2008-04-05.log:16:58 < fenn> you dont have the heuristics for protein folding
2008-04-05.log:16:59 < kanzure> you don't need protein folding
2008-04-05.log:18:24 < kanzure> heh, and you mentioned protein folding earlier
2008-04-05.log:18:24 < kanzure> so this is where you'd need protein folding
2008-05-21.log:01:03 < fenn> origami strikes me a lot like protein folding
2008-06-03.log:22:08 < kanzure> Vedestin: would require protein folding, or lots of selection
2008-06-03.log:22:11 < kanzure> probably avoiding protein folding
2008-06-03.log:22:14 < kanzure> (2) biomolecular engineering with protein folding. let's hope we can model the folding of polymerase.
2008-06-03.log:22:15 < fenn> it's the old 'protein folding' problem
2008-06-03.log:22:29 < kanzure> re: protein folding, see CASP, the protein folding prediction compettition
2008-06-09.log:22:03 < fenn> kanzure: i think with easy enough access to DNA synthesis, the protein folding problem will become obvious
2008-06-09.log:22:04 < kanzure> fenn: has DNA synthesis been our barrier to protein folding characterization ?
2008-06-09.log:22:05 < kanzure> fenn: ok, so with DNA synthesis then we can try to do protein folding I guess
2008-06-10.log:00:24 < kanzure> protein folding, I mean.
2008-06-10.log:00:29 < kanzure> unless we solve some other problems re: protein folding of course
2008-06-16.log:21:31 < kanzure> You know, fold.it really makes sense in light of the compartmentalized evolution hypotheses of protein folding, and then the reverse of that evolutionary encoding (having everybody play around) is an interesting reverse attempt.
2008-06-24.log:11:07 < kanzure> Also, the fact that it's all protein folding really, really sucks, and means that you're not really going to be laying down exact wires for electronics ever, that's unacceptable. But one possible method -- if we really, really wanted the self-replicable, evolvable ribosomal manufacturing unit, is to use an STM tip to inject mRNA into a ribosome at the end of the tip, and then the output channel would go into another nanotubule as 
2008-06-24.log:11:22 < kanzure> the problem that evolution was solving probably wasn't protein folding
2008-06-24.log:11:22 < kanzure> since protein folding is of the amino acids themselves, no?
2008-06-24.log:11:23 < kanzure> and "solving" protein folding would mean something like being able to digitally specify what protein to make
2008-09-24.log:23:32 < kanzure> do they even understand how to use the protein folding information for their goals?
2008-10-07.log:14:26 < nsh> for making constructive games such as the protein folding one
2009-03-09.log:23:38 < xp_prg> guys I need help a PHD in protein folding is coming over in 30 minutes he wants to see some papers on embryonic forming in xenopus, can you guys help me find some real quick?
2009-03-09.log:23:58 < xp_prg> a guy with a phd in protein folding is coming over in 8 minutes help!
2009-08-21.log:16:41 < genehacker> cool talk on protein folding here live
2009-10-10.log:14:20 < fenn> it's a biology game about protein folding
2010-09-30.log:09:39 < dbolser> we tried to learn some from protein structures to solve  the protein folding problem
2011-06-08.log:14:46 < fenn> though lately he's gone off on this crusade to solve the protein folding problem
2012-01-23.log:22:42 < yashgaroth> I do agree with klafka that we're waiting for physics, to help solve the protein folding problem
2012-01-23.log:22:42 < kanzure> the protein folding problem is irrelevant
2012-03-20.log:22:08 < Steel2> figure out a protein folding equivalent game
2012-03-20.log:22:08 < lichen> protein folding on phones sounds really slow
2012-04-21.log:18:52 < yashgaroth> ooh a computational solution to the protein folding problem
2012-04-21.log:19:06 < roksprok> yashgaroth: is it an enabling technologies problem (like faster quicker gene synthesis and sequencing and higher resolution imaging and protein folding predication) or a 'we need to characterize these 20000 proteins better'
2012-04-21.log:19:06 < yashgaroth> protein folding would solve 90% of biology
2012-04-21.log:19:07 < kanzure> i suspect that none of you are qualified to comment since you haven't written any protein folding algorithms
2012-04-21.log:19:08 < diginet> I don't have to have: protein folding is like the travelling salesmen problem, it can't be solved efficienctly on a turing machine
2012-04-21.log:19:11 < roksprok> so then why is 'the protein folding problem' still a problem that will solve 90% of biology when it is solved
2012-04-21.log:19:12 < AdrianG> thats a completely different problem from predicting protein folding
2012-04-21.log:19:12 < roksprok> yashgaroth: isn't that more kinematics? or does protein folding encompass it all
2012-04-21.log:19:16 < AdrianG> well, godel or no godel, eventually we would be able to work out protein folding.
2012-04-21.log:19:18 < kanzure> yashgaroth: you suck for bringing up protein folding
2012-09-20.log:09:53 < nmz787> i just posted the other day about an idea for protein folding experiments
2012-11-29.log:13:24 < docl> protein folding news: http://www.nature.com/news/proteins-made-to-order-1.11767
2013-01-14.log:10:10 < eleitl> inverse protein folding is almost solved
2013-01-14.log:10:11 < eleitl> strangely, it is easier than straight protein folding
2013-01-14.log:12:14 < rigel> i have no skills to try and implement this but are fpgas useful for solving protein folding problems?
2013-01-22.log:12:17 < eudoxia> are they collaborating on a design for a rod logic computer? solving the protein folding problem? helping eleitl with his cryo org?
2013-06-21.log:13:42 < kanzure> there are a number of protein folding motifs that we already have heuristics for
2013-12-24.log:14:44 < delinquentme> I know we've got all kinds of variations in protein folding ... so say sticking a single cell in a minturized bioreactor such as http://marblar.com/technology/US8133722
2014-05-15.log:19:09 < QuantumG> protein folding, computation
2014-06-02.log:14:13 < dbolser> based on the physics of protein folding
2014-08-11.log:23:23 < thundara> Last I'd heard, they made their own supercomputer to specially do protein folding / MD stuff
2014-10-11.log:21:07 < kanzure> protein folding http://xkcd.com/1430/
2014-10-31.log:02:06 < fenn> genehacker: just solve the protein folding problem and you're good to go
2014-10-31.log:02:08 < genehacker> fenn: for what we're doing we don't need to solve the protein folding problem
2014-10-31.log:15:29 < fenn> i was reading science magazine while eating breakfast. i've been teasing genehacker about how solving the protein folding problem would put him out of a job
2014-11-28.log:11:09 < delinquentme> quantum compute: what are the machanimss by which it solves the protein folding problem?
2014-12-27.log:12:15 < yoleaux> "Chaperonins are proteins that provide favourable conditions for the correct folding of other proteins, thus preventing aggregation. Newly made proteins usually must fold from a linear chain of amino acids into a three-dimensional form. Chaperonins belong to a large class of molecules that assist protein folding, called molecular chaperones." — http://en.wikipedia.org/wiki/Chaperonin
2014-12-31.log:20:24 < kanzure> "Synthetic tRNAs—those that are transcribed in vitro from a DNA sequence that has been ligated into an appropriate plasmid—have a wide variety of applications that range from testing tRNAs for their requirements for amino-acylation (1) to providing reagents for the investigation of protein folding (2). The examples that will be described here involve specific tRNAs with altered anticodons. The construction of several of these tRNAs ...
2015-01-28.log:15:39 < nmz787_i> justanotheruser: so what, we need to figure out protein folding
2015-01-28.log:15:40 < justanotheruser> nmz787_i: donate bitcoin to a solution. You can pretty much consider using this currency to be a donation to solving protein folding (but a very ineffecient one where most of the money goes to scammers/pumpers and a little bit going towards solving the problem)
2015-01-28.log:15:41 < nmz787_i> but why not use an actual function (protein folding) which is as-hard-as-crypto-is-today?
2015-01-28.log:15:44 < nmz787_i> for verifying your protein folding algo was right
2015-01-28.log:15:48 < kanzure> oh this isn't the one that talks about a separate blockchain of protein folding consensus hrm
2015-01-28.log:15:52 < kanzure> but to be fair, there have been some proposals in the past to do protein folding to achieve consensus
2015-01-28.log:15:52 < justanotheruser> so basically this is a speculative asset that loses value (considered donations) by funding protein folding
2015-01-28.log:15:53 < justanotheruser> so in the end you can donate $100 in bitcoins to protein folding, or you can lose $100 in foldingcoins to donate to protein folding.
2015-01-28.log:15:54 < kanzure> on a related note, i'm also not very convinced that folding@home is a good idea, even if it takes 30 years to achieve stronger results or something, i suspect that protein folding is either the wrong problem or we're going about proteins wrong
2015-01-28.log:15:54 < kanzure> i sent this long-ass email to freitas the other day complaining about the focus on protein folding
2015-01-28.log:15:59 < kanzure> eh, it is worth knowing that "protein folding algorithms are presently incapable of correctly predicting all types of folds"
2015-01-28.log:16:15 < ParahSailin> have there been any papers resulting from protein folding other than the distributed computing papers
2015-01-28.log:16:18 < kanzure> not very sure. i haven't even been able to find a super-good review of latest protein folding "stuff".
2015-01-29.log:13:49 < kanzure> (and was also based on protein folding)
2015-03-14.log:22:56 < delinquentme> kanzure, why cant blockchain be adapted to protein folding problems?
2015-03-30.log:13:18 < kanzure> gene_hacker: someone should do nanotech things with the subset of protein folding that happens to be predictable (e.g., strike out the useless unpredictable amino acids)
2015-05-30.log:05:51 < Urchin[emacs]> FourFire: have you considered colaborating with one of the protein folding distributed projects?
2015-05-30.log:15:18 < kanzure> ... themselves). To some degree, “inflated fitness advantages” occur in theories of runaway sexual selection (where everyone tries to mate with whoever seems even nominally smartest). To whatever extent such sexual selection was occurring, we should decrease our estimate of the sort of cognitively produced fitness advantage that would carry over to a machine intelligence trying to work on the protein folding problem (where you do not ...
2015-05-30.log:15:21 <@fenn> even a marginally better algorithm would be selected for in a protein folding competition
2015-07-19.log:14:34 < kanzure> FourFire: you should just use subunits to make structures, problem solved, no protein folding needed
2015-07-19.log:15:26 < kanzure> mgin: are you asking about protein folding, or are you asking about the "central dogma"?
2015-07-19.log:15:26 < mgin> right, isn't protein folding a huge problem by itsefl??
2015-09-22.log:19:16 < kanzure> "Ribosomal robots: evolved designs inspired by protein folding" http://creativemachines.cornell.edu/sites/default/files/Ribosomal%20Robots.pdf
2015-10-08.log:14:30 < kanzure> since we don't know how to do all protein folding predictions, you can use bruteforcing with direct dna synthesis of a million different variants and then look at the shape of each resulting protein
2015-10-08.log:15:00 < fenn> i think the right strategy is to skip the protein folding problem entirely by using rationally engineered proteins instead of giant blobs of shit
2015-10-09.log:17:25 < kanzure> on page 25 i'm not sure why they think a markov model is an evolutionary search algorithm (they are using a markov model for protein folding predictions)
2015-10-30.log:12:20 < kanzure> dna origami predictability is only slightly better than protein folding predictability
2015-11-07.log:13:34 < kanzure> "you should feel bad because you are over-estimating the utility of protein folding simulations" with slightly more teeth
2015-11-16.log:13:17 < fenn> andares i think you'll have to either be very clever with multiple unusual DNA polymerases, or solve the protein folding problem and build a light-directed DNA synthase from scratch
2016-01-04.log:15:07 < kanzure> iirc the last goal you set was "solve protein folding"
2016-01-04.log:15:28 < xentrac> yeah, protein folding in general is a much harder problem than that
2016-04-24.log:15:24 < kanzure> "Actually, protein design is fairly routine now. We used Exacycle, an idle-cycle computer at Google, to show that given reasonable amounts of computer time, design of proteins is a straightforward process. You don't need advanced supercomputers or even GPUs to solve this problem- just classic clusters- although GPUs (not supercomputers) speed up the rate significantly. The important reason why this works is that while "protein folding is ...
2016-07-03.log:17:39 < kanzure> "CONFOLD: Residue-residue contact-guided ab initio protein folding" http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.714.5591&rep=rep1&type=pdf
2016-07-21.log:15:00 < kanzure> re: restriction of protein engineering to only those spiral shapes or whatever.   i think that needs to be double checked. i think some of the protein folding is solved for. so we should find some good chases in existing proteins. then paste catalytic protein domains to that, run selection experiments and get improved proteins with known-simulatable backbones and folding.  and next most of biology should be replaced with only those ...
2016-07-21.log:15:43 < kanzure> i think foldit is assuming that all protein folding predictions can be solved by trying to predict and score the lowest minimum energy configuration or something
2016-07-21.log:16:15 < kanzure> slide 8 is pretty interesting (simulation vs experiment for protein folding) http://web.stanford.edu/class/cs279/lectures/lecture5.pdf
2016-07-21.log:18:53 < kanzure> protein design is commonly said to be hard due to protein folding-- but often that's handwaving because discussion of the actual problems don't occur afterwards.
2016-07-22.log:14:16 < kanzure> dna origami is probably more predictable than protein folding
2016-09-08.log:19:14 < kanzure> "It's worse than that though.  Never forget that the true "compiler" of DNA is Physics, specifically that of protein folding, conformational dynamics, and catalysis.  It won't be simulated anytime soon with anything approaching useful kinetic accuracy.  (It's not clear if we'll ever get kT-accurate quantum simulations of correlated electron wavefunctions that scale to protein-sized systems, though there is some hope in the far future ...
2016-12-18.log:14:09 < kanzure> dna origami is a much more solvable folding problem than protein folding
2016-12-18.log:14:10 < kanzure> someone is going to find a subset of proteins that fold in a highly specific way, and then use sgRNA or something to force them to bind together in certain ways, and then protein folding will become a non-problem as long as you don't use natural shapes
2017-04-05.log:19:00 < yashgaroth> I don't get how it would allow one to "consolidate all the good genetic mutations"...if you wanted to do it without solving the protein folding problem, there's plenty of mismatch endonucleases that'll cleave incorrect bases
2017-04-07.log:12:29 < kanzure> protein folding simulation stuff just isn't going to be ready for a very long time
2017-05-17.log:20:43 < yashgaroth> getting a polymerase to spit out a conversion will take a long time, until we solve the protein folding problem, which will take an even longer time
2017-08-12.log:15:31 < yashgaroth> even if you solve the protein folding problem I doubt that'd be possible, especially when many errors will just introduce a stop codon
2017-08-12.log:15:33 < kanzure> ...transcription, protein folding or it folds and the protein just fails to do its job) and the cell dies. Whoops."
2017-09-27.log:20:47 < heath> btw, re: protein folding, check out protein origami: 10.1016/j.cbpa.2017.06.020
2017-10-28.log:15:41 <@nmz787> and that this sort of signal-maximizing algorithm of moving the bot swarm around, might have the same effect as protein folding would
2017-10-28.log:15:44 <@nmz787> and thus solve protein folding problems
2017-10-28.log:17:25 < fltrz> Im quite skeptical about many-body problems being faster to solve with "physics/real systems", in the end computers are physical systems too. I can imagine certain restricted problems being faster to solve with analog electronics and/or optics (like 4F correlators), but I would be lying if I had any idea of how to efficiently translate protein folding to analog electronics for example
2017-10-28.log:20:31 <@nmz787> fltrz: well computers are pretty good these days, and computers still aren't solving protein folding problems quickly or well
2018-03-06.log:14:28 < emeraldgreen> Also: do you think there could be some use of opensource implementation of this protein folding RNN ? biorxiv.org/content/early/2018/02/14/265231
2018-04-13.log:16:18 < nmz787> then top that with we dont knot protein folding... secondary/tertiary effects probably can come into play... i.e. self-catalytic stuff that only works during setup phase (i.e. it breaks itself, or hides the catalytic domain inside the protein core)
2019-03-11.log:14:48 < fenn> importance to pharmaceuticals than it is to Alphabet." (protein folding prediction)
2019-05-14.log:17:18 < yoleaux> GitHub - POVME/POVME: A pocket volume analyzer for use in computational protein folding.
2020-01-20.log:07:59 < kanzure> do we have a protein folding wish list?
2020-02-12.log:07:56 < kanzure> specifically i'm excluding the question of protein folding, and i'm willing to have a much more limited environment of folds and peptide-peptide or protein-protein glue
2020-02-12.log:08:00 < kanzure> molecular nanotechnology doesn't necessarily need to solve protein folding, just building stuff out of protein and amino acids (e.g. the engineering thing where the underlying science doesn't matter as long as you have predictable effects)