--- Log opened Tue Sep 19 00:00:22 2017 00:17 -!- entity8421[m] [entity8421@gateway/shell/matrix.org/x-euaimwnhgbkaiodf] has joined ##hplusroadmap 00:28 -!- Rmesil8O4b[m] [rmesil8o4b@gateway/shell/matrix.org/x-qstxfkeyzakbvxpy] has joined ##hplusroadmap 00:34 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-fjnilpjwuekgauqu] has joined ##hplusroadmap 00:40 -!- RebelCoder [~Yuriy@95.143.115.254] has joined ##hplusroadmap 01:21 -!- jaboja [~jaboja@jaboja.pl] has joined ##hplusroadmap 01:23 -!- poppingtonic [~brian@unaffiliated/poppingtonic] has quit [Ping timeout: 264 seconds] 01:33 -!- aeiousomething [~aeiousome@gateway/vpn/privateinternetaccess/aeiousomething] has joined ##hplusroadmap 01:40 -!- poppingtonic [~brian@unaffiliated/poppingtonic] has joined ##hplusroadmap 01:58 -!- preview [~quassel@2407:7000:842d:4000::3] has joined ##hplusroadmap 02:33 -!- bluebear_ [~dluhos@80.95.97.194] has joined ##hplusroadmap 02:41 -!- fltrz [d4470ea2@gateway/web/freenode/ip.212.71.14.162] has quit [Ping timeout: 260 seconds] 02:57 < kanzure> blort 02:57 -!- Gurkenglas [~Gurkengla@dslb-178-000-177-083.178.000.pools.vodafone-ip.de] has quit [Ping timeout: 246 seconds] 02:57 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has quit [Ping timeout: 264 seconds] 03:31 < archels> kanzure: lasers! 03:35 -!- preview [~quassel@2407:7000:842d:4000::3] has quit [Ping timeout: 264 seconds] 03:35 -!- darsie [~username@84-113-55-42.cable.dynamic.surfer.at] has joined ##hplusroadmap 03:43 -!- poppingtonic [~brian@unaffiliated/poppingtonic] has quit [Remote host closed the connection] 03:48 -!- poppingtonic [~brian@unaffiliated/poppingtonic] has joined ##hplusroadmap 03:55 -!- poppingtonic [~brian@unaffiliated/poppingtonic] has quit [Ping timeout: 260 seconds] 04:11 -!- traumschule [~traumschu@185.104.184.179] has joined ##hplusroadmap 04:13 -!- poppingtonic [~brian@unaffiliated/poppingtonic] has joined ##hplusroadmap 05:03 -!- jaboja [~jaboja@jaboja.pl] has quit [Ping timeout: 240 seconds] 05:05 -!- jaboja [~jaboja@jaboja.pl] has joined ##hplusroadmap 05:33 -!- CRM114 [~urchin@unaffiliated/urchin] has quit [Ping timeout: 258 seconds] 05:33 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-fjnilpjwuekgauqu] has quit [Quit: Connection closed for inactivity] 05:52 -!- jtimon [~quassel@199.31.134.37.dynamic.jazztel.es] has quit [Ping timeout: 248 seconds] 06:12 < kanzure> what about them? 06:26 < kanzure> "mouse cloud" http://www.vium.com/technology/ 06:49 < cluckj> that is less cool than I was hoping 06:49 < cluckj> oh nevermind, it does have actual mice 06:50 < cluckj> it's exactly as cool as I was hoping. 06:51 < cluckj> like facebook but for mice! 06:51 -!- esmerelda [~mabel@unaffiliated/jacco] has quit [Ping timeout: 252 seconds] 06:54 -!- justanotheruser [~justanoth@unaffiliated/justanotheruser] has quit [Ping timeout: 255 seconds] 07:27 < archels> they might be faster at moving liquids than piezos 07:31 < poppingtonic> more like big brother for mice 07:32 < cluckj> ^ neither of us are wrong 07:38 < poppingtonic> gasp... 07:39 < poppingtonic> does that mean... that facebook is big brother? 07:39 < cluckj> :o 07:47 < pompolic> micebook 08:10 < archels> pricey https://www.seeedstudio.com/Hamamatsu-C12880MA-MEMS-u-Spectrometer-and-Breakout-Board-p-2916.html 08:19 -!- Gurkenglas [~Gurkengla@dslb-178-000-177-083.178.000.pools.vodafone-ip.de] has joined ##hplusroadmap 08:20 -!- poppingtonic [~brian@unaffiliated/poppingtonic] has quit [Quit: Leaving.] 08:34 -!- Gurkenglas [~Gurkengla@dslb-178-000-177-083.178.000.pools.vodafone-ip.de] has quit [Ping timeout: 240 seconds] 08:37 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-gcncexkzthirrkyj] has joined ##hplusroadmap 08:37 -!- jaboja [~jaboja@jaboja.pl] has quit [Ping timeout: 240 seconds] 08:43 < pasky> kanzure: chido has a good point - nowhere on the site it talks about mice (besides the infographics) 08:43 < pasky> it's totally for people! 08:45 < kanzure> it's multi-mammalious, it works for both mice and people. 08:46 -!- Gurkenglas [~Gurkengla@dslb-178-000-177-083.178.000.pools.vodafone-ip.de] has joined ##hplusroadmap 09:12 -!- mf1008 [~mf1008@unaffiliated/mf1008] has quit [Quit: Cave quid dicis, quando, et cui] 09:12 -!- mf1008 [~mf1008@unaffiliated/mf1008] has joined ##hplusroadmap 09:19 -!- delinquentme [~delinquen@2602:306:ceb7:990:c5cf:32a3:c15:357b] has joined ##hplusroadmap 09:26 -!- jaboja [~jaboja@jaboja.pl] has joined ##hplusroadmap 09:37 -!- delinquentme [~delinquen@2602:306:ceb7:990:c5cf:32a3:c15:357b] has quit [Ping timeout: 252 seconds] 09:50 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has joined ##hplusroadmap 09:55 -!- jaboja [~jaboja@jaboja.pl] has quit [Ping timeout: 240 seconds] 10:06 < kanzure> archels: not sure what you mean. laser-pulsed jet thing? 10:25 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has quit [Ping timeout: 260 seconds] 10:38 -!- jaboja [~jaboja@jaboja.pl] has joined ##hplusroadmap 10:39 -!- c0rw1n_ [~c0rw1n@206.41-244-81.adsl-dyn.isp.belgacom.be] has joined ##hplusroadmap 10:43 -!- delinquentme [~delinquen@2602:306:ceb7:990:cdf7:d527:d537:2e8c] has joined ##hplusroadmap 10:47 < kanzure> delinquentme: you gonna do gwas stuff 10:47 < kanzure> ? 11:04 < delinquentme> i was just looking through some of the tutorials on NIH before I got distracted 11:04 -!- traumschule [~traumschu@185.104.184.179] has quit [Ping timeout: 240 seconds] 11:09 < kanzure> yep 11:16 < delinquentme> wrapped up in trying to get effing sound to work on my new desktop 11:17 < delinquentme> in a perfect world academics would push out their everyday queries to programmers wanting to learn things about biotech :P 11:17 < delinquentme> theres no repository of "interesting questions" which programmers can take up to learn biotech + do this kinda stuff... so yeah 11:17 < delinquentme> my short answer is "what do I do w gaws information ... IE what do I want to test?" 11:18 < kanzure> ask gwern in #lesswrong he has lots of stupid gwas questions 11:27 < kanzure> http://www.uah.edu/otc/available-technologies/biotech/microfluidic-reactor 11:35 -!- augur [~augur@198-27-215-123.static.sonic.net] has joined ##hplusroadmap 11:47 -!- augur [~augur@198-27-215-123.static.sonic.net] has quit [Remote host closed the connection] 12:08 -!- augur [~augur@198-27-215-123.static.sonic.net] has joined ##hplusroadmap 12:23 -!- delinquentme [~delinquen@2602:306:ceb7:990:cdf7:d527:d537:2e8c] has quit [Ping timeout: 255 seconds] 12:25 < archels> kanzure: it was re your wanting to move liquids faster than a piezo could 12:27 < kanzure> but.. i'm not convinced. you want to do laser pumping? 12:28 -!- bluebear_ [~dluhos@80.95.97.194] has quit [Quit: Leaving.] 12:30 -!- jtimon [~quassel@199.31.134.37.dynamic.jazztel.es] has joined ##hplusroadmap 12:39 -!- augur [~augur@198-27-215-123.static.sonic.net] has quit [Remote host closed the connection] 12:42 -!- augur [~augur@198-27-215-123.static.sonic.net] has joined ##hplusroadmap 12:45 -!- jaboja [~jaboja@jaboja.pl] has quit [Remote host closed the connection] 12:55 -!- jaboja [~jaboja@jaboja.pl] has joined ##hplusroadmap 12:55 -!- NikopolSohru [~NSohru@89.38.96.187] has joined ##hplusroadmap 13:10 -!- fltrz [d5db905c@gateway/web/freenode/ip.213.219.144.92] has joined ##hplusroadmap 13:20 -!- NikopolSohru [~NSohru@89.38.96.187] has quit [Ping timeout: 248 seconds] 13:32 -!- NikopolSohru [~NSohru@37.48.65.48] has joined ##hplusroadmap 14:18 -!- augur [~augur@198-27-215-123.static.sonic.net] has quit [Remote host closed the connection] 14:25 < kanzure> hmm bill peck (twist) is in austin this weekend 14:27 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has joined ##hplusroadmap 14:32 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has quit [Ping timeout: 248 seconds] 14:34 -!- preview [~quassel@103.23.18.14] has joined ##hplusroadmap 14:39 -!- c0rw1n_ [~c0rw1n@206.41-244-81.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 240 seconds] 14:39 -!- NikopolSohru [~NSohru@37.48.65.48] has quit [Quit: Leaving] 14:51 -!- kwgwk [~kwgwk@boole.london.hackspace.org.uk] has quit [Ping timeout: 248 seconds] 15:06 -!- augur [~augur@198-27-215-123.static.sonic.net] has joined ##hplusroadmap 15:51 -!- jaboja [~jaboja@jaboja.pl] has quit [Remote host closed the connection] 16:31 -!- augur [~augur@198-27-215-123.static.sonic.net] has quit [Remote host closed the connection] 16:34 < kanzure> https://github.com/astorfi/3D-convolutional-speaker-recognition 16:35 -!- Malvolio [~Malvolio@unaffiliated/malvolio] has quit [Ping timeout: 248 seconds] 16:37 -!- augur [~augur@198-27-215-123.static.sonic.net] has joined ##hplusroadmap 16:52 -!- augur [~augur@198-27-215-123.static.sonic.net] has quit [Remote host closed the connection] 16:58 -!- streety_ [~streety@li761-24.members.linode.com] has quit [Quit: ZNC 1.6.5 - http://znc.in] 17:11 < nmz787> https://www.google.com/patents/US9707535 17:11 < nmz787> .title 17:11 < yoleaux> Patent US9707535 - Microfluidic reactors for oligonucleotide synthesis - Google Patents 17:12 -!- Malvolio [~Malvolio@unaffiliated/malvolio] has joined ##hplusroadmap 17:18 < kanzure> there's a thesis floating around, ybit has it 17:18 < nmz787> apparently someone else too 17:20 < nmz787> err, I guess same person actually 17:21 < kanzure> ybit: ping 17:23 -!- augur [~augur@198-27-215-123.static.sonic.net] has joined ##hplusroadmap 17:23 < heath> nmz787: pdf sent 17:24 < kanzure> nmz787: heath is interested in an integrated microfluidic dna synthesizer + sequencer 17:25 < nmz787> I mentioned I am aiming for single-molecule (or scaling up by orders of magnitude from there til it works) 17:26 < kanzure> i sort of forget whether you two talk with each other frequently 17:26 < kanzure> it is difficult for me to synchronize everyone 17:26 < heath> i think if you're going to build an archival storage system such as https://www.youtube.com/watch?v=xfFfQzELYHE then you'll need syn/seq integrated 17:27 < kanzure> .title 17:27 < yoleaux> HPE StoreEver ESL G3 with LTO-7 - YouTube 17:27 < kanzure> needing read+write doesn't make it any more easy or achievable or realistic in the near-term 17:27 < heath> true 17:29 < kanzure> it's interesting that they were able to avoid valves in that paper, that's kinda nice 17:29 < kanzure> 16 spots is really underperforming on the density front though? 17:34 < nmz787> that thesis, basically they are limited by their glass tech, I think, and also that they are using solid-phase synthesis on microbeads 17:35 < nmz787> but it's pretty much a great start otherwise 17:36 < nmz787> I love that the simulation section is so detailed 17:36 < nmz787> they used a demo version of COMSOL :P 17:47 < kanzure> microbeads shouldn't be the limiting aspect? 18:08 -!- darsie [~username@84-113-55-42.cable.dynamic.surfer.at] has quit [Ping timeout: 240 seconds] 18:09 -!- preview [~quassel@103.23.18.14] has quit [Ping timeout: 240 seconds] 18:14 -!- yashgaroth [~yashgarot@2606:6000:cd4d:3300:f5e0:f867:a11d:8d52] has joined ##hplusroadmap 18:22 < nmz787> sup yashgaroth 18:22 < yashgaroth> yo 18:24 < nmz787> how's the protein biz? 18:41 < yashgaroth> meh, same old 18:43 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-gcncexkzthirrkyj] has quit [Quit: Connection closed for inactivity] 18:57 -!- cluckj [~cluckj@static-173-59-27-112.phlapa.ftas.verizon.net] has quit [Read error: Connection reset by peer] 18:58 -!- cluckj [~cluckj@static-173-59-27-112.phlapa.ftas.verizon.net] has joined ##hplusroadmap 18:58 < kanzure> yashgaroth: can you think of any conceivable way to make biology more immune to frameshift errors? 19:01 < yashgaroth> I sincerely hope it's easier to fix synthesis than to re-engineer the way the central dogma functions 19:02 < kanzure> well... errors. 19:05 < kanzure> and it takes too long to go from dna -> chromosome -> cell -> cell survival is impacted by some sort of stupid error in the original dna molecule. 19:05 < kanzure> if the dna molecule could be disposed of more quickly (without dna sequencing), then the situation would look better 19:06 < yashgaroth> well there is the stuff with oligo annealing for assembly where any mismatched pieces get degraded, not sure how well it functions in practice 19:06 < fenn> you could theoretically add more constraints to the sequence you're trying to synthesize such that any errors are recognized by an enzyme or precipitated out 19:07 < fenn> like "there are no CTAAGGG" 19:07 < kanzure> you mean: synthesize a-b-c and a-c gets recognized as an error by the enzyme (and also a-c is not intended to be anywhere in the molecule)? 19:07 < fenn> right 19:07 -!- justanotheruser [~justanoth@unaffiliated/justanotheruser] has joined ##hplusroadmap 19:08 < yashgaroth> still need an enzyme for each possible error, none of which can occur elsewhere either 19:08 < fenn> mismatched dsDNA detection would be a more sensible solution 19:09 < kanzure> this would be: make 10^9 molecules at the same time, finish your synthesis, heat 'em up, anneal them, degrade anything that has significant mismatch, repeat 30 times? 19:10 < yashgaroth> any mismatch, but yes...hopefully one entire fragment is perfect and remains undegraded, and with 10^9 molecules surely there's at least one that survives 19:10 < kanzure> if you're doing inkjetting then the microwells should have piezos or other detectors to complain when a drop was not delivered per expectation (this is another source of error) 19:11 < kanzure> yashgaroth: in that scheme don't you need more than one in the 10^9 that is a match? otherwise your signal is too low and you can't extract any relevant survivors, right? 19:12 < yashgaroth> depends how you do your purification, if you assume every incorrect strand is degraded into dNTPs then you could separate an intact strand with like a microfluidic size exclusion gel; as long as you have one copy survive it's enough to amplify 19:13 < yashgaroth> any DNA larger than a dNTP is assumed to be correct 19:13 < kanzure> ok. well i could imagine some enzymes could be made that help stabilize that sort of degradation outcome. 19:14 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has joined ##hplusroadmap 19:14 -!- anachronick [~kvirc@a85-138-233-38.cpe.netcabo.pt] has joined ##hplusroadmap 19:14 < kanzure> (my general intuition is that you can easily find enzymes that improve things by at least 10% to 300%, after enough search and mutagenesis. and 100x improvements require a lot more search.) 19:14 < yashgaroth> or you have the mismatch recognition enzymes tag the mismatches so they can't amplify, hell make them covalently stick to the strand an inactivate it 19:15 < kanzure> do these enzymes work? in particular i am wondering about how they know that they are doing the alignment correctly. 19:15 < kanzure> you need both strands to be at the same position 19:15 < kanzure> i guess that's the magic of annealing 19:16 < yashgaroth> I think they're pretty good at only recognizing a single base change in an otherwise double-strand, but I haven't looked into em much 19:17 < kanzure> can you get me a ref on this oligo annealing assembly technique 19:18 < yashgaroth> this was the patent I found while doing the oligo assembly stuff https://patents.google.com/patent/US20150329890A9/en 19:18 < kanzure> thank you 19:19 < kanzure> there was also this thing: 19:19 < kanzure> "A systematic comparison of error correction enzymes by next-generation sequencing" http://www.biorxiv.org/content/early/2017/01/15/100685 19:20 < kanzure> but this is correction... not degradation. 19:21 < yashgaroth> there's mismatches, but I suppose the most common error found would be deletions 19:21 < yashgaroth> mismatches are difficult since the enzymes don't know which strand is the correct one 19:22 < kanzure> yes... and if your synthesis technique is sufficiently bad, that might mean 10s of bp at a time go missing 19:22 < kanzure> mismatches should just cause the molecules to be degraded by the enzyme 19:22 < kanzure> or the mismatch-degradation stuff can happen at another step (i could see problems with a one pot one reaction strategy here) 19:23 < yashgaroth> you want it early so you can amplify as early as possible 19:24 < yashgaroth> ideally every imperfect molecule gets destroyed, but the more oligos you attempt to anneal in one pot, the lower the efficiency 19:24 < kanzure> "lower the efficiency" only means "you need (perhaps exponentially?) more cycles of this reaction" right? 19:25 < fenn> the matching oligos can't find each other? 19:25 < fenn> or will anneal with the wrong complementary sequence yield a lot of unnecessary mismatches at a lower temperature 19:25 < fenn> s/yield/yielding/ 19:26 < yashgaroth> mostly the second one 19:26 < fenn> it's the same thing 19:26 < yashgaroth> yeah I suppose so 19:28 < kanzure> long deletions (or additions) can be filtered pretty easily by gel electrophoresis 19:29 < yashgaroth> needs to be a pretty big deletion, you can separate 40bp from 50, but 4950 from 5000, not so much 19:29 < kanzure> a future enzyme that would be helpful to make would be one that degrades short oligos, and another one that only degrades oligos above a certain size. 19:30 < nmz787> can you do mismatched annelaing somehow, where the length of the annealed segment is more important than matching? 19:30 < kanzure> 50 bp deletion *total* is very likely on the scale of 10kmers 19:30 < kanzure> oh you mean.. wait. er. i got this wrong. 19:31 < kanzure> basically i am imagining a really bad synthesis error rate of like >5% 19:31 < yashgaroth> at a certain point the overall matching drives annealing enough to overcome mismatches, but it's rough with indels since they fuck up the strand alignment more 19:33 < kanzure> is it OK for me to assume things like "isolated single bp mismatches sprinkled throughout a large genome are less likely to influence annealing compared to large chunks of mismatches that are all clustered together" ? 19:33 < kanzure> s/large genome/large dna molecule 19:33 < nmz787> seems reasonable 19:33 < nmz787> but if the large chunk fractured off 19:34 < nmz787> then the average annealed % would be higher in the now-broken segments 19:34 < kanzure> 10-100kmers 19:35 < kanzure> alignment seems really weird to me. this is dna physics right? 19:35 < nmz787> s/large chunk/mismatched chunk/ 19:39 < nmz787> there isn't a clear table of bp/nm^3 for different organizational levels (linear fragments, circular, nucleosomes, chromatin, chromosomes, etc) 19:41 < nmz787> "Structural states of the nucleosome that are likely to be interchangeable. These include the tetrasome, which is formed by the wrapping of ~80 bp DNA around an (H3–H4)2 tetramer." 19:42 < kanzure> i'm reasonably happy with an idea like automatic destruction/degradation of mismatched, unaligned and imperfect molecules. 19:43 < kanzure> the errors from phosphoramidite chemistry originate from chemical yield issues across all the molecules randomly, not in aggregate, right? assumin your reagents are good and delivery and environment was good. 19:43 < kanzure> *assuming 19:44 < kanzure> this actually encourages you to synthesize short molecules again, i believe, because if you're synthesizing 10^13 molecules at a time (per spot), you still run into information theoretic problems where there's not even a single molecule likely to be correct as you continue to push its total length upwards 19:44 < kanzure> so you still need to keep the molecules short-- and increase the total number of molecules you're building per spot 19:45 < kanzure> (and then you combine all the spots together at the end and do homologous recombination assembly magic stuff, or whatever other assembly technique you've developed) 19:45 < nmz787> maybe you don't even need to do that, since you want them to be fragments when you sequence them 19:46 < nmz787> nanopore even requires adapters, I guess to attract it to the nanopore to begin 19:46 < kanzure> well i guess it doesn't matter... you can keep pushing the length up, and your destroy-incorrect-molecules scheme should cause the total dna content to drop to zero, in which case you know you need to try again... but this doesn't help because you might get stuck trying a million billion times (due to probability and statistics). 19:46 < nmz787> then you just need an ASIC to assemble the DNA reads 19:46 < kanzure> you're not sequencing them. sequencing is extreme cost. that's why you're doing annealing mismatched molecule destruction things. 19:47 < nmz787> well they need a read for their write 19:47 < kanzure> i'm not convinced. you only need to remove wrong molecules, you don't need to know what their content was. 19:47 < nmz787> I'm just saying, if that's what is required for reading efficiently, and the assembly in-silico is already existing, it is a route 19:48 < nmz787> "optimize the write to your read" was someone's quote 19:48 < kanzure> no that was about developing compatible chemistries if you're reading 19:48 < nmz787> kanzure: that's exactly what I said at the meeting 19:48 < nmz787> kanzure: re only remoiving the offenders 19:48 < kanzure> then why are you always insisting on integrated sequencing sigh 19:49 < nmz787> I specifically said "you don't need to know what the base was, just whether it not it was extended by N requested" 19:49 < nmz787> if you say add, it better have added 19:49 < nmz787> otherwise try again 19:49 < nmz787> (or filter the bad ones) 19:50 < nmz787> I can search for the verbatim transcript 19:50 -!- Gurkenglas [~Gurkengla@dslb-178-000-177-083.178.000.pools.vodafone-ip.de] has quit [Read error: Connection reset by peer] 19:51 < nmz787> " Realistically you don't need to know what the sequence is. You only need to know, "did it get added". You only need length discrimination." 19:51 < nmz787> you typed 19:58 < heath> nmz787: do you have a reference to the tape industry being $500MM/year? the company publishing this wants $2450 for a copy: https://www.persistencemarketresearch.com/market-research/tape-storage-market.asp 19:59 < nmz787> what is MM? 19:59 < nmz787> megamillion? 19:59 < nmz787> I don't think I mentioned industry market $ per year 20:00 < kanzure> MM is a common way to write "million" in the finance industry. 20:02 -!- Gurkenglas [~Gurkengla@dslb-178-000-177-083.178.000.pools.vodafone-ip.de] has joined ##hplusroadmap 20:03 -!- esmerelda [~mabel@71-212-29-109.tukw.qwest.net] has joined ##hplusroadmap 20:04 < nmz787> what do they consider M to be? 20:05 < kanzure> yashgaroth: i think the mismatch-degradation stuff only works if you can guarantee the correct version will be the "highest plurality". and if you can guarantee that it's in there somewhere (1 Gbp would require 4^(1 Gbp choices) molecules at least..) 20:07 < yashgaroth> no way we're getting near a gigabase in one reaction 20:07 < kanzure> i'm just illustrating the problem :P 20:08 < yashgaroth> fair, but we still need to find and extend the limit on how much we can assemble in a single reaction, either free-floating or continuous assembly off a support, then use that as the basis 20:09 < kanzure> if the probability of a correct molecule existing in the batch is is 1/(4^(length)) then you are still length constrained... which was the same thing as the original problem with phosphoramidite chemistry (errors). 20:09 < nmz787> 3 charged sites, doesn't that mean that theorhetically you can only have up to 8 distinct electrical charge discrimination elements for i.e. detecting with nanopore ionic current? 20:10 < yashgaroth> say we have a chunk length of 10kbp where you get a few molecules that are entirely correct, that's a reasonable size where you can implement QC/amplification/storage 20:10 < kanzure> there's no way you have molecules that are correct if you are building 10kbp molecules though :( 20:11 < kanzure> .wam 4^(10000) 20:11 < kanzure> .wa 4^10000 20:11 < yoleaux> kanzure: Sorry, no result! 20:11 < kanzure> .gc 4^10000 20:11 < yoleaux> kanzure: Sorry, that command (.gc) crashed. 20:11 < yashgaroth> yeah that is rather optimistic, but still the error rate isn't nearly that high right? 20:12 < kanzure> whatchamean. am i overestimating the error rate of this chemistry? 20:12 < yashgaroth> hang on I'm bad at math 20:12 < kanzure> wait yes i am also bad at math 20:13 < kanzure> 1/(4^200) (so, you're building a 200 bp molecule) would be 1 / 2582249878086908589655919172003011874329705792829223512830659356540647622016841194629645353280137831435903171972747493376 20:13 < yashgaroth> but like with regular pamidite chemistry the error rate isn't 1/(4^(length)) 20:13 < kanzure> yes i am looking at 2582249878086908589655919172003011874329705792829223512830659356540647622016841194629645353280137831435903171972747493376 and i'm starting to agree with you :-) 20:14 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has quit [Ping timeout: 248 seconds] 20:14 -!- aeiousomething [~aeiousome@gateway/vpn/privateinternetaccess/aeiousomething] has quit [Ping timeout: 248 seconds] 20:15 < kanzure> i think it might be a linear error rate of "1/200"... but what the hell does that mean. 20:15 < heath> .title https://www.hpcwire.com/2017/09/14/darpa-pledges-another-300-million-post-moores-readiness/ 20:15 < yoleaux> 502 Bad Gateway 20:15 < heath> "DARPA Pledges Another $300 Million for Post-Moore’s Readiness" 20:16 < yashgaroth> so we take like .99^(length) 20:16 < kanzure> sure 20:17 < yashgaroth> 1000bp has a reasonable correct fraction of 0.000043, though 10000bp is like 2^-44 20:17 < yashgaroth> err, 2e-44 20:19 < kanzure> 4^length would have to assume you are randomly making molecules, why would i think that. weird. 20:20 < yashgaroth> a thousand monkeys sitting at a thousand synthesizers will eventually make a genome 20:21 < kanzure> 0.000043 is not very encouraging either, fwiw 20:21 < yashgaroth> meh, one in 23255 molecules ain't bad if you have a million of them 20:21 < kanzure> i guess the question is whether you're in a situation where the errors agree with each other more than your 0.000043 agrees with itself 20:22 < kanzure> you have 1e10 to 1e13 of them not just a million 20:22 < yashgaroth> oh sick 20:22 < kanzure> well ok fine maybe a million. 20:22 < yashgaroth> yeah I suppose both strands potentially having a complimentary error could happen, but the fully correct molecules should outnumber them maybe 20:23 -!- aeiousomething [~aeiousome@gateway/vpn/privateinternetaccess/aeiousomething] has joined ##hplusroadmap 20:23 < kanzure> yes, your best outcome is one where you have your 0.000043 that are correct, and then literally all the other molecules are wrong but they are each randomly wrong 20:24 < nmz787> it is .99^N 20:24 < nmz787> where N is length i.e. 200 20:24 < nmz787> 0.13397967485 20:24 < nmz787> correct at end 20:24 < nmz787> % 20:25 < nmz787> oh, nvrmmind 20:25 < nmz787> I skipped a few lines 20:25 < nmz787> after being blinded by BIG HUGE NUMBER 20:27 < nmz787> but that is also open-loop 20:27 < nmz787> close the loop and things change 20:28 < yashgaroth> so your only problem is the (1%^2)*((1/3)^2) chance at each position to have two complimentary errors, which in a 1000bp assembly, 1% will have that? idk 20:28 < kanzure> well maybe with some enzymes 99% can be increased to 99.9%. and then you get some good numbers for 10kbp 20:30 < kanzure> we need a polymer chemist so we can ask him to attach a quencher/fluorophore to a phosphoramidite that causes it to burn for like 200 milliseconds and then we can just watch it with a camera like nmz787 wants 20:34 < kanzure> yashgaroth: if there are two complementary errors on the same molecule, but other random errors, doesn't that qualify those two molecules for destruction in our above scheme? 20:34 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has joined ##hplusroadmap 20:34 < kanzure> whoops "on the same molecule" i meant "on two error molecules" 20:38 < yashgaroth> true, but if we're looking at attrition that applies to otherwise-correct molecules as well, also I'm not good at statistics 20:39 < yashgaroth> oh wait yeah they'd also still have to attach to each other by chance 20:39 < kanzure> 1e18 1kbp ssDNA molecules per gram ... not sure if looking at it in terms of grams is the right way to go here. i think we need... a chart (of ssDNA molecule length vs number of moleculs we can have per gram) 20:41 < yashgaroth> (10^23)/327 molecules per gram, divided by the length 20:42 < yashgaroth> x6.022 20:42 < yashgaroth> oh wait you already did that 20:44 < kanzure> alright chemistry is just downright screwed 20:45 < kanzure> unless you get those magic stabilizing enzymes, i'll admit defeat and say you need a camera monitoring the reaction and just redo if nothing happens, even in an oligo library ligation-extension scheme 20:45 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has quit [Ping timeout: 260 seconds] 20:47 < kanzure> and this limits the total number of molecules you can have to like, one 20:47 < kanzure> also you need a fluorescent event that fires upon reaction at a rate of like 99.9999999999999% 20:48 < kanzure> and it needs to last long enough for your crappy camera to see it 20:58 < kanzure> fenn: what should i tell people my targeted inkjet step size is 21:00 < nmz787> this seems to be the best estimate for realistically starting a big-ass DNA management system https://en.wikipedia.org/wiki/Nucleosome 21:00 < kanzure> i don't thikn you can do better than 200 nm even with microstepping 21:00 < nmz787> "The nucleosome core particle consists of approximately 146 base pairs (bp) of DNA[12] wrapped in 1.67 left-handed superhelical turns around a histone octamer consisting of 2 copies each of the core histones H2A, H2B, H3, and H4.[13] Core particles are connected by stretches of "linker DNA", which can be up to about 80 bp long. Technically, a nucleosome is defined as the core particle plus one of 21:00 < nmz787> these linker regions; however the word is often synonymous with the core particle." 21:01 < nmz787> "The current understanding[26] is that repeating nucleosomes with intervening "linker" DNA form a 10-nm-fiber, described as "beads on a string", and have a packing ratio of about five to ten.[19] A chain of nucleosomes can be arranged in a 30 nm fiber, a compacted structure with a packing ratio of ~50[19] and whose formation is dependent on the presence of the H1 histone." 21:02 < nmz787> so I think 146 + 80 for address per 10 nm cube is OK to start 21:02 < nmz787> and maybe a bit of a fudge factor as you get more length since packing will have some overheard since this isn''t a perfect crystal 21:03 < nmz787> kanzure: re:stepers that's why I'm saying just use direct plumbing and forget the gantry 21:07 < kanzure> you want a screw-top? 21:10 < nmz787> electroosmostic/electrophoresis would work for many things, if not all if using enzymes 21:11 < nmz787> yashgaroth: any info on electrphoresis on proteins causes disruptions after being in the electric field for too long? 21:11 < nmz787> (not touching the electrodes, i.e. running out of a 'gel' though) 21:11 < yashgaroth> nah should be fine 21:12 < nmz787> so moving them around, or moving DNA away from them shouldn't affect their reusability in-vitro 21:12 < nmz787> ;) 21:13 * nmz787 just went to get twinkies and now things are good 21:13 < nmz787> radiohardened GMO-produced twinkie fruit is now on my list 21:14 < nmz787> I shall fulfill the family guy prophecy 21:22 < kanzure> either of you gonna be on the gp-write call tomorrow 21:22 < kanzure> ? 21:23 < yashgaroth> but it's eeeearly 21:23 < kanzure> oh yeah 21:23 < kanzure> well at least comment on the documents 21:25 < nmz787> I will 21:26 < nmz787> which "the documents" 21:33 < kanzure> check inbox, he mentions two for tomorrow 21:34 -!- jtimon [~quassel@199.31.134.37.dynamic.jazztel.es] has quit [Ping timeout: 240 seconds] 21:47 -!- jtimon [~quassel@199.31.134.37.dynamic.jazztel.es] has joined ##hplusroadmap 21:48 < nmz787> hmm, they are pretty thin 21:51 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has joined ##hplusroadmap 22:01 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has quit [Ping timeout: 260 seconds] 22:10 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has joined ##hplusroadmap 22:11 -!- abetusk [~abe@68.175.143.22] has quit [Ping timeout: 240 seconds] 22:20 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has quit [Ping timeout: 255 seconds] 22:21 -!- TC [~talinck@cpe-174-97-113-184.cinci.res.rr.com] has joined ##hplusroadmap 22:21 -!- TC is now known as Guest88555 22:24 -!- abetusk [~abe@68.175.143.22] has joined ##hplusroadmap 22:25 -!- hehelleshin [~talinck@cpe-174-97-113-184.cinci.res.rr.com] has quit [Ping timeout: 255 seconds] 22:25 -!- abetusk is now known as Guest97162 22:28 -!- maaku [~mark@173.234.25.100] has quit [Quit: Lost terminal] 22:33 -!- maaku [~mark@173.234.25.100] has joined ##hplusroadmap 22:37 < nmz787> "The inability to carry out protein synthesis means that no virus can evolve to target mammalian red blood cells." 22:49 -!- yashgaroth [~yashgarot@2606:6000:cd4d:3300:f5e0:f867:a11d:8d52] has quit [Quit: Leaving] 22:58 -!- jtimon [~quassel@199.31.134.37.dynamic.jazztel.es] has quit [Ping timeout: 240 seconds] 23:09 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-hnthemkxyzsyaqxz] has joined ##hplusroadmap 23:11 -!- traumschule [~traumschu@185.104.184.179] has joined ##hplusroadmap 23:16 -!- Gurkenglas [~Gurkengla@dslb-178-000-177-083.178.000.pools.vodafone-ip.de] has quit [Ping timeout: 240 seconds] 23:32 < fenn> i think dot density is a losing game, rather you want to move faster and have more precise drop timing 23:32 < fenn> reel to reel printing substrate can give you huge amounts of surface area to print upon 23:33 < fenn> 600 dpi is a conservative dot density, that's droplets 42 microns apart 23:33 < fenn> nanometers are too hard to think about 23:38 < fenn> "for wide format Epson that magic number is 360 ppi" 23:38 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has joined ##hplusroadmap 23:45 -!- delinquentme [~delinquen@108-235-112-153.lightspeed.sntcca.sbcglobal.net] has quit [Ping timeout: 240 seconds] --- Log closed Wed Sep 20 00:00:23 2017