Developer Chat

@inkycatz Welcome to our first chat for 2015! Thank you all for coming :) 15:00
@rav3n_pl *drums* 15:00
@inkycatz All chats are logged! If you miss something, or have to leave, check afterwards. Thanks to jflat's magic logging systems, those logs get posted pretty quickly. 15:00
@inkycatz Our next chat is in April, for those who really like to plan ahead. Also a big shout out today to Bruno K, for bringing a bunch of questions to our thread ahead of time. We love when that happens, so thanks! 15:00
@inkycatz (We appreciate everyone's questions but he really went above and beyond) 15:01
@inkycatz This is the end of the announcement portion of our chat and you should start paying attention again. :) 15:01
@inkycatz Today's agenda starts with a drug design update, then we'll kick it over to beta_helix who should be here by then, and then he and bkoep will dive into some of these great questions we have piling up. Or we'll just wing it after the drug design stuff. We're all cool. 15:01
bkoep Hi, everyone! 15:01
@inkycatz free_radical, you awake? 15:02
free_radical yeah 15:02
free_radical howdy :D 15:02
@inkycatz Take it away! :D 15:02
free_radical Hi guys, so I wanted to give you an update on the drug design implementation 15:02
free_radical we had a meeting two weeks ago to talk about the game 15:02
free_radical What we are working on now is to get the interface cleaned up 15:03
free_radical so that it is easier to use 15:03
free_radical we also would like to announce our first puzzle 15:03
free_radical We will be looking at HIV protease 15:03
free_radical Unfortunately I dont have any new videos 15:04
@inkycatz Aww :( 15:04
free_radical but you should look to the blog next week for the interface changes 15:04
@rav3n_pl *interface tip* (selection) get rid of animated show and go buttons, make them stationar and grey out when can not be used. 15:04
free_radical and hopefully in a month we will release the first puzzle! 15:04
free_radical @rav3n_pl we have implemented some of the greying out things you mentioned 15:05
free_radical for hiding chemistry that is non-existent 15:05
@rav3n_pl cool :) 15:05
@TimovdL As this will be a whole new kind of designing, we might need a few very simple puzzles to get the hang of it 15:06
free_radical @inkycatz, I was at a conference last week and didnt get any chance to create new videos :( 15:06
@inkycatz it's okay I just love videos :) 15:06
BletchleyParkirc Will we be given introduction puzzles ? 15:06
@rav3n_pl yea, some fast small ones to warm up and get new skills 15:06
free_radical @Timovdl we are planning on introduction puzzles 15:06
@TimovdL And will there be new LUA interfacing tools? 15:07
@rav3n_pl new? do we have some? *looks I`m missed something* 15:07
free_radical I do not have an answer for that, but I cant imagine why not. We are using the same modules fuond in Foldit 15:08
BletchleyParkirc for my understanding: this is an extension of the current foldit client, correct ? 15:08
free_radical correct 15:09
BletchleyParkirc thank you 15:09
@TimovdL Rav, as i understand we are designing non proteins, ligands 15:09
free_radical There will be a small update to the client when the ligand stuff comes around 15:09
@rav3n_pl uh, ok, I really missed something :D got it :) 15:09
@inkycatz Fortunately the regular series of blogs should catch you up! 15:10
@rav3n_pl so we will just see new type of puzzles, rught? 15:10
free_radical correct. you will see a new classification of puzzles, small-molecule design 15:10
free_radical we will start off slow for everyone to get the hang of things, then we will add more puzzles 15:10
MuckeMcFly sounds good 15:10
Brow42IRC two things that Lua needs to work with ligand bonding is 1) a way to find out if a bond exists (and a quick way to list existing bonds) and 2) a way to find out if an atom is a donor or acceptor (and a quick way to list them). This can be preprogrammed for AAs but not for ligands. 15:11
free_radical Just so everyone knows, we will be testing the designs 15:11
free_radical @Brow42IRC, hydrogen bond donors/acceptors are identified on the atom type level, which the ligand design stuff will use 15:12
free_radical Additionally, bonded atoms are identified too, I am not sure if those modules are in LUA, but I can check 15:12
@rav3n_pl So it will be like reverse-design puzzle. Locked part of protein and we play on ligand 15:12
free_radical @rav3n_pl, this is the general idea. Medicinal chemists generally lock the protein and design the ligand; however, it is up to you how you go about design 15:13
free_radical we will not lock the protein, but you can freeze it 15:13
@TimovdL And can you post in the blog an explanation of the design principles/tools? 15:13
free_radical when you design small molecules, protein flexibility is extremely important, albeit hard to model. Thats why medicinal chemists generally lock the protein 15:14
free_radical I would be more than happy to make a blog post for the new tools and principles 15:14
@inkycatz Sounds fantastic. 15:15
@rav3n_pl Score will be based on current price of used atoms... :D 15:15
@rav3n_pl cheaper - higher score 15:16
free_radical lol 15:16
jmbrownlee333 are there medicinal chemists that are part of this project to actually synthesize stuff ? 15:16
@TimovdL If so I ll add water 15:16
free_radical @rav3n_pl, we are actually going to start charging for foldit. We are looking at the freeium model 15:17
free_radical jk 15:17
BletchleyParkirc can you please elaborate on that ? 15:17
free_radical @jmbrownlee333, we do have medicinal chemists, organic chemists, and computational chemists all at our disposal 15:17
@rav3n_pl jmbrownlee333: I remmeber "fuel puzzles" which was actually tested in real 15:17
@MikeCassidytoo Homeopathic medicine is the best ;-) 15:18
free_radical @BletchleyParkirc, I was making a tounge in cheek comment about freemium 15:18
BletchleyParkirc ok 15:18
jeff101 will some of the Med Chem, Org Chem, and Comp Chem folks be at Eli Lilly? 15:18
@inkycatz Sorry, folks - text doesn't always spin those jokes as well as they sound in our heads. :) 15:19
free_radical We will try our best to synthesize as many of them as we can 15:19
free_radical @jeff101, no 15:19
free_radical This is a very important distinction that needs to be made 15:19
free_radical Eli Lilly pays for me to program and helps offset the cost of synthesizing the small molecules 15:20
free_radical however, the people who will synthesize and test the small molecules will be academic institutions 15:20
free_radical such as Vanderbilt, where I am at right now 15:20
free_radical by program, I mean write code in Foldit for drug design 15:21
@rav3n_pl looks like we all need change our CV at point "I do drugs" :D 15:21
Susume2 do the smalll molecules fall under the foldit commitment to make all discoveries by foldit players public? 15:21
vmulligan Hi -- sorry I'm late. 15:22
free_radical here is the program that pays my bills 15:22
free_radical 15:22
free_radical @Susume2, in my contract we specifically made an amendment for discoveries to be part of the rosetta commons, which is what Foldit falls under 15:23
Susume2 thx 15:23
jmbrownlee333 can you tell us more about scoring ligands. 15:23
free_radical sure 15:24
free_radical for scoring ligands, we will be using the same scoring function that Foldit uses, coined the Talaris2013 score function 15:24
@inkycatz (After this question we'll start diving into the pile of waiting questions, and I believe beta_helix is also on the way.) 15:24
free_radical as a side note, I was at the original Talaris2013 conference where we made the decisions about the score function 15:25
* rav3n_pl preparing drums 15:25
free_radical You will be given one score, just like in Foldit, but it is slightly different 15:25
free_radical We will be normalizing the score based on the number of atoms designed 15:25
free_radical this is because you can infinitely make a residue score better by just adding atoms 15:26
free_radical we will also fold some more things into the score function 15:26
@rav3n_pl first puzzles will be abused for sure :) 15:26
@inkycatz ...well I do have questions, but hey beta! hey vmulligan! we're glad you guys could make it. 15:27
free_radical these will be reported separately and include things like synthesizability, ADME properties 15:27
@TimovdL That is our job, finding holes in the scoring functions 15:27
@betahelix Hi everyone, sorry I'm late (getting home in the snow here too longer than I thought! :-) 15:27
free_radical @Timovdl, you are correct. finding anamolies in the score function help us out a lot 15:27
vmulligan Hi! :) 15:28
@inkycatz Thanks for the update, free_radical, can't wait to see that next blog in my inbox soon! :D 15:28
free_radical de nada :D 15:29
@inkycatz beta, I believe you had some stuff to chat about? :) 15:29
@betahelix I wanted to talk about Puzzle 1035: Incredibly Difficult 95-residue Electron Density Design Puzzle 15:29
@inkycatz (I know, you just get here and BAM right away I put you to work) 15:29
@betahelix The first thing I want to mention is how hard we thought this puzzle was going to be (hence the puzzle title!) 15:29
@betahelix We honestly believed that this was so hard, that we would have to provide you with some additional information (such as the location of the first residue, for example). At the last second we decided to see what you could do with no extra help... 15:30
@betahelix and as you Foldit players never cease to amaze us, in less than 19 hours some of you had already solved it! 15:30
@betahelix These are the final plots: 15:30
@betahelix 15:30
* rav3n_pl ba-dum-tsss 15:30
@rav3n_pl gtd=1 equals native? 15:31
@betahelix Can everyone see the GDT and RMSD plots? Yes 1.0 is perfect, 0.0 is not. 15:31
Brow42IRC I see 2 minima 15:31
@rav3n_pl ha! we won :D 15:32
@betahelix RMSD is the other way around: 0 is a perfect match 15:32
@rav3n_pl give us good ED and we solve everyting :) 15:32
@betahelix @brow42 you mean on the RMSD plot? 15:32
Brow42IRC yes, that's right, on RMSD, it's a little disconcerting. 15:33
BletchleyParkirc Hi Beta, what was the significance of this particular puzzle ? 15:33
@betahelix Anything under 1 Ang RMSD is really good. 15:33
@betahelix Great question, BP! 15:33
jeff101 GDT is supposed to be a better measure than rmsd 15:33
@betahelix You might have been wondering why we posted such a whacky puzzle, and how did we dream something like that up, right?! 15:34
@betahelix Turns out that we didn't... but some researchers at the University of Michigan ran a similar "puzzle" with an undergraduate class using Coot. 15:35
@betahelix They used the exact same protein that we gave you, along with the exact same density map... the only difference is that in their case the researchers provided the students with the exact location of the first Glycine. 15:36
@jflat06 also - there will always be local minima away from the native. that isn't necessarily a bad thing, unless the local minima scores better than the native, while also being far from it. 15:36
@betahelix I'm sure anyone who played this puzzle would have liked having that information (and this is why we thought it would be so tough for you without having that glycine to anchor your model in the density) 15:37
BletchleyParkirc It was just a better density model 15:37
BletchleyParkirc than usual 15:37
BletchleyParkirc I think that made a big difference 15:38
frood2IRC agreed BP 15:38
@betahelix What is exciting about this result is not just the fact you were able to recreate how the students did (given less information) but also that Foldit seems to be sufficient to solve this type of problem, whereas they used the crystallographic software: Coot. 15:38
jmbrownlee333 I keep wondering ,do foldit players beat automated software, like arp/Warp ? 15:39
@betahelix For anyone who is interested in reading their paper, it's freely accessible on their website: 15:39
@betahelix 15:39
@betahelix @jmbrownlee that is a great question and one that we would really like to answer very soon. 15:40
Brow42IRC I didn't do the puzzle, but I've been feeling that, since the premise of foldit is to rely on human visualization, then what foldit needs is some way for humans to adjust and smooth the density themselves, even if it doesn't preserve certain properties. Image processing has contrast, gamma, edge detection, blurring kernals, etc. ED currently has threshold, and pre-smoothing by the puzzle poster. 15:40
@betahelix One of the tricky aspects to test that is that we need blind cases (proteins that are about to be solved, or haven't been publicly released yet) 15:40
@betahelix Such cases are very difficult to find, which is why this particular puzzle was interesting... since there was no way to know the native since we didn't give you the sequence. 15:41
@betahelix We are working with the researchers at UMich to find some crystallographers that are willing to share their density with us (either for models they haven't solved yet, or haven't published) 15:42
@betahelix bkoep, did you want to add anything? (did I forget something!) 15:43
bkoep I'll add that Foldit players can expect to see plenty more electron density puzzles in the future 15:43
Susume2 how well did foldit players do on sequence identity? 15:43
bkoep Usually, with much poorer-quality density 15:43
@betahelix Great question, Susume! 15:43
jeff101 these electron densities come from crystals, so aren't they periodic with unit cells and things? how do you know if the density you give us is from one unit cell with one copy of the protein? is it possible the density you give us has part of one unit cell and part of another? 15:44
@betahelix One player that nailed the structure within 19 hours actually sent the correct sequence as a "Share with Scientist" 15:44
BletchleyParkirc Congratulations with the results :) 15:45
bkoep @jeff101, you're correct that the crystallographic density is periodic, but we can very easily determine the size of the unit cell 15:45
Alve 14% of UM students outperformed the published structure did foldit players also do that? 15:45
@betahelix But we still need to do more analysis of all the top solutions (now that the puzzle ended) w.r.t. sequence identity. 15:45
bkoep However, sometimes it's not so easy to tell how many copies of a protein are in a unit cell 15:45
@betahelix Congrats to all of you, BP :-D 15:45
Susume2 the score function preferred a different sequence in places than the ED indicated 15:46
@TimovdL I think I nailed half the sequence but ran out of time for the second half 15:46
@inkycatz Truly some impressive work by the community :) 15:46
@betahelix @Alve that is another interesting question, because I don't think % would be a fair comparison. The students in the class HAD to complete the assignment, but any new Foldit players that joined could have tried the puzzle (even though we made it "Advanced difficulty") 15:46
jeff101 and don't some crystals have copies of the protein in different orientations? 15:46
@rav3n_pl that should get ribbon :) 15:46
@rav3n_pl *that player 15:47
jmbrownlee333 I want that player to show me how they did it ;) I could see structures but not make foldit behave. 15:47
@betahelix Yes, I think another Black Belt Folding would be great... 15:47
jeff101 make a video of the steps that folder took? 15:47
jmbrownlee333 yep ! 15:47
@inkycatz I was just thinking that very thing. 15:47
@betahelix particularly a "Design into Density" session 15:47
jeff101 you have the data already to make such a video 15:48
@rav3n_pl do it! 15:48
@betahelix I want to make it clear that many foldererers nailed this one! 15:48
@rav3n_pl do many videos :D 15:48
@betahelix tracks makes that a bit trickier... but we have talked about implementing this many times. 15:49
jeff101 early foldit puzzles had videos, even with parts done by different players 15:49
@betahelix Any last questions about this? 15:49
BletchleyParkirc This result is great, but We have 10 minutes left, can we quickly go through the questions ? 15:49
@inkycatz We do have some great questions! Let's get to them. :) 15:50
@betahelix Yep, that's why I wanted to move on :-) 15:50
vmulligan 'Twould be interesting for those of us writing algoritms, too.  Emulating the human players is something worth doing. 15:50
@inkycatz Bruno has some questions regarding antigen/antibodies, and I'm just going to throw them out in order. 15:50
@inkycatz 1) What do you try to do, is it almost "exactly" what our body does (in 2-3 weeks ?) when trying to construct an antibody against an antigen? 15:50
@inkycatz (just holler done and I'll post the next one :) 15:51
bkoep The end goal is similar, in that the body wants to produce proteins (antibodies) that bind a particular target (antigen) 15:51
@betahelix (@vmulligan I completely agree and tried to manually do this for the M-PMV case... but my computer crashed with all of models I tried loading in! ) 15:51
bkoep But the body goes about it with a very different strategy 15:51
vmulligan Yes.  Here's an analogy: 15:51
vmulligan Let's say you want a comfortable chair.  You go to IKEA.  They have a zillion chairs that you can sit in, at least one of which will fit you well.  That's the body's strategy: make a zillion types of antibodies, all the time, and when something sticks to one, make more of that one. 15:52
vmulligan Our strategy is different.  You want a comfortable chair.  We carefully measure your backside, and build the chair for you. 15:52
bkoep *A self-folding chair 15:53
vmulligan Well, yes, that's the complicated part.  :P 15:53
@rav3n_pl but colour is wrong.... :F 15:53
Brow42IRC a self-assembling IKEA chair, that'll be worth millions. 15:53
vmulligan If only we could cheat and use Allen wrenche... :P 15:54
@betahelix (it would still be missing a bolt ;-)) 15:54
jeff101 use that special foam 15:54
BletchleyParkirc done 15:54
vmulligan *wrenches 15:54
@inkycatz haha ok chair talk aside :) 15:54
@inkycatz 2) Is there an idea how our immune system "recognizes" a potential target on an antigen? How does our body know what to synthesize? (we folders "see" a hole and we try to stiff something in it, but our body?) 15:54
@inkycatz (I'm not translating that into chairs but you may feel free to.) 15:54
vmulligan I think we kind of answered this one.  The body doesn't know how to design something, but it does know how to make a zillion variations on a design in the hope that one happens to have the right shape. 15:55
MuckeMcFly Evolution! :) 15:55
@rav3n_pl it is explained in "Il ├ętait une fois... la vie" - white cells call for HLA card ;] 15:55
Alve And can deal with the subsequent mutations 15:56
@inkycatz Okay. Then the last one on this topic is: 3) Are there examples of successes in medicines synthesized by the way you investigate? (folding target, folding potential medicines in function of this, synthesizing, then of course in vivo tests)? or is it still fundamental research? 15:56
vmulligan It's like evolution, but without going through generations of organisms.  Within your own body, many immune cells are produced, each expressing a different antibody.  Those antibodies that stick to something signal for the cells that produced them to multiply, thus ensuring production of more of that antibody in response to an antigen.  (That's simplified, of course.) 15:57
vmulligan Done. 15:57
bkoep To my knowledge, there are no "designer proteins" currently available as therapeutics 15:57
@inkycatz Bruno, love your question about getting a visual look behind the scenes - I think this question merits a blog, or video blog, or something, but it will probably be too much to go into today. :) 15:57
@inkycatz Here's another question from our thread: "There have been top solutions for abeta, is there a need for further refinement of those results (even if no points would be given) ? Are there other science puzzles that would benefit from further refinement ?" 15:58
bkoep There are naturally-derived proteins that are used in some treatments and vaccines 15:58
@inkycatz (just queueing em up for you guys, no pressure :D 15:58
bkoep But for now, protein design is still the cutting-edge of medicine research 15:58
vmulligan The best example we have right now of a designer protein that's going into the clinic is KumaMax, an artificial enzyme designed in our lab to break down gluten to help people with Celiac disease. 15:58
vmulligan It's still early stages for clinical tests, though. 15:58
@betahelix ...which was partly created by undergrads using Foldit for the iGem competition! 15:59
vmulligan But it's at the translational research stage, not the basic research stage.  Most protein design is still basic research, though. 15:59
@betahelix paper for that: 16:00
@inkycatz <- more info about  KumaMax 16:00
@betahelix 16:00
@betahelix 16:00
@betahelix Ok gang, I have to go now... but I wanted to congratulate you all again! 16:01
@inkycatz thanks for dropping by! 16:01
BletchleyParkirc could you please answer the last question ? 16:01
bkoep Re: refinement of previous puzzles -- we would like to revisit some of your Abeta designs 16:01
vmulligan We also have some Abeta designs of our own that we'll be asking you to refine. 16:02
@betahelix Thank you for all your hard work, as usual. Keep up the great folding... bye everyone. 16:02
BletchleyParkirc very good 16:02
bkoep Bye, betahelix! 16:02
@inkycatz thanks for all the great links :D 16:02
vmulligan Bye, betahelix! 16:02
jeff101 thanks for having this Chat 16:02
BletchleyParkirc bye and thank you 16:02
@rav3n_pl cu betahelix 16:02
@betahelix Thank you! :-D Bye bye! 16:02
bkoep Yes, the Abeta binder is a really interesting project that is still in early stages; you can expect to see more Abeta-related puzzles in the future 16:03
BletchleyParkirc The sooner the better... 16:03
@rav3n_pl just not make 100+ residues one w/filters.... 16:04
bkoep @BletchleyParkirc, duly noted 16:04
BletchleyParkirc if that gives better science results I'm even for bigger models 16:04
Brow42IRC for vmulligan and bkoep: Players, I think, are very dedicated to hiding hydrophobes. How do they perform on proteins that are dimers, have binding sites, or have domains that transform in response to another binding site elsewhere?  These proteins have a lot of exposed. Do folders still sucessfully find the natives? Some relevance to drug design too. 16:04
@rav3n_pl big ones are unplayable :( 16:05
@inkycatz Hey scientists, do you have time to stick around for a few more questions, or do you have to dash off? :) Be good to get a few from the floor here. 16:05
bkoep I have a few more minutes 16:05
@inkycatz (don't cut the log yet, jflat!) 16:05
bkoep @Brow42, yes, you're right that players are very intent on hiding hydrophobic residues 16:06
@TimovdL Do the repeats of old puzzles really serve a purpose or are they just to keep us busy? 16:06
bkoep But that's a good think--the burial of hydrophobics is generally the strongest force in protein folding 16:06
@inkycatz (for those of you that do have to dash off, thank you for coming!) 16:07
jeff101 do you see noticeable improvement on the repeat puzzles? 16:07
@rav3n_pl as far I read description, revisiting old puzzles are to check that new tools we have are helping us and making higher score possible 16:07
bkoep Natural proteins often do have exposed hydrophobic residues that make structure prediction difficult 16:07
alwen tyroxine seems to often be outside 16:08
alwen *s 16:08
bkoep But we've seen that designed proteins without exposed hydrophobics are super-stable, which is one reason I think Foldit players seem so adroit at protein design 16:08
@TimovdL And another question important for script design: do the slow filters interact with the tools (wiggle, shake, rebuild, mutate)? 16:09
bkoep @Timo, the revisited puzzles are useful to us 16:10
@rav3n_pl if we do stuff when filters are disabled or enabled tools make different resoults? (ie shape) 16:10
BletchleyParkirc I find I can get better scores with filters always on than with bypassing them 16:10
bkoep Often, we don't have time to closely examine the results, but they can serve as a kind of "canary in a coal mine" when we make a debilitating change in the score function 16:12
@rav3n_pl in meaning: start position, filters enabled, wiggle -> end position and strat position, filters disabled, wiggle, filters enabled, -> end position. It will be same in both runs? 16:12
BletchleyParkirc no 16:12
@rav3n_pl thats bad 16:13
Brow42IRC Wiggle has behaved very strangely since NC.  It just gets stuck. 16:13
bkoep also @Timo, the filters are not as well-integrated with some of the core tools as we would like 16:13
Alve Revisits are good practice as well, plus seeing a variety of proteins 16:13
Brow42IRC So it is hard to do a controlled test of wiggle. 16:13
alwen wiggle sidechains mostly does nothing now 16:13
bkoep However, we're actively working on that and you should see improvements in the near future 16:13
Susume2 for performance, it would be best if the filters did not affect the tools at all; then we could run the tools with filters off for speed and not fear missing out on points 16:14
jmbrownlee333 curious about how many genes were ordered for expression. a few a dozen? a hundred? 16:15
@TimovdL It would be nice if there would be no interaction at all, like sus says, I can edit almost all scripts to get more speed then. 16:15
bkoep @jmbrownlee333, in the most recent batch, we ordered just those 8 designs 16:15
BletchleyParkirc I think wiggle evaluates the result of a sub-wiggle and then decides its next move. if filters are on, that decision will be influenced. (my observation) 16:15
bkoep Overall, we've ordered several dozen Foldit designs 16:15
jeff101 just recently or ever? 16:16
@TimovdL And a new kind of running mode (filters off unless the recipe asks for the score) would be possible. 16:16
bkoep In addition to a few thousand Foldit player-designed Ebola-binders, which we were able to screen in very high throughput 16:16
BletchleyParkirc @bkoep, what was the result of that screening ? 16:16
Brow42IRC jmbrownlee333's question just reminded me of news stories in the last month that basically implied people would soon be doing bioengineering in their new technology coming that will make it easier for you to produce custom proteins? 16:16
jeff101 3D protein printer? 16:18
jmbrownlee333 will that ebola binder library study be published 16:18
bkoep The ebola-binding work is still under way 16:18
bkoep The hotspots found by Foldit players have been extremely helpful to some of us in lab designing our own peptides 16:19
BletchleyParkirc that is good to read ! 16:19
bkoep The full peptide designs have had trouble competing with some of the designs from our scientists 16:19
BletchleyParkirc You might put some of those up for refinement 16:20
jeff101 I posted a question right before the Chat began. It has links to Feedbacks. 16:21
jeff101 One Feedback it links to is about comparing revisit puzzles with their originals. 16:21
@inkycatz Thanks for adding to the thread, jeff! Saw that feedback. 16:22
@inkycatz We don't want to take up all of your day, bkoep & vmulligan :) 16:23
@inkycatz Thank you both for coming by though! 16:23
bkoep I should get back to the lab 16:23
bkoep Thanks everyone for a great chat! 16:23
vmulligan Sorry I was in and out a bit -- lots on the go! 16:23
@inkycatz We appreciate you hanging out for a little overtime here 16:24
MuckeMcFly have a nice day bkoep 16:24
jeff101 thanks again! 16:24
BletchleyParkirc thank you for your time and asnwers, bye ! 16:24
@inkycatz Thanks to EVERYONE for coming. This is the end of our chat 16:24
@inkycatz Our next chat is in April! 16:24

Generated by 2.13.1 by Marius Gedminas
- find it at!

User login
Download links:
  Windows    OSX    Linux  
(10.12 or later)

Are you new to Foldit? Click here.

Are you a student? Click here.

Are you an educator? Click here.
Social Media

Only search
Other Games: Mozak
Recommend Foldit
Top New Users

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