thing nature comes up with. Not so much for copying but rather as a source of inspriration. 13:36
@beta_helix that is a great suggestion, spvincent. 13:37
@JavierTheUnfolded <@yrl > and me have been looking at the structures that you guys have been producing looking for designs that we think might work when we express them in bacteria 13:38
@JavierTheUnfolded And a lot of your designs look really cool! 13:38
@beta_helix We will make sure to add positive examples to the blog/wiki post that Dr_Lemming, Javier, and yrl are writing up. 13:38
@JavierTheUnfolded But we are worried about their stability 13:38
@Dr_Lemming @spvincent: Yeah, we can point to some examples.  So for starters, there's a great example of a 27-amino acid trimer that works really well and is natural; it's called Foldon.   13:39
@yrl they look cool as a whole complex but not as a monomer 13:39
mottiger cool = promising or just looks cool 13:39
mimi2 it was fun to make pretty shapes but hard to find ones that score well 13:39
@yrl just looks cool but we don't think they would be stable enough 13:40
spvincent Can't click on or even read that link I'm afraid. 13:40
mimi2 works if you are on irc 13:40
marie_s http://tinyurl.com/ccytoal 13:40
spvincent tx marie 13:40
@beta_helix thanks marie, you beat me to it 13:40
@Dr_Lemming Well... we have tried making two of these designs (still just a tiny sample, I realize.)  No luck... in fact, not even soluble.       @spvincent: sorry!  Working from Mibbit here so...  oh wow, thanks marie_s, that was very fast. 13:40
@beta_helix (we'll make sure to post that on the foldit site/wiki) 13:40
@JavierTheUnfolded Looking cool, some of the sequences need more cooking. 13:40
@Dr_Lemming One of the big themes we'd like to hit on here is the importance of monomer stability. 13:41
mottiger isn't that something rosetta could do? 13:41
@Dr_Lemming Foldon (the 27 residue trimer I linked to) is a good example, because it has an independently folded beta-hairpin as the monomer.  (Monomer = independent subunit; just one chain alone.) 13:42
@yrl individual subunit has to be stable enough so that the whole complex could form 13:42
@Dr_Lemming Essentially, these symmetric things have to fold at least a little bit on their own, so they're mostly the right shape when they come together and dock to form the trimer or dimer or whatever symmetric complex they form. 13:42
@yrl unstable monomer would be degraded during synthesize and we would not be able to get the complex 13:42
@JavierTheUnfolded Living cells have mechanisms of detecting unfolded proteins so if your protein complex is made of disordered loops it is most probably going to be degraded 13:43
@Dr_Lemming And we haven't really emphasized that enough... plus, it's tricky to provide in-game feedback using the energy scores, though we're trying.  (Future puzzles might have scores weighted partly for the monomer or partly for the full complex.) 13:43
@Dr_Lemming @mottiger: do you mean the cooking of the sequences? 13:44
mottiger yeah, brute force the sequence with the power of so many computers 13:44
AndrewBartlett001 is yellow hydrophobic? 13:45
AndrewBartlett001 Philic* 13:45
@JavierTheUnfolded Yes we can improve the sequences with Rosetta but we think human intuition work better. We have add a new code to the game to improve the design that we hope will help fix this issue. 13:46
Madde http://foldit.wikia.com/wiki/Puzzle_644 13:46
Madde so the first example there^^ would be too complex as a monomer? 13:46
@beta_helix @AndrewBartlett001 are you asking a general Foldit question? If so, please use the Puzzle Chat rather than Global Chat for the next 15 minutes... we're in the middle of a chat with Scientists. Thanks! 13:46
@Dr_Lemming @Madde: yes. 13:47
@JavierTheUnfolded Yes! 13:47
@Dr_Lemming That is, in fact, a great example. 13:47
@Dr_Lemming the monomer on its own would be... well, it has nothing holding it together on its own, sadly.   HOWEVER, here's the exception: 13:48
@Dr_Lemming In this case, you have a covalent hub attaching the chains together. 13:48
@Dr_Lemming In this specific case (covalent attachment; amino acid chains "glued" permanently together) you don't have to worry at all about the monomer being stable on its own, because it will never be on its own. 13:48
@Dr_Lemming So, if this puzzle didn't have a hub holding the chains together, it would be a great example of something that wouldn't work.  HOWEVER, this particular case (the hydrogenase puzzle) works great because the chains are glued together. 13:50
@beta_helix first image in this one, for example: 13:50
@beta_helix http://foldit.wikia.com/wiki/Puzzle_638 13:50
@yrl this is a perfect example of cool-looking complex but unstable monomer 13:50
mottiger how is your approach to synthesize these complexes? do you synthesize the monomers and then attach it to the haem or do you grow the protein from the haem 13:50
@Dr_Lemming We'll make sure to let you know in the puzzle descriptions which symmetric design puzzles are "glued together" (and can have wacky monomers) and which are not (and have to fold on their own first before they find each other). 13:51
@yrl if the monomers are stable then there are experiments we can do to determine if the designed complex work or not 13:51
@Dr_Lemming @mottiger: we synthesize the peptide then attach to the porphyrin, using an excess of peptide.  (Admittedly somewhat inefficient, but it works.) 13:51
@yrl complex is going to form automatically in solution 13:51
@JavierTheUnfolded The problem of this structures is that in order for them to work the monomers have to find each other in space and interact, we you are depending on several monomers to need to come together to fold you have a slow process where the protein is going to be unfolded for a long time and bacterias don't like that. 13:52
@yrl if monomer is not stable, we'll only get a bunch of protein crash out of solution -> insoluble 13:53
ComputerMage guys, will you implement AA electrostatic charges attraction - repulsion? 13:53
ComputerMage That how it works in the real proteins when they fold. 13:53
@JavierTheUnfolded Or nothing because the monomer gets degraded. 13:53
spvincent CAn you give any guidelines for making structures that are likely to be stable? 13:54
@JavierTheUnfolded Yes! 13:54
spvincent I'm all ears 13:54
@beta_helix @spvincent They have actually written up a very long and detailed summary about this. 13:55
@Dr_Lemming @spvincent: yep.  It will be in the wiki/blog, but as a preview: aside from keeping the score good, also: keep independent chains well-folded, make sure there's a good deal of secondary structure, and... 13:55
@beta_helix We were going to post it as a blogpost, but I think it's so detailed that we'll put it in the wiki, with a shorter blogpost pointing to it. 13:55
@beta_helix We'll try to have that posted by the end of this week. 13:56
spvincent tx: look forward to reading it 13:56
@JavierTheUnfolded If you follow the guidelines your designs have better chances to pass our filters and get ordered, express and purified. 13:56
@Dr_Lemming a nice rule for folding rates is to focus on local structure.  Protein chains have to fold together mostly sequentially, so nearby structures assemble first and "build out" from there.  It's really tough to simulate this in Rosetta/Foldit, and it doesn't always matter anyway (since a protein chain will usually find the best energy even if it takes a long time 13:57
@beta_helix @ComputerMage unfortunately that slows things down considerably. 13:57
@beta_helix there have been many attempts to add electrostatics to Rosetta... 13:57
@Dr_Lemming but it's still something to keep in mind: interactions between opposite ends of an amino acid chain are "riskier", since there's a LOT of flexibility between the ends of a long chain, but amino acids that are close to each other on the chain can find each other pretty easy. 13:58
ComputerMage I will try to implement it on pyrosetta code. Will see if it fast enough. Just need more time off the main job. 13:58
@JavierTheUnfolded We have ways of testing the stability of the monomer but it will slow things down if we add them to foldit 13:59
ComputerMage yep. becuse of the chares the will fold into the SS and then ss will have they local charges that will atract each other. 13:59
ComputerMage they = their. 13:59
marie_s can you explain why you want to have symmetric proteins? 14:00
@JavierTheUnfolded Like for example Ddg delta delta G 14:00
ComputerMage Ok. Gotta run for the train. Would love to talk about AA charges more. 14:01
@beta_helix Thanks ComputerMage, don't miss your train!!! 14:01
@Dr_Lemming @marie_s: many reasons, but here are a few: they can be used as starting points to design functional materials (self-assembly), cages (origami!) and other really neat higher-order designs 14:01
ComputerMage but it's time to run. I am open for private mails in fold.it 14:01
mottiger and it is easier to synthesize, not? 14:01
@beta_helix Awesome recent paper published in Science showing a really cool example of using symmetry: 14:02
@beta_helix http://tinyurl.com/cq7uzks 14:02
@beta_helix you can download the pdf at the bottom of that page 14:02
@Dr_Lemming @mottiger: are symmetric proteins easier to synthesize?  Sort-of.  If the chains are short enough they can be made as peptides, but the overall complex is still pretty big... 14:02
mottiger nope, i wanted to say it is easier if you have 4 identical proteins on each porphyrin than 4 different ones 14:03
BletchleyParkirc What is self-assembly used for ? 14:03
marie_s in 638, the pauldunn one is not good? 14:04
@JavierTheUnfolded Cells use self assembly for everything! 14:05
@Dr_Lemming @mottiger: Oh - yeah!  Absolutely.  That makes things WAY easier.  Attaching 4 separate, different chains isn't just linearly-harder, it involves a whole different kind of synthesis, with selective protection and deprotection of the chain-attachment points and a whole bunch of other horribly tedious chemistry. 14:05
@JavierTheUnfolded That is how they build a lot of their internal machinery 14:05
@yrl @marie_s that one is not bad 14:05
@Dr_Lemming @marie_s: we like that one, actually. 14:06
@yrl but again, you can that one monomer is not really stable 14:06
pauldunn 14:06
@yrl but the whole complex indeed quite nice 14:07
@Dr_Lemming @yrl: the monomer isn't too bad, though - it has secondary structure helping to hold it together 14:07
BletchleyParkirc Could a protein be made to assist another protein in achieving symmertry by stabilizing it ? 14:08
@Dr_Lemming (and there are independently-folded hairpins and three-stranded sheets, so starting from a hairpin or three-stranded sheet can be a good way to go.) 14:08
@Dr_Lemming @BP: sort-of, yes.  One idea we have is two-component cages: essentially you could think of them as an assembly of trimers and dimers that "help" each other assemble.  These could be neat because they could use opposing charges. 14:09
@JavierTheUnfolded Here is a PDF showing a lot of self assembly components poster_quickref.pdf 14:09
@beta_helix http://tinyurl.com/bznyb94 14:10
@Dr_Lemming @BP: but if you mean chaperones, then that's a bit tricky for synthetic stuff.  Cells have some really interesting, complex chaperone systems that can help proteins fold in ways we can't always replicate outside of cells. 14:10
@beta_helix Any last questions for our Symmetry Specialist Scientists? 14:11
BletchleyParkirc are you using any of the tools devised in eterna ? 14:11
mottiger what about other point groups? 14:12
mottiger we had... C4 ? 14:13
@beta_helix @BP not yet... 14:13
BletchleyParkirc thank you 14:13
@Dr_Lemming @mottiger: we're focusing mostly on the simple stuff now (C2, C3, D2, C4) BUT we're also going to try some tetrahedral/icosahedral/octahedral symmetric puzzles.   Those are a bit hard though, both for us and for you.  (Dealing with 24 symmetric chains is CRAZY.) 14:13
@Dr_Lemming What we might do is start you with pre-made trimers or dimers that already work, and have you assemble those... but we're still considering. 14:14
mottiger dimer of dimer is D2 ? 14:14
@Dr_Lemming yep, dimer of dimers = D2 14:14
mottiger shouldn't this also be a promising approach, especially for the design of cages? 14:14
@Dr_Lemming yes! 14:14
@Dr_Lemming That's one of the things we're definitely working towards. 14:14
mottiger then i have a serious question now.... where is this D2 puzzle 14:14
@Dr_Lemming Er, unless you mean D2 specifically.  D2 is useful (we've done a D2 puzzle) but actually not on-path for a cage... 14:15
@Dr_Lemming since it's a sort-of "two-up, two-down" type of symmetry and it can't close up on itself like the C2, C3, C4, and C5 can.  (Tricky to explain without pictures...) 14:15
@Dr_Lemming The D2 puzzle was: puzzle #627. 14:16
@Dr_Lemming ...and 636 14:16
mottiger just in the porphyrine case... we had 4 up which was C4 14:16
@Dr_Lemming We do like D2; it's found in nature quite a lot.  (And pseudo-D2; for example, hemoglobin.) 14:16
@Dr_Lemming Ah, gotcha.  Yeah, we'll probably be releasing a D2 porphyrin puzzle; going to finish analyzing the C4 results first. 14:17
@beta_helix Ok everyone, thank you so much for joining our Scientist Chat today. 14:17
@beta_helix We'll post a transcript of this chat for those who missed it. 14:17
marie_s thanks 14:17
@JavierTheUnfolded Nice talking to you guys! 14:17
BletchleyParkirc thank you 14:17
spvincent tx all 14:17
boondog thank you 14:17
@Dr_Lemming   bye! 14:17
mimi2 thanks 14:17
mottiger stay safe and keep the pictures coming 14:18
@JavierTheUnfolded Keep an eye on the instructions an well express your designs! 14:18
@beta_helix you all stay safe as well! Bye! 14:18
@JavierTheUnfolded And don't forget about monomer stability! 14:18
mottiger point it out in the next puzzle description 14:18
@JavierTheUnfolded Bye! 14:18

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

Get Started: Download
  Windows    OSX    Linux  
Windows
(7/8/10)
OSX
(10.7 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
Recommend Foldit
User login
Soloists
Evolvers
Groups
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