--- Log opened Mon Jul 13 00:00:09 2015
00:57 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has quit [Ping timeout: 250 seconds]
01:56 -!- Porb_ [~Porbus@CPE-124-176-219-129.lns5.win.bigpond.net.au] has quit [Read error: Connection reset by peer]
02:33 -!- Viper168_ [~Viper@unaffiliated/viper168] has joined ##hplusroadmap
02:35 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 250 seconds]
02:37 -!- Viper168_ is now known as Viper168
02:55 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap
03:08 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has joined ##hplusroadmap
03:24 -!- soylentbomb [~k@unaffiliated/soylentbomb] has quit [Quit: leaving]
03:32 -!- zadock [~outsider@cthulhu.tuiasi.ro] has joined ##hplusroadmap
03:34 -!- zadock [~outsider@cthulhu.tuiasi.ro] has quit [Remote host closed the connection]
03:38 -!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-vqbxlinqdnrfvdvb] has quit [Quit: Connection closed for inactivity]
03:43 -!- drewbot_ [~cinch@ec2-54-226-221-122.compute-1.amazonaws.com] has joined ##hplusroadmap
03:49 -!- drewbot [~cinch@ec2-54-211-61-160.compute-1.amazonaws.com] has quit [Ping timeout: 256 seconds]
03:58 -!- knobuddy is now known as ghostcaptain
04:06 -!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-duyxyyetgupmyrla] has joined ##hplusroadmap
04:22 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Ex-Chat]
04:23 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap
05:30 < dpk> .privacy
05:30 < yoleaux> dpk: This channel is public. When I am asked when I last saw you, I may repeat things you say and what time it was when you said them.
05:30 < ghostcaptain> ʘ_ʘ
05:33 -!- Porb [~Porbus@CPE-124-176-219-129.lns5.win.bigpond.net.au] has joined ##hplusroadmap
05:34 < fenn> ghostcaptain: if that was a surprise, may i humbly submit that you please read the channel topic
05:40 -!- Porb [~Porbus@CPE-124-176-219-129.lns5.win.bigpond.net.au] has quit [Ping timeout: 265 seconds]
05:41 -!- Porb [~Porbus@CPE-124-176-219-129.lns5.win.bigpond.net.au] has joined ##hplusroadmap
05:46 -!- eudoxia [~eudoxia@r167-56-36-161.dialup.adsl.anteldata.net.uy] has joined ##hplusroadmap
06:01 < ghostcaptain> that topic is a mess
06:01 < ghostcaptain> plus i forgot
06:01 < ghostcaptain> ^_^
06:02 < streety> probably best to assume every channel you are on is logged by someone
06:03 < fenn> best to assume all traffic everywhere is logged by someone
06:03 < chris_99> mmm
06:03 < fenn> but this channel is google-searchable
06:04 < ghostcaptain> i know i'm logging
06:05 < streety> poorly searchable
06:06 < fenn> if you want better searchability you can do this and then grep:
06:06 < fenn> wget -r -np -nc  --exclude-dir="logs/html/,logs/hplusroadmap-logs/" --reject "2008-03-25.log" gnusha.org/logs/
06:06 < chris_99> heh, why reject that one out of interest?
06:06 < fenn> because it changes every time and it's dumb to download it when it's just a duplicate of the individual log files
06:07 < chris_99> ah
06:10 < streety> thanks
06:10 < fenn> note that this will give a bad file for the day you downloaded it because -nc doesn't try to re-download files that exist (use -c for that, which is slower)
06:11 < fenn> i mean the next time you download logs, the "head" from last time will be incomplete
06:11 < fenn> so you could just delete the last log file before downloading
06:15 -!- ebowden [~ebowden@101.177.244.210] has quit [Remote host closed the connection]
06:17 -!- proofoflogic [uid65184@gateway/web/irccloud.com/x-trgyjteenmtzpizq] has quit [Quit: Connection closed for inactivity]
06:33 -!- nottimschmidt [~timschmid@35.10.220.83] has joined ##hplusroadmap
06:34 -!- nottimschmidt [~timschmid@35.10.220.83] has left ##hplusroadmap []
06:35 -!- nottimschmidt [~timschmid@35.10.220.83] has joined ##hplusroadmap
06:35 -!- nottimschmidt [~timschmid@35.10.220.83] has quit [Client Quit]
06:35 -!- nottimschmidt [~timschmid@35.10.220.83] has joined ##hplusroadmap
06:36 -!- Madplatypus [uid19957@gateway/web/irccloud.com/x-ckarasonpysmdemc] has quit [Quit: Connection closed for inactivity]
06:46 -!- ebowden [~ebowden@101.177.244.210] has joined ##hplusroadmap
06:48 < kanzure> i guess we could throw up a search form and an index, like solr or something
06:48 < kanzure> but then i would feel bad about not indexing special characters
06:51 < kanzure> CaptHindsight: airlocks are important
06:52 -!- ebowden [~ebowden@101.177.244.210] has quit [Ping timeout: 248 seconds]
06:53 < kanzure> so who here wants to house the dna synthesizer once it's complete? any takers.
06:55 -!- JayDugger [~jwdugger@108.19.186.58] has joined ##hplusroadmap
07:04 < ParahSailin> this one labs quote?
07:05 < kanzure> yes
07:06 < kanzure> also- i think the air jet dryer thingy will blow away the dna unless it's tied to the surface. so that's more chemistry that needs to be figured out.
07:06 < ParahSailin> what likelihood would you assign to it working
07:06 < kanzure> motion control elements, 100% likely
07:06 < kanzure> 100 bp strands per spot, no clue..
07:06 < fenn> mumble mumble acetonitrile toxicity flammability
07:07 < fenn> mumble large heavy fragile expensive object that needs to be treated carefully
07:07 < kanzure> well the primary reason for the size is the number of spots, i think
07:08 < kanzure> oh plus whatever form factor we can purchase argon in
07:08 < ParahSailin> 60cm travel in one dimension?
07:08 < fenn> the size is mostly because it's made on an optical breadboard i think
07:09 < kanzure> this isn't going to happen if i can't find somewhere to put it
07:10 < ParahSailin> can put it here
07:10 < fenn> going once
07:10 < fenn> twice
07:10  * nottimschmidt is asking about storing it in lab space at Michigan State University and Washington University.
07:11 < nottimschmidt> Genspace might be interested in NYC
07:11 < kanzure> genspace hates us
07:11 < nottimschmidt> I see :-/
07:11 < kanzure> but nice idea
07:12 < kanzure> posam (another inkjet dna microarrayer thingy) is actually housed in washington in the hood laboratory
07:13 < kanzure> https://www.systemsbiology.org/bio/leroy-hood/
07:13 < kanzure> http://techdev.systemsbiology.net/posam/
07:13 < nottimschmidt> I know a postdoc there, in the Kerr and Klavins labs
07:13 < ParahSailin> you going with this instead of cmos because its cheaper?
07:13 < nottimschmidt> https://twitter.com/LuisZaman
07:13 < kanzure> when you say cmos what do you mean
07:13 < nottimschmidt> He does 3D printing and microfluidics stuff
07:14 < ParahSailin> electrochemical array
07:14 < kanzure> i believe the electrochemical cmos array thingy still used inkjet stuff- it just had electrodes at each spot for generating acid or something
07:14 < ParahSailin> the thing azco sells?
07:15 < ParahSailin> pretty sure no inkjets involved
07:15 < kanzure> oh was this the light valve one? with LCDs or DMDs?
07:16 < fenn> there was a paper about photoelectrochemical synthesis and another one about just electrochemical synthesis
07:16 < fenn> they used the same reaction mechanism
07:17 < ParahSailin> CustomArray
07:17 < ParahSailin> no optics, just electronics
07:18 < kanzure> "My only other comments are that this looks like a great V1 machine.  LinuxCNC as a requirement is a barrier to entry - because it necessitates a dedicated control computer.  And a reduced-cost and reduced-wiring-complexity v2 might strive toward single-board electronics"
07:18 < kanzure> "FPGAs are similarly a barrier to entry, due to the required non-free tooling, and increased cost.  But those barriers are actively being removed at the moment, at least for certain FPGAs.
07:18 < kanzure> https://wiki.debian.org/FPGA "
07:18 < kanzure> ParahSailin: i wasn't aware that they were using *only* electrodes to control the chemistry. how do they isolate which strands get capped/uncapped?
07:19 < fenn> the benefit of FPGA is higher performance/throughput and it enables servo motors instead of stepper motors
07:19 < ParahSailin> addressing each electrode individually
07:19 < kanzure> but.. so .. the claim is that each electrode is able to control the capping/uncapping of a long oligo?
07:20 < ParahSailin> its not a claim, its been a commercial product for almost 10 years
07:20 < fenn> i'm not sure either of these features are necessary for a proof of concept, but it's not a super huge cost relative to the overall quoted cost of the system
07:20 < ParahSailin> you can even order a test run from azco for a couple k, to see if the error rate is good enough
07:20 < nottimschmidt> fenn: servos can be done in software, on CPUs too.  LinuxCNC does that, for instance.  Nothing special about FPGAs enables them.
07:21 < fenn> nottimschmidt: no, in fact you will not get good results doing servo control in software without some kind of dedicated interface card to watch the optical encoders
07:21 < nottimschmidt> In CPUs, these are called "external interrupts"
07:22 < fenn> even a dedicated microcontroller can not keep up with a servo motor with common encoder resolutions and speeds
07:22 < fenn> there is usually some kind of hardware counter to add up the encoder transitions since last interrupt
07:23 < nottimschmidt> One servo per, little interrupt handler to dispatch control I/O, bob's your uncle.  You want to be running a realtime OS, but something like http://bertos.org/ manages this on AVR, ARM, and x86 with a sane interface.
07:23 < ParahSailin> even less than a couple k https://azcobiotech.com/oligo-array-12k-pool-10-to-130-mer-in-length-80-0012.html
07:23 < fenn> that only works for low resolutions and/or slow speeds, which negates the point in using servo motors in the first place
07:23 < nottimschmidt> Certainly speed is affected by the choice of solution.  Have you done the math on what speed and resolution you need?
07:24 < fenn> modern PC CPUs are terrible at realtime interrupts
07:24 < kanzure> ParahSailin: eventually i want to look into the physics or chemistry of that reaction mechanism, because i was not aware that it was electrode-only control....
07:24 < ParahSailin> isnt that just an os issue?
07:24 < fenn> mumble mumble pipeline context switch mumble
07:25 < ParahSailin> i think its this http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1762329/
07:25 < kanzure> nottimschmidt: well, we would like to have millions of spots- which would get us pretty close to the size of a genome
07:25 < fenn> i don't know the specifics at a low level but i do know that people more experienced than me all agree that it's impossible
07:25 < kanzure> .title
07:25 -!- Porb [~Porbus@CPE-124-176-219-129.lns5.win.bigpond.net.au] has quit [Quit: Leaving]
07:25 < yoleaux> Electrochemically Generated Acid and Its Containment to 100 Micron Reaction Areas for the Production of DNA Microarrays
07:25 -!- nsh [~lol@wikipedia/nsh] has quit [Ping timeout: 246 seconds]
07:25 < ParahSailin> ask a cmos custom prototyper for a quote on that apparatus
07:26 < nottimschmidt> fenn: I've written / heavily modified some (admittedly slow, at 16Mhz) realtime motor control interrupt handlers.  Seems very possible to me.
07:26 < kanzure> i guess fenn is right, here, i was probably mistakening this for the "photolithography dna synthesis" method that used photogenerated acid or something for one of the steps, or electrode-generated acid for one of the steps.
07:26 < kanzure> ParahSailin: how are the strands tied to the electrode?
07:26 < fenn> nottimschmidt: was 16MHz the speed of the cpu or the speed of the encoder transitions?
07:27 < ParahSailin> photolithography with chrome masks is how affy is stuck doing it
07:27 < nottimschmidt> fenn: cpu speed
07:27 < kanzure> ParahSailin: *masked* photolithography! ouch. what the fuck?
07:27 < fenn> nottimschmidt: roughly how fast did the encoders go?
07:28 < nottimschmidt> they topped out around 40,000 pulses per second
07:28 < kanzure> not bad; i think this piezo printhead does 16 kHz drops
07:29 < ParahSailin> cleavable linker http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1762329/figure/pone-0000034-g006/
07:29 -!- nsh [~lol@wikipedia/nsh] has joined ##hplusroadmap
07:29 < nottimschmidt> I think you could do a lot more with a 100Mhz ARM M3 let alone something cell-phone class.
07:29 < kanzure> right... but i mean the initial placement of the linker.
07:29 < fenn> nottimschmidt: ok that's about the max rate for outputting stepper motor pulses from a parallel port
07:30 < nottimschmidt> yeah, parallel ports are not great
07:30 -!- nottimschmidt [~timschmid@35.10.220.83] has quit [Remote host closed the connection]
07:30 < kanzure> he.. died?
07:30 < fenn> not sure what just happened
07:32 < kanzure> so wouldn't it be better to isolate each reaction in a drop instead ofj ust hoping the acids don't brownian motion themselves over to the other electrode areas
07:32 < ParahSailin> some sort of a reactive chemical reagent?
07:33 < ParahSailin> kanzure: which is why i suggest spending the $1650 on an azco array run, and then a few k more on a nextseq run to see how the error rate is
07:34 < fenn> can't you just ask them what the error rate is
07:34 < fenn> surely someone has quantified this already
07:34 < ParahSailin> well, the application for these is generally just as hybridization probes
07:35 < kanzure> i also wonder if individual dna could be harvested from that approach (oops i mean specific sequences); because i think it's all in the same water bath
07:35 < ParahSailin> like stick biotin on your whole pool and do X
07:35 < kanzure> or er, not water...
07:35 < fenn> since they're bound to the surface (right?) you could 3d print wells around them or something, assuming they give you the plate
07:36 < ParahSailin> "In contrast, in the electrochemical case reagent generated by the electrode is able to diffuse into the PRL but is prevented from excessive diffusion by the presence of a buffering agent in the surrounding solution."
07:36 < kanzure> then how does it migrate to the tip of the oligo?
07:36 < ParahSailin> so i think they figured out a solution
07:36 < ParahSailin> " Near the electrode, the buffer is overwhelmed, allowing reaction of the ECG reagent with the PRL bound substrate; but further away, the concentration of the diffusing reagent is insufficient to overcome the concentration of the buffering agent, and no reaction occurs."
07:37 < fenn> the oligo is only ... (0.2nm * number of base pairs) long
07:37 < kanzure> i suppose the oligo is very short- oh fenn is on it
07:37 < fenn> i think stereolithography/DLP 3d printing works the same way
07:37 -!- nottimschmidt [~timschmid@35.10.220.177] has joined ##hplusroadmap
07:37 < fenn> wb nottimschmidt
07:38 < nottimschmidt> thanks, sorry, power adapter issues
07:38 < ParahSailin> get a bouncer
07:38 < nottimschmidt> Will do
07:39 < ParahSailin> microsoft gives away $750/mo in azure to practically anyone
07:39 < nottimschmidt> I've got a cloudatcost machine already running some things.  Will set it up there.
07:41 < kanzure> so i agree that an electrode microarray would be a good and sane thing to work with in the near future
07:41 < kanzure> although, an inkjet plus motion control is also useful for other reasons
07:41 < kanzure> and can possibly be coerced into higher densities
07:42 < nottimschmidt> So anyway, FPGAs can be a great solution.  http://papilio.cc/ and other low-cost FPGA boards have been starting to grow a larger community of knowledgeable folk.  And they can run AVR and ARM soft cores for offloading some of the realtime processing, just like a microcontroller.
07:43 < nottimschmidt> I'm really happy to see some Free tools start to show up for some of these boards.
07:43 < nottimschmidt> They're still early days, though
07:43 < ParahSailin> kanzure: it seems like you have a bias to have something that anyone can build themselves at home, whereas you should go with what is actually the viable technology
07:44 < CaptHindsight> the whole PC vs uController thing will cost you more money, there isn't a single board microcontroller board that can handle the closed loop and printhead control on the market
07:44 < CaptHindsight> you'll have to reinvent that wheel
07:44 < kanzure> CaptHindsight: i misunderstood the electrode acid approach, it's a microelectrode array and no LCD or DMD or inkjet
07:45 < nottimschmidt> CaptHindsight: if I can do it (with some speed limitations) on a 16Mhz AVR (and I have), what makes you say that no uC board on the market can do it?
07:46 < kanzure> ParahSailin: well i'm biased towards the inkjet design because we have someone that is willing to build it
07:46 < CaptHindsight> nottimschmidt: I know that the Linuxcnc works well. Try and do it with your AVR
07:46 < ParahSailin> kanzure: and thats only because you emailed them before you emailed a cmos prototype shop
07:47 < nottimschmidt> CaptHindsight: I hack CNC firmwares for the AVR.
07:48 < kanzure> nottimschmidt: what's wrong with linuxcnc anyway? your argument is that it's "bulky" right?
07:48 < CaptHindsight> kanzure: have a link to clarify the "electrode acid approach"?
07:49 < ParahSailin> scroll
07:50 < CaptHindsight> http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1762329/  ? lots of noise in here today
07:51 < nottimschmidt> kanzure: it shoves a lot of the realtime work onto the big desktop CPU, precisely where we all agree it shouldn't be.  That necessitates running a realtime OS on a machine you interact directly with, which means external interrupts are being triggered not just by the machine, but by the unpredictable user.  The whole thing requires additional hardware over a uC or self-hosted FPGA-based solution, setup is more complex, ladder logic isn't wi
07:51 < nottimschmidt> dely understood, etc etc.
07:52 < nottimschmidt> LinuxCNC works and is very flexible.  Hence my comment about this design being a good v1 machine.
07:54 < juri_> ok, i've caught up.
07:55 < juri_> I'd be coordinating a batch of propeller MCUs, but the FPGA chosen for this project is my personal favorite.
07:56 < juri_> it's supported by a free software toolchain, as well. we could load a MCU into it, and make it appear to the PC as a serial port accepting GCODE.
07:57 < nottimschmidt> juri_: that's awesome
07:57 < juri_> we could even use pronterface for delivering the gcode.
07:58 < CaptHindsight> I thought we'd go the electrode acid approach next
07:58 < nottimschmidt> So long as Free tools work well with it, and cost isn't multiple orders of magnitude different, the FPGA sounds like the best solution.
07:58 < juri_> linuxcnc might get us across the finish line faster, if we approach the problem from the right angle, break it into jobs we can test seperately, and apply all the brainpower we have.
07:58 < nottimschmidt> Heat and power consumption aren't likely to be issues
08:00 < CaptHindsight> Linuxcnc is guaranteed to work, then you can't use the working example to make whatever else you'd like
08:01 < CaptHindsight> BBB, smoothie, parallel propellers, stm32 etc etc
08:01 < CaptHindsight> sorry can't/can
08:01 < nottimschmidt> juri_: I'm not wedded to gcode either, there are other machine control languages like OpenSBP, and I'm sure some genomics-specific domain specific languages.  Open to ideas.
08:02 < juri_> I think gcode is the right tool. it's just a question of how you use it.
08:02 < nottimschmidt> k
08:02 < juri_> Honestly, i think you should be modeling the machine more like a 3d printer in how you speak to it.
08:03 < juri_> this makes the "slicer" more complicated, but removes a lot of the realtime interaction's complexity during the debugging stage.
08:04 < ParahSailin> kanzure: and if you figured out how to electrochemically cleave the linker, then you have a way to suck up individual oligos
08:04 < nottimschmidt> Yes.  Doing as much work ahead of time as possible, in the "slicer" is an excellent idea.
08:04 < CaptHindsight> I think we should make more tools so that people can easily make things like the CMOS array
08:05 < juri_> I don't think the gcode interpreter should have any idea what the (x,y) offsets of the nozzles are.
08:06 < juri_> we should treat them as extruders. "go to position XYZ" "go from that position to this new position while firing a dropplet from nozzles (42-85)"... etc.
08:07 < juri_> complicated "slicer", but simple gcode interpreter.
08:07 < kanzure> i'm confused about the eat-gcode-by-serial-port thing, is it the fpga that is interpreting the gcode? why... gcode interpreters suck, why not just interpret it before sending data to the equipment?
08:07 < fenn> in the current design the gcode doesn't say anything about nozzles at all, that's handled by a preprocessor (that has yet to be written)
08:07 < kanzure> yes i'm not sure the optimal gcode modification method here. still looking at linuxcnc docs.
08:07 < kanzure> i understand the HAL stuff, but not how it interfaces with gcode stuff.
08:08 < kanzure> ParahSailin: we can still use electrode arrays with this approach
08:08 < juri_> kanzure: that's prospective, and down the line. i think implementing in linuxcnc is probably faster.
08:08 < nottimschmidt> kanzure: the FPGA would do the interpreting, yes.  Reason is that the machine control signals happen in real-time which desktops suck at.
08:08 < nottimschmidt> Desktops want to do large batch job type things.
08:08 < ParahSailin> sounds like a lot of reduplicated effort
08:08 < kanzure> nottimschmidt: so who is going to write the vhdl for a gcode interpreter? isn't that a fate worse than death
08:08 < fenn> i'm thinking of it like the print head nozzle firing is slaved to the movement; the fpga fires the nozzles at the correct position but otherwise has no control over the machine
08:09 < nottimschmidt> kanzure: soft-AVR core on the FPGA w/ a firmware?
08:09 < CaptHindsight> fenn: thats the approach I'd recommend
08:09 < juri_> kanzure: you don't. you write a gcode interpreter for an MCU, then upload that MCU to the fpga.. again, in an ideal world. ;)
08:09 < kanzure> so the fpga is not moving the gantry. ok.
08:09 < CaptHindsight> that is how most inkjet printers operate
08:09 < nottimschmidt> kanzure: it's better than it sounds.  Opencores has a 100Mhz soft-avr core.
08:10 < kanzure> nottimschmidt: i have not looked at opencores in a while, that's neat.
08:10 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Ping timeout: 252 seconds]
08:10 < juri_> nottimschmidt: the propellers get 200MIPS at 32Mhz. 8 core microcontroller. crazy design. ;)
08:10 < nottimschmidt> Yes.  I love the prop.
08:10 < kanzure> ParahSailin: you still need nozzles, you still need a neutral atmosphere, it's a lot of similar equipment
08:11 < fenn> it's sort of hard to talk about because the fpga is performing multiple functions in parallel; it's also watching the servo encoders and generating motor control signals etc
08:11 < nottimschmidt> juri_: also love that it was designed for production on metalized gate array.
08:11 < kanzure> fenn: ah i see.
08:11 < ParahSailin> kanzure: actually you dont even need to go to a cmos place https://azcobiotech.com/oligo-array-12k-chip-cleavable-19-5012.html
08:11 < ParahSailin> just buy it and breadboard some shit around it
08:11 < kanzure> isn't this the same link
08:12 < ParahSailin> nah, this is a "consumable" cmos array
08:12 < fenn> the thing that fires the nozzles can see what position the motors are at, but has no control over them
08:12 < ParahSailin> this is much cheaper than the quote you got
08:12 < fenn> but this "thing" only exists as a pattern of activated logic gate cells
08:12 < CaptHindsight> yes the printhead just fires as it get gated encoder pulses
08:13 < kanzure> ParahSailin: btw CaptHindsight made the quote
08:13 < juri_> CaptHindsight: you're not wanting to control motion from the same FPGA?
08:13 < CaptHindsight> yeah, we looked at those 12k chips last week
08:14 < fenn> FPGA only does motion interpolation at a low level
08:14 < fenn> like linear interpolation (maybe not even circular interpolation)
08:14 < kanzure> ParahSailin: you make some good arguments.
08:15 < CaptHindsight> juri_: I want it to work in a few weeks. I'm not spending any time on reinventing the wheel on the first rev
08:15 < juri_> CaptHindsight: so, describe the motion control to me.
08:16 < ParahSailin> kanzure: so the steps i would do is order a run done on one of these customarray chips, send it to ucdavis to run on a miseq for another 1500, then if satisfactory error rate, get a 12k chip and design electronics and fluid handling so you dont have to buy the official 300k machine
08:16 < CaptHindsight> juri_: if you want to spend the time stuffing it all the FPGA you will be able to
08:16 < CaptHindsight> it's an open design
08:16 < juri_> because i'm really concerned that if you try to expose the 'blast a bitmap' code to the CPU in realtime, you'll lose those weeks to testing and defects.
08:17 < kanzure> ParahSailin: send the array to ucdavis?
08:17 < ParahSailin> any cal students could get a $500 discount from davis it looks like
08:17 < kanzure> or send the cleaved oligos
08:17 < ParahSailin> kanzure: cleaved oligos
08:17 < kanzure> this seems to claim to have a cleaving function. hm.
08:17 < CaptHindsight> juri_: the FPGA on the Mesa card is about 10% full with the 3 servos
08:17 < ParahSailin> azco will send a tube of pooled, cleaved oligos
08:18 < kanzure> oh
08:18 < ParahSailin> you just pass it on to davis
08:18 < CaptHindsight> so we have enough room left to control the printhead with some buffering
08:18 < fenn> CaptHindsight: it's been a while and i forget the details.. does mesa's hotmot thing do PID on the FPGA or in the PC CPU?
08:18 < CaptHindsight> it's how we have been making printhead controllers for years
08:18 < fenn> hostmot*
08:18 < CaptHindsight> with the FPGA it's in the FPGA
08:18 < juri_> CaptHindsight: ah, that's fine.
08:19 -!- nmz787_i [ntmccork@nat/intel/x-nxxygqsseduszlkg] has joined ##hplusroadmap
08:19 < CaptHindsight> juri_: if we only had 10% of the gates left I wouldn't recommend this or what to spend the time hand packing gates
08:20 < CaptHindsight> what/want
08:20 < juri_> CaptHindsight: if you run out, you can always just add another FPGA board, and tie the I/O between the two boards.
08:21 < juri_> that makes it a compiler problem, nothing more.
08:21 < nmz787_i> CaptHindsight: I saw your quote, you listed solenoids... but I couldn't tell what for... what are they hooked up to?
08:21 < CaptHindsight> we have large FPGA cards that are Linuxcnc compatible and we also can plug in multiple FPGA cards
08:22 < juri_> CaptHindsight: i like this model. stick to it. ;)
08:22 < CaptHindsight> I have systems with 7 PCIe FPGA cards on one motherboard
08:22 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap
08:22 < CaptHindsight> running multiple printheads with 100x the Epson nozzles
08:23 < juri_> Impressive.
08:23 < nottimschmidt> yeah, very cool
08:24 < CaptHindsight> http://global.kyocera.com/prdct/printing-devices/inkjet-printheads/  think about trying to supply data to 16 or 32 of these
08:24 < fenn> how do you get the bitmap data into the fpga at the right time?
08:24 < kanzure> HAL
08:24 < fenn> presumably there's some sort of buffer somewhere
08:24 < juri_> I seriously reccomend you just encode it into the gcode.
08:24 < CaptHindsight> >100 million drops per printhead per second
08:24 < juri_> it will make your world much easier to debuh.
08:24 < fenn> i recommend against that because of how linuxcnc passes g-code around internally
08:25 -!- ghostcaptain [~IceChat9@unaffiliated/knobuddy] has quit [Remote host closed the connection]
08:25 < juri_> fenn: howso?
08:25 < kanzure> the one major argument i have against not-in-the-gcode is that then you're not /really/ using gcode to talk with the machine. but this is only a weak argument.
08:26 < juri_> it'll make simulation harder.
08:26 < juri_> and debugging.
08:26 < kanzure> fenn: i think the fpga reports encoder stuff back to the pc
08:26 < fenn> well it generates a message for each g-code command and puts it on a queue and then something else takes things off the queue and looks at them and determines their C++ type in order to figure out what kind of message it is and then passes it to another thing that figures out what to do with this particular message type... it ends up generating a lot of work if you do it literally a million times
08:26 < fenn> a second
08:26 < kanzure> yes i noticed that linuxcnc seems to pass around gcode by stdout? it was amusing
08:26 < nottimschmidt> I implemented fast (in the interrupt handler) bitmap raster support for lasers in Marlin, and encoded the bitmap data into the gcode.  Works a treat.  Single "nozzle" though.
08:27 < fenn> kanzure: if it's using stdout it's not technically part of emc but rather some preprocessor script
08:27 < kanzure> ah okay
08:27 < CaptHindsight> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?EMC_Components  Linuxcnc with block diagrams of how it works
08:27 < ParahSailin> i give up
08:27 < nottimschmidt> we base64-endoded the bitmap data, and shoveled it in with a custom G code.
08:28 < juri_> not a bad approach.
08:28 < fenn> ParahSailin: what are you giving up?
08:29 < nmz787_i> nottimschmidt: x86 I/O that isn't pipelined is extremely laggy
08:29 < nottimschmidt> nmz787_i: yep
08:29 < CaptHindsight> using the FPGA's we can do the PID in the FPGA, there are Ethernet FPGA cards that only need a 1mS loop vs few uS loop in the PC
08:30 < CaptHindsight> and do it over ethernet
08:30 < kanzure> fenn: he's giving up on trying to convince me to use a microelectrode array
08:30 < nottimschmidt> nmz787_i: deep caches, pipelines, competing threads, boatloads of I/O, and non-realtime OSes all contribute to this problem.  That's why I prefer to offload g-code processing to a uC or soft-core in an FPGA running in real time.
08:30 < nmz787_i> kanzure: you didn't realize that was electrode-only?
08:30 < nottimschmidt> Keeps the realtime code small and fast.
08:31 < fenn> if you offload g-code to a uC you dont really need a pc at all
08:31 < nottimschmidt> fenn: added benefit
08:31 < nmz787_i> nottimschmidt: it's mostly the lag to come out of (kernel mode, I think) and go into (system, I think)... regardless of OS... just thinking of getting something from the CPU to the outside
08:32 < fenn> part of why people are resistant to using pc-based controllers in the first place i guess
08:32 < nmz787_i> it's a very x86 specific reason, at the bare-metal layer
08:32 < kanzure> nmz787_i: correct, i did not realize
08:32 < nottimschmidt> nmz787_i: there's nothing specific about x86 that causes this problem.
08:32 < nmz787_i> that's why the galileo/edison can only do a few 10s or 100s of kilohertz busy-loop GPIO toggling
08:32 < nottimschmidt> Attempting to do same on a big multi-core ARM would be just as bad.
08:32 < kanzure> fenn: linux everywhere is pretty nice....
08:33 < nmz787_i> nottimschmidt: it definitely is, I've talked to experts
08:33 < fenn> nmz787_i: x86 io involves the ISA bus which has extremely slow timing for legacy reasons... 1ms round trip i think
08:33 < nottimschmidt> nmz787_i: experts, eh?
08:33 < nmz787_i> fenn: does a modern x86 even have ISA?
08:33 < fenn> there may be no actual ISA bus anywhere outside of the chip, but it's still used *shrug*
08:34 < fenn> this is what i was told several years ago and it's probably still true
08:34 < juri_> for some parallel ports. this is PCIE signaling. i think that problem is now dead.
08:34 < nottimschmidt> fenn: ISA doesn't exist any more in most x86 chips.  The system management bus is logically compatible, but implemented on top of I2C
08:34 -!- Netsplit *.net <-> *.split quits: rigel, padz, dpk, mf1008, BobaMa, Betawolf, sh
08:34 < nottimschmidt> or SPI, forget.
08:35 -!- Netsplit *.net <-> *.split quits: heath, EnabrinTain, eleitl
08:35 -!- Netsplit over, joins: heath
08:35 -!- Netsplit over, joins: padz
08:35 < juri_> plus, linux has come up with some solutions to the X86 slowness too. kernel threads, for instance.
08:35 -!- Netsplit over, joins: Betawolf, eleitl, sh, mf1008
08:35 < juri_> <--- linux kernel hacker
08:36 -!- BobaMa [bobama@kapsi.fi] has joined ##hplusroadmap
08:36 -!- Netsplit over, joins: dpk
08:36 < fenn> juri_: any insight as to why linuxcnc still doesn't run on a beagleboard after all these years?
08:37 < nottimschmidt> fenn: I'm told there's some trouble with the beaglebone's GPIO implementation being done through sysfs
08:37 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 264 seconds]
08:37 < nottimschmidt> It's possible to flip GPIOs without going through the FS layer, by peeking and poking various memory locations, but those change from kernel version to kernel version
08:38 < juri_> no good answer. i don't use it. it's based on the realtimelinux patches which disable ACPI (even on a realtime compliant COREBOOT system), so i'd start there...
08:38 < nottimschmidt> So far, no one's implemented a library to do it the fast way
08:39 < fenn> nottimschmidt: that's... not the answer i expected at all
08:40 < nmz787_i> CaptHindsight: I saw your quote, you listed solenoids... but I couldn't tell what for... what are they hooked up to? (Do I have to read through that POSAM paper to decipher the parts list?)
08:40 < CaptHindsight> juri_: http://openlunchbox.com/open-sbc/  imx6 with FPGA on the same board
08:40 < juri_> which FPGA? ;)
08:40 < nmz787_i> agreed --> 07:46 < ParahSailin> kanzure: and thats only because you emailed them before you emailed a cmos prototype shop
08:40 < CaptHindsight> nmz787: drying gas and reagents
08:41 < CaptHindsight> juri_: something that Mesa supports so that it can use all their open cores
08:42 < nmz787_i> CaptHindsight: please explain a bit more, are you using them as single-stroke piston pumps? valves?
08:42 < fenn> seems sort of backward to pick your open hardware to be able use existing semi-closed source software
08:42 < CaptHindsight> nmz787: it's the POSaM design
08:42 < nmz787_i> all reagents in the ABI 391 are under pressure, and valves open/close to let them around the machine
08:42 < juri_> CaptHindsight: I'm only interested in the xilinx spartan 6 slx9 or slx45. ;)
08:43 < fenn> juri plz pester the lunchbox people about this issue
08:43 < CaptHindsight> juri_: http://www.mesanet.com/fpgacardinfo.html
08:44 < juri_> yea, i have to judge from the jpegs, but it looks like they are using the right stuff.
08:44 < CaptHindsight> fenn: I'll use whatever hardware people pay for on the openlunchbox designs
08:44 < fenn> huh i thought it was supposed to have an fpga on board
08:44 < fenn> maybe i am mixing it up with some other open hardware laptop thingy
08:44 < fenn> novena?
08:45 < CaptHindsight> yeah, novena
08:45 -!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-duyxyyetgupmyrla] has quit []
08:45 < fenn> The general purpose FPGA on the Novena mainboard is a Xilinx Spartan6 XC6SLX45
08:45 < nmz787_i> x86 will be soon seen with FPGAs on-chip
08:45 < chris_99> from intel?
08:45 < CaptHindsight> I offered to make the whole laptop but there isn't much interest outside a few paranoid or grey market folks
08:46 < fenn> really there isn't interest in an open laptop design?
08:46 < fenn> seems hard to believe
08:46 < kanzure> lkcl wants an open laptop, he dumped like $100k into some production runs
08:46 < CaptHindsight> actually had some people talk about evading law enforcement with the designs in IRC
08:47 < juri_> there's interest, but free hardware/software hackers are also cheap by necessity.
08:47 < fenn> if "evading law enforcement" means not having any factory-installed backdoors, i'm fine with that
08:47 < CaptHindsight> fenn: well it's up to AMD to reopen the AGESA firmware for the recent chips
08:47 < CaptHindsight> fenn: they were talking about smuggling
08:48 < fenn> smuggling what?
08:48 < fenn> actually nevermind, i don't care
08:48 < CaptHindsight> I tried to work with lkcl a few years ago, no thanks
08:49 < kanzure> nmz787_i: show me a single-nucleotide pump, or single-molecule pump from the literature, really.
08:49 < nmz787_i> chris_99: only speculation, but seems extremely likely
08:49 < kanzure> nmz787_i: i don't like you going around to out-of-band communication to try to convince me that you have this figured out
08:49 < chris_99> nice, i saw something about them supposed to make FPGA x86 chips a while ago iirc
08:49 < nmz787_i> kanzure: they do it all the time, dilute something to 1-molecule per milliliter... suck out 1 milliliter
08:49 < kanzure> nmz787_i: you can't handwave
08:49 < fenn> i would like an actually open laptop design, but i am cheap also
08:49 < kanzure> nmz787_i: he's not handwaving when he says "i'm using the posam valves and solenoid actuators".
08:50 < juri_> chris_99: there were FPGAs that fit into socket940.
08:50 < chris_99> http://www.extremetech.com/extreme/184828-intel-unveils-new-xeon-chip-with-integrated-fpga-touts-20x-performance-boost is what i was thinking of
08:50 < nmz787_i> kanzure: he didn't really reply though... I don't want to go read that paper when he already has the answer
08:50 < chris_99> juri this is an x86 with FPGA though
08:50 < juri_> nmz787_i: i read the paper yesterday. ;)
08:50 < nmz787_i> juri_: and I read it months and years ago
08:50 < juri_> it's a good read. i recommend it.
08:50 < fenn> still waiting for "reconfigurable computing" - the FPGA patent bog has delayed this miracle of technology for 20 years at least
08:51 < kanzure> CaptHindsight: hah, glad to hear you have met lkcl
08:51 < nmz787_i> juri_: meh, it's kind of boring to me outside the chemistry section
08:51 < juri_> lkcl follows my 3d printing work, but doesn't talk to me on irc, only email.
08:52 < juri_> weird.
08:52 < CaptHindsight> kanzure: have also met flyback, hours of fun
08:52 -!- EnabrinTain [sid11525@gateway/web/irccloud.com/x-vyhcaaiofprbcbch] has joined ##hplusroadmap
08:52 < juri_> flyback is troubled.
08:52 < kanzure> CaptHindsight: nmz787_i thinks that you haven't put enough thought into the valves and solenoids.
08:52 < juri_> I've known him for.. 15+ years.
08:52 < CaptHindsight> fenn: yeah, I've been hearing about it since the 80's
08:53 < fenn> but now we have "apps"
08:53 < fenn> woo
08:53  * fenn cries
08:53 < nmz787_i> kanzure: that isn't true... I simply asked if they were pumps or acting as valves
08:53 < kanzure> nmz787_i: gravity-assisted phosphoramidite dispense from the vials into the inkjet printhead
08:54 < nmz787_i> kanzure: so a siphon?
08:54 < CaptHindsight> nmz787: there's a manifold with solenoid vales, it's pretty simple
08:54 < nmz787_i> or this is still argon-pressure driven?
08:54 < CaptHindsight> it's like having multiple faucets going to one spout
08:55 < nmz787_i> that isn't an answer though to the question
08:55 < CaptHindsight> the paper jumps between argon and nitrogen, the point is no oxygen and no water
08:55 < juri_> I can talk to some people at BUGS, in case the 'where te put the machine' problem is still a problem. they have a wet lab there.
08:55 < CaptHindsight> nmz787: what you you really want to know?
08:55 < nmz787_i> you just answered it
08:55 < CaptHindsight> you/do
08:55 < nmz787_i> pressure-driven pumping, with the solenoids acting as valves
08:56 < juri_> a few ideas:
08:56 < juri_> you should be using the same laser for droplet detection as you use for head position checking. just make the mechanics interfere.
08:57 < CaptHindsight> electrically operated on/off vales
08:57 < juri_> the first priority should be droplet detection and analisys.
08:57 -!- PatrickRobotham_ [uid18270@gateway/web/irccloud.com/x-qfnipffmztvesdfv] has joined ##hplusroadmap
08:57 < kanzure> juri_: there's a droplet detector in the spec
08:57 < kanzure> specification
08:57 < CaptHindsight> juri_: we can also add a camera to look at drops
08:58 < juri_>  we should be doing volumetric work as soon as we can verify we can fire the jets. so we can adapt the control signal to the fluids in question.
08:58 < nmz787_i> camera needed anyway for tritly color detection
08:58 < nmz787_i> trityl*
08:58 < CaptHindsight> and yes, the laser would tell you when the printhead is at a Home position
08:58 < fenn> lol yeah right "It is probably a lot easier to rewrite a small segment of critical code to run on an FPGA than to rewrite the entire application in OpenCL."
08:58 < juri_> ok, good.
08:58 -!- PatrickRobotham_ is now known as PatrickRobotham
08:58 < nmz787_i> unless you were underdosing the acid, and taking the efficiency hit for simplcitys sake
08:58 < fenn> openCL is super portable and has plenty of uses outside of GPGPU
08:58 < CaptHindsight> 2 lasers can give you XY
09:01 < CaptHindsight> laser are also often used to prevent the heads from crashing into the print substrates
09:02 < CaptHindsight> you don't want an idiot to crash a $8k printhead by not setting the lower limit of the Z
09:02 < CaptHindsight> or only checking one edge of the substrate when the other end is 5mm thicker
09:03 < juri_> it's a little different from what i do, where the worse damage is realigning, and scratches in the bed.
09:04 < CaptHindsight> we still don't want to smack a $200 printhead
09:04 < fenn> i want an all-out televised printhead brawl, with lasers and rockets
09:05 < CaptHindsight> it takes a few hours to get the printer back up again if you do it when you're ready to print with all the vials primed
09:06 < juri_> so, what do you have for signaling specs on these new heads?
09:07 < CaptHindsight> juri_: https://www.youtube.com/watch?v=hq-r0vAMURw  watching drops in flight
09:08 < nmz787_i> fenn: battlebots -> printrbattle? or battleprintrs?
09:08 < kanzure> i'm still getting random email from the central bank of ecuador; they are still flustered that i asked for an sdk.
09:10 < kanzure> i need to decide about electrode array vs not electrode array. bleh, i don't want to have to manufacture increasingly smaller electrode arrays. that's going to be a pain in the butt.
09:11 < CaptHindsight> anyone know these people? https://www.youtube.com/watch?v=SloUgM6k77g BioDot BioJet Elite™: Drop-on-Dro
09:11 < kanzure> .title
09:11 < yoleaux> BioDot BioJet Elite™: Drop-on-Drop - YouTube
09:11 < nmz787_i> kanzure: fab is probably more accessible than those printheads, worldwide
09:11 < CaptHindsight> they use microvales for dispensing
09:12 < nmz787_i> well, at least in terms of more-than-one-off-pricing
09:13 < CaptHindsight> the heads may be removed from a new Epson 1430 $300 or for ~$200 from the Asian suppliers that assemble the printers
09:14 < CaptHindsight> there are a few importers that also sell just heads locally in the US and in the EU
09:15 < fenn> electrode array would be more appealing if it weren't exactly as slow as every other reaction chemistry out there
09:15 < fenn> can you even buy those chips?
09:15 < nmz787_i> no but the scales are huge in terms of fab
09:15 < CaptHindsight> there is a demand for just heads since people started hacking the Epson printers to make T-shirt printers
09:16 < kanzure> fenn: seem to be purchasable here https://azcobiotech.com/oligo-array-12k-chip-cleavable-19-5012.html
09:16 < nmz787_i> they'd be able to be produced on the oldest/cheapest process available at any place, most likely (outside academia, where they still like to train kids on 100+ micron features, just because it works and is easier)
09:17 < nmz787_i> CaptHindsight: do you think the plastic housing materials have changed since the POSAM paper? that would require re-tweaking of chemistry
09:17 < fenn> uh so is that reusable?
09:17 < kanzure> unclear if reusable
09:17 < kanzure> if you can cleave everything off of it....
09:17 < CaptHindsight> DLP lithography can get us few micron features
09:17 < nmz787_i> not to mention every time the manufacturer decides to change materials, since that likely affects ink less than DNA synth chems
09:17 < fenn> .c 375/(12000*100)
09:18 < yoleaux> 375/(12000×100) = 1/3200
09:18 < fenn> bah
09:19 < fenn> well it's not a huge marginal cost even if it's not reusable, assuming you use the whole chip every time
09:19 < nmz787_i> kanzure: you realize you will have to do some damn costly sorting to get verbatim sequences out of an array right? you realize that was cambrian genomics whole ballgame, right?
09:20 < kanzure> sorting?
09:20 < fenn> still it would be annoying to have to wait until you have 12000*n bases to synthesize
09:20 < fenn> the point of having your own synthesizer is to shorten the development cycle right?
09:20 < nmz787_i> kanzure: well your 12k spots off 100bp would not be /only/ 100bp... it's going to be a spectrum of lengths
09:21 < nmz787_i> kanzure: you're still going to need to sort/filter the bad sequences, if you aren't doing some evolutionary work where some error is OK
09:21 < kanzure> fenn: interesting point
09:21 < kanzure> nmz787_i: that's the same case with cambrian's method anyway. it's just strands on beads. the beads can have many molecules that each have different errors.
09:22 < nmz787_i> well except that they were sequencing each bead, and only using the ones with low/no error
09:22 < nmz787_i> also, their beads may have been quite tiny, or primed in very dilute conditions so that few oligos grew per bead
09:23 < nmz787_i> kanzure: my point is... they weren't/aren't/haven't been so successful with their train of half-mil-$ machines
09:24 < kanzure> what?
09:24 < fenn> i really doubt the machine itself cost very much
09:24 < fenn> vs paying salaries for people for years, renting a space etc
09:25 < CaptHindsight> from what I saw of the printers they aren't that complex, not sure of that patent situation to make a DIY version
09:25 < nmz787_i> they had an azco machine, then a mi or hiseq, then their special Helicos-tech-like sequencer with laser desorption, probably an HPLC
09:28 < kanzure> fenn: if it's reusable then i think those concerns are nulled
09:29 -!- rigel [~yourmom@c-24-21-52-83.hsd1.or.comcast.net] has joined ##hplusroadmap
09:30 < CaptHindsight> oligo array 12K chip 50 pk  	Our price $16,250.00
09:30 < fenn> i'm unclear on how much reagents the array chip uses and how that compares to inkjet
09:30 < kanzure> i wonder why they would have a 50 pack
09:30 < kanzure> maybe they are reusable but people ship the arrays to other people for long-term use or something
09:31 < kanzure> (in which case, they should be rented to people, rather than just given)
09:31 < juri_> lead time to manufacture?
09:31 < fenn> you can use the arrays as hybridization detectors
09:31 < fenn> instead of cleaving them off
09:31 < kanzure> yes, i believe most people use these arrays for hybridization stuff
09:31 < fenn> say you want to detect if genes 1 to 12000 are present in a cell extract
09:35 < kanzure> ah they have a non-cleavable version? http://azcobiotech.com/oligo-array-12k-chip-non-cleavable-19-5013.html
09:35 < nottimschmidt> back, catching up
09:37 < kanzure> here are some weird ways the customarray microarray has been used by others http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3231185/
09:38 < ParahSailin> nmz787_i: and the good part is, he doesnt even need to go to a cmos shop because it can be bought off the shelf as a disposable
09:38 < ParahSailin> heh, rice people bugging me again to pay off my parking tickets so that i can pick up my diploma after how many years?
09:39 < kanzure> here is church mentioning customarray http://arep.med.harvard.edu/pdf/Kosuri_Church_2014.pdf
09:39 < kanzure> "$0.00001/nucleotide"
09:41 < kanzure> "Array-based gene synthesis. Even though microarray-based oligo pools are cheap, there are several challenges in using them for gene synthesis. First, although the numbers of oligos that can be produced in a pool are large, their individual concentrations are quite low for most existing gene synthesis protocols. Second, the error rates for oligo pools are usually higher than those for column-synthesized oligos. Finally, the sheer number ...
09:42 < kanzure> ... of oligos produced leads to interference between gene assemblies, making it difficult to scale up."
09:42 < kanzure> "Tian et al.35 were the first to show how these problems could be overcome. They used PCR amplification to increase the concentration of the oligos before assembly, error-corrected them by hybridization to reverse complementary oligos (also constructed on the chip), and designed protein sequences computationally to avoid potential mishybridizations of the sequences."
09:42 < kanzure> "However, this work and contemporaneous reports36,37 used only dozens to hundreds of oligos at once. Scaling these methods proved difficult beyond pool sizes of 1,000 oligos38. At greater pool complexities, where the advantages in cost would come to play, constructing any individual gene became difficult, presumably owing to spurious cross-hybridization during the assembly process. In addition, the methods required sufficient sequence ...
09:42 < kanzure> ... orthogonality between synthesized genes, which limited potential applications."
09:42 < kanzure> "To alleviate these and other issues, two approaches were used that first isolated subpools of oligos required for any single assembly, thereby overcoming the concerns about both pool complexity and sequence orthogonality (Fig. 3). Kosuri et al.39 used predesigned barcodes that allowed PCR amplification of oligos participating in only a particular assembly and then removed the barcodes by digestions, which was followed by standard ...
09:42 < kanzure> ... assembly of the genes. Quan et al.40 used a custom inkjet synthesizer20 that synthesized subsets of oligos in physically separated microwells, where amplification and assembly were then done in situ. Both methods used much larger oligo pools (>10,000 oligos) and enzymatic error correction, which paved the way for commercialization in recent years (Gen9)."
09:43 < kanzure> well there's no reason to use a single pool
09:43 < fenn> the barcode thing is cool (and relevant to gene synthesis)
09:43 < kanzure> i don't know why people are focusing on single pool
09:44 < fenn> what's "enzymatic error correction"?
09:44 < ParahSailin> theres that dnase that attacks non-duplex dna
09:45 < kanzure> "More general methods of error correction usually depend upon a number of enzymatic techniques to reduce errors. All of these techniques rely on the fact that at any given position, most molecules possess the correct base. Heating and reannealing can force the formation of heteroduplexes that will contain disruptions to the canonical helical DNA structure. Such disruptions can be recognized and acted upon by several proteins."
09:45 < kanzure> "MutS binds heteroduplexes and can be used to filter errors by reverse purification13. Certain polymerases with exonuclease activity, endonucleases and resolvases can all cut or nick at such heteroduplex sites and, upon reamplification, can help filter errors12,26,45–47. Commercial enzymatic cocktails such as ErrASE have been commonly used to help reduce errors in synthetic genes as well30,39. Such error reduction can greatly reduce ...
09:45 < kanzure> ... the cost and time of gene synthesis by bringing error rates low enough that the genes can be directly used for functional assays without in vivo cloning and sequence verification30,48."
09:45 < fenn> ParahSailin: if you had a deletion mutation, wouldn't the dnase just destroy the opposing strand then?
09:46 < fenn> (not the mutated strand)
09:46 < kanzure> instead of single pool why isn't everyone using micropipette of each well?
09:46 < fenn> kanzure because it's a pain to synthesize things in wells
09:47 < kanzure> "More recent approaches have leveraged NGS technologies to screen and then select for perfect sequences at either the oligo or gene level. Matzas et al.49 used 454 sequencing (Roche) combined with a robotic pick-in-place pipette to mechanically pull reads with perfect sequence off the sequencing array and used these oligos for synthetic genes. Kim et al. also used 454 but marked individual molecules with random tags and then amplified ...
09:47 < kanzure> ... constructs with perfect sequence42. Both approaches help reduce error rates, although they still suffer from sequencing errors and require more expensive long-read platforms. Schwartz et al.41 used Illumina sequencing of barcodes to pull out perfect sequences but overcame limitations in length and sequencing errors through a tag-based consensus-sequencing approach. All three approaches substantially cut error rates and will continue ...
09:47 < kanzure> ... to improve with the rapid progress in DNA sequencing technologies. These NGS based error-correction approaches are also especially exciting for library-based constructions and synthesis methods because they allow correction without having to separate each gene assembly into individual reactions."
09:47 < kanzure> so.. that's wells/pores right there.
09:47 < kanzure> as far as i know
09:49 < kanzure> http://image.slidesharecdn.com/ngsintrov6public-110320062557-phpapp01/95/ngs-introv6public-15-728.jpg?cb=1300602396
09:49 < kanzure> (image is a cross-section of picotiter plates)
09:50 < kanzure> that's what we should be doing
09:50 < kanzure> this method is also compatible with inkjet stuff
09:50 < ParahSailin> thats ancient
09:50 < fenn> the pick and place thing is wells, the other two methods are not
09:50 < ParahSailin> do it like an illumina flow cell, its already attached to a glass slide
09:51 < fenn> i really dont get the microtiter thing
09:51 < ParahSailin> all you need then is lasers and microscopes
09:51 < kanzure> fenn: it's just a high-density pore array
09:51 < kanzure> and then they use pick-and-place
09:51 < fenn> they mechanically position a pipette tip to within 40 microns?
09:51 < kanzure> well, not for sequencing, but yes
09:52 < CaptHindsight> fenn: what do you see as the problems with wells?
09:52 < fenn> and then do that a zillion times?
09:52 < kanzure> fenn: that's my understanding of the last quote i pasted
09:52 < fenn> doesn't that seem ... crazy?
09:52 < kanzure> it's more specific than single pool
09:53 < juri_> CaptHindsight: is that an answer to my question?
09:53 < ParahSailin> what you really want is oligo clones
09:53 < CaptHindsight> heh yeah
09:53 < kanzure> matzas reference is "High-fidelity gene synthesis by retrieval of sequence-verified DNA identified using high-throughput pyrosequencing"
09:53 < ParahSailin> so load your pooled library into flowcell and do the bridge amplification step
09:53 < kanzure> http://arep.med.harvard.edu/pdf/Matzas_10.pdf
09:53 < ParahSailin> then sequence all the clusters on flow cell and pick and place
09:54 < fenn> CaptHindsight: have to add reagents to each well, and the wells get in the way of your deposition method (?)
09:55 < ParahSailin> destroy all bad clusters with uv laser to make it easier to pick good ones off
09:55 < fenn> there's more area available for synthesis if there are no well walls
09:55 < fenn> i'm not really claiming these are obstacles
09:56 < ParahSailin> no, that probably wouldnt work
09:56 < kanzure> i agree that pore/well walls are not entirely necessarily
09:56 < nmz787_i> kanzure: micropipette of each well would still be a single pool though
09:56 < kanzure> especially if you are doing single drop isolation
09:56 < kanzure> nmz787_i: what??
09:56 < CaptHindsight> fenn: ok the walls take up space, so wall-less wells use surface tension
09:56 < ParahSailin> illumina has gone to patterned flow cells with wells
09:56 < nmz787_i> you can have a kiddie pool or an olympic sized pool, both are pools
09:56 < kanzure> nmz787_i: yes, each drop is a single pool, so what?
09:57 < fenn> CaptHindsight: i hadnt considered that
09:57 < nmz787_i> I'm guessing their  assumptions hold for all sized pools, minus the pool where there are few enough molecules to count them on your fingers and maybe toes also
09:57 < kanzure> ParahSailin: can you please find me a useful diagram of an illumina flow cell :-( i don't understand
09:58 < ParahSailin> http://www.illumina.com/company/video-hub/pfZp5Vgsbw0.html
09:58 < kanzure> .title
09:58 < yoleaux> Patterned Flow Cell Technology | Illumina Video
09:58 < juri_> CaptHindsight: so, is that answer "i've got it"?
09:58 < fenn> .title http://youtu.be/pfZp5Vgsbw0
09:58 < yoleaux> Patterned Flow Cell Technology | Illumina Video - YouTube
10:00 < ParahSailin> you can get scrap used flow cells in the trash
10:00 < nmz787_i> kanzure: you were complaining about people talking about single pools... I am guessing they are concerned with single pools because they need this info to be able to say whether or not the process can be scaled to multiple parallel pools, or sequential, etc
10:00 < fenn> ooo aah
10:01 < kanzure> nmz787_i: the point is that you don't want all dna from all pores in the same pot or pool. you can't assemble genes easily in that scenario. but if you separately handle each pore, you can combine strands together at your own pace.
10:01 < fenn> they ensure only one sequence per well simply by reaction kinetics eh?
10:01 < kanzure> nmz787_i: no they are doing typical dna hybridization stuff.... they aren't doing gene assembly.
10:01 < kanzure> fenn: yes that sort of surprises me
10:02 < fenn> no it makes sense
10:02 < fenn> alright i am ready to unlock the power of the genome
10:03 < ParahSailin> i dont know what fraction of the wells randomly start with one oligo
10:03 < archels> "Hi all, I’m writing to inform you all that in the cell culture lab I dark reared some mice in the drawer where the syringes are stored. Please do not open or attempt to open the above mentioned drawer."
10:03 < ParahSailin> thats more of a matter of tinkering with the library dna concentration
10:03 < fenn> "a high percentage"
10:04 < fenn> there was some kind of initiator molecule, so if you have a low concentration of that, and a low concentration of dna, it's extremely unlikely that you'd have two dna's and two initiators at the same place and time
10:04 < kanzure> "with enough scale the error rate doesn't matter!~~~"
10:04 < fenn> once there's one initiator and one dna, amplification happens quickly and the substrate gets used up
10:04 < nmz787_i> fenn: likely some non-polymer, so they can ensure a monolayer?
10:05 < fenn> nmz787_i: what?
10:05 < nmz787_i> the initiator
10:05 < fenn> oh, it looked like an enzyme
10:05 < fenn> it was a red blob
10:05 < ParahSailin> guys illumina publishes protocols
10:06 < fenn> all i know is what we just saw  http://youtu.be/pfZp5Vgsbw0#t=54s
10:06 < ParahSailin> http://support.illumina.com/content/dam/illumina-support/documents/myillumina/ab41723c-129f-47e1-8000-7282fb605c87/clusterstationuserguide_15018818_d.pdf
10:07 < nmz787_i> .youtube gcta pcr commercia
10:07 < kanzure> pick-and-place bead protocol could be done on the same microarray plate. you could probably run lots of simultaneous gibson assemblies or something.
10:07 -!- nmz787_i [ntmccork@nat/intel/x-nxxygqsseduszlkg] has quit [Quit: Leaving.]
10:08 < CaptHindsight> exclusion amplification
10:08 < CaptHindsight> what is this magic?
10:10 < CaptHindsight> http://www.google.com/patents/WO2013188582A1?cl=en
10:10 < CaptHindsight> Kinetic exclusion amplification of nucleic acid libraries
10:11 < fenn> i wonder why this took until 2013 to figure out
10:11 < fenn> pretty sure it is older than that
10:11 < ParahSailin> illumina has been commercial since 2008
10:12 < CaptHindsight> maybe they just decided to patent after the lawyers combed through everything
10:13 < fenn> illumina is a company; are you saying they only use this specific sequencing method based on exclusion amplification?
10:13 < ParahSailin> oh, its not clear what this kinetic exclusion bla is
10:13 < ParahSailin> clearly not essential to flowcell sequencing
10:14 < kanzure> all the cells are under the same flow?
10:14 < kanzure> so they are experiencing aqueous contamination with each other?
10:14 < ParahSailin> there are only 8 different channels in a flow cell
10:17 < kanzure> so... yes?
10:19 < ParahSailin> each cluster on the flow cell is in the same aqueous solution
10:20 < kanzure> fenn: here is the "crazy" micropippete + microarray pick-and-place approach, see figure 2 on page 2 http://www.nature.com/nbt/journal/v28/n12/extref/nbt.1710-S1.pdf
10:21 < kanzure> oh is that the same matzas paper. i see.
10:21 < CaptHindsight> would it help to have a femtoliter drop size vs pL?
10:21 < fenn> that's a really thick block of granite
10:21 < fenn> don't sneeze
10:22 < CaptHindsight> heh
10:23 < fenn> so this actually works in practice? not just like one or two sequences
10:23 < kanzure> i seem to recall that they had a person operating the pipette tip, and they were manually looking at the video to determine when they were crashing the tip into the edge of a well or pore
10:24 < kanzure> but i don't remember why they were doing this
10:24 < fenn> i gotta say i like the pcr barcode way just in principle
10:24 < kanzure> because at 4096x4096 that's.. a lot of video game playing.
10:25 < fenn> from DNA select * where sequence=^AGCT.*
10:26 < CaptHindsight> fenn: here's a 5-axis we did for a large silicon valley fruit company  http://imagebin.ca/v/28Vp4S0dYt00
10:26 < kanzure> is this just dna hybridization wings technique
10:26 < kanzure> that should be labeled nsfw, it's clearly machine porn
10:26 < CaptHindsight> 5um repeatability with inkjet, lasers and microscope
10:27 < fenn> i can almost smell the ethylene
10:28 < fenn> ok i think my brain is melting bbl
10:29 < CaptHindsight> we can pick and place using a 0.5um tip pretty quick using linear servos
10:30 < kanzure> i wonder why they needed manual operator control of that
10:30 < CaptHindsight> budget, skill ?
10:31 < CaptHindsight> time
10:33 < CaptHindsight> heh pages 23- 135 are all just the sequences in text format
10:33 < kanzure> yes, biologists are silly
10:33 < CaptHindsight> I winder if that would fit on a floppy?
10:35 < kanzure> "Kosuri et al.39 used predesigned barcodes that allowed PCR amplification of oligos participating in only a particular assembly and then removed the barcodes by digestions, which was followed by standard assembly of the genes."
10:35 < kanzure> "Quan et al.40 used a custom inkjet synthesizer20 that synthesized subsets of oligos in physically separated microwells, where amplification and assembly were then done in situ."
10:35 < CaptHindsight> maybe inkjet directly to beads
10:35 < kanzure> yes
10:35 < kanzure> totally doable
10:35 < kanzure> ( those quotes are from http://arep.med.harvard.edu/pdf/Kosuri_Church_2014.pdf )
10:36 < CaptHindsight> beads + laser launcher
10:36 < CaptHindsight> lets get the chems worked out and the timing
10:36 < CaptHindsight> then we can print on beads
10:37 < CaptHindsight> how did Cambrian make their beads?
10:38 -!- nottimschmidt [~timschmid@35.10.220.177] has quit [Ping timeout: 265 seconds]
10:40 < juri_> bbiaf.
10:42 < kanzure> CaptHindsight: there's lots of places to buy beads
10:45 -!- nmz787_i [~ntmccork@192.55.54.42] has joined ##hplusroadmap
10:49 -!- nmz787_i1 [~ntmccork@134.134.139.72] has joined ##hplusroadmap
10:50 -!- nmz787_i [~ntmccork@192.55.54.42] has quit [Ping timeout: 255 seconds]
10:51 < CaptHindsight> hmm, printing right onto beads might only help if you have very low error rates
10:53 -!- nottimschmidt [~timschmid@13-195-195.client.wireless.msu.edu] has joined ##hplusroadmap
10:54 < kanzure> the beads is more for easy transport
10:54 < kanzure> afaik it does not materially impact things like reaction kinetics
10:54 < CaptHindsight> a laser can launch them off anything
10:56 < kanzure> btw you might be unaware but there are some pcr techniques that involve laser heating of picoliter droplets on a surface
10:57 < CaptHindsight> makes sense
10:57 < nmz787_i1> well beads can have a multitude of surface material modifications
10:58 < nmz787_i1> CaptHindsight: you can take a look at bangslabs.com
10:59 < kanzure> "Parallel on-chip gene synthesis and application to
10:59 < kanzure> gah
11:00 < kanzure> "Parallel on-chip gene synthesis and application to optimization of protein expression" http://www.researchgate.net/profile/Nicolas_Negre/publication/51073109_Parallel_on-chip_gene_synthesis_and_application_to_optimization_of_protein_expression/links/0046352567745e4375000000.pdf
11:00 < kanzure> they are claiming oligo amplification and assembly on the same inkjet-printed array?
11:00 -!- delinquentme [~delinquen@74.61.157.78] has joined ##hplusroadmap
11:02 < kanzure> ah, polymerase cycling assembly
11:03 < kanzure> er, "polymerase chain assembly"
11:10 < kanzure> .title http://www.nature.com/ncomms/2015/150202/ncomms7073/full/ncomms7073.html
11:10 < yoleaux> A high-throughput optomechanical retrieval method for sequence-verified clonal DNA from the NGS platform : Nature Communications : Nature Publishing Group
11:10 < kanzure> "pulse laser retrieval system"
11:10 < kanzure> "Writing DNA plays a significant role in the fields of synthetic biology, functional genomics and bioengineering. DNA clones on next-generation sequencing (NGS) platforms have the potential to be a rich and cost-effective source of sequence-verified DNAs as a precursor for DNA writing. However, it is still very challenging to retrieve target clonal DNA from high-density NGS platforms. Here we propose an enabling technology called ...
11:10 < kanzure> ... ‘Sniper Cloning’ that enables the precise mapping of target clone features on NGS platforms and non-contact rapid retrieval of targets for the full utilization of DNA clones. By merging the three cutting-edge technologies of NGS, DNA microarray and our pulse laser retrieval system, Sniper Cloning is a week-long process that produces 5,188 error-free synthetic DNAs in a single run of NGS with a single microarray DNA pool. We ...
11:10 < kanzure> ... believe that this technology has potential as a universal tool for DNA writing in biological sciences."
11:11 < kanzure> a more general article, http://www.biotechniques.com/news/biotechniquesNews/biotechniques-356987.html#.VaP_rlIX3RY
11:11 < CaptHindsight> ooh
11:11 < kanzure> oh this seems to be the same person (bang)
11:11 < kanzure> nope nevermind
11:12 < kanzure> "Fortunately, they came across a 2010 Nature Biotechnology paper (1) from George Church’s group describing megacloning, which uses NGS to directly select sequence-verified clones from a complex pool of mixed oligonucleotides after NGS without the need for subsequent selective PCR amplification. This was possible using Roche 454 Life Sciences’ GLS platform, where each individual bead in a well in the open-top 454-Picotiterplate (PTP) ...
11:12 < kanzure> ... that corresponds to a sequenced oligonucleotide clone could be directly physically extracted with a micropipette using a robotic imaging and bead picking system."
11:12 < kanzure> "“The concept of direct utilization of clones on the NGS substrate was a refreshing jolt,” said Kwon, but as an engineer, he felt that there was room for improvement to the method. The throughput was constrained by the need for physical contact-based retrieval of the beads, making the method inadequate for generating the large number of oligonucleotides necessary for writing megabase-sized DNA pieces."
11:12 -!- ebowden [~ebowden@101.177.244.210] has joined ##hplusroadmap
11:13 < kanzure> "Kwon and Bang’s improved version of the megacloning method, Sniper Cloning, was recently published in Nature Communications (2). Their key modification is the use of a fully automated pulse laser system for the non-contact and very rapid retrieval of the desired beads from the PTP. They cleverly took advantage of the structure of the PTP, where each well that holds one bead is formed by etching of the end of an individual fiber optic ...
11:13 < kanzure> ... core that conducts the light released from the bead during pyrosequencing back to the CCD camera. By inverting the PTP, a low-energy laser pulse targeted to the back side (i.e. top) of a selected well is transmitted by the fiber optic core to the bead, pushing it out by radiation pressure down into the well of a 96-well microtiter plate positioned underneath the PTP. This approach greatly reduces the possibility of ...
11:13 < kanzure> ... cross-contamination inherent to megacloning and, by using an automated motorized stage, is orders of magnitude faster than retrieving the selected beads from the PTP."
11:14 < CaptHindsight> laser prod
11:14 < CaptHindsight> non-contact
11:14 < CaptHindsight> except for the well material
11:14 < kanzure> hmm so what doesn't make sense to me here is that the 96 array doesn't allow you much room... sure, you can punch any bead you want into any one of the 96 slots, then you can run further chemistry to combine the oligos. but then what if you want to combine two of those sequences together (to make a larger dna fragment)?
11:16 < CaptHindsight> it's 96 beads since the beads are large
11:16 < kanzure> they launch the smaller beads into the larger 96 plate
11:17 < CaptHindsight> inkjet pools to laser lanes
11:17 < kanzure> right, still need the inkjet to print on to the tiny beads anyway
11:18 < CaptHindsight> can't you just push the next oligo onto the last one combined to make it longer?
11:18 < CaptHindsight> what makes it stop at only 2? anything?
11:19 < kanzure> 96 plate is incompatible with laser transportation of beads into other 96 plate holes
11:20 < kanzure> there are 2 plates. super-tiny-pore plate is on top. 96 plate is below. stage moves 96 plate. laser pokes a micro bead out of tiny plate, drops into known pore on 96 plate.
11:20 < kanzure> you oculd use pipettes to transport material from one 96 plate pore to another pore on the 96 plate, i guess
11:24 < delinquentme>  If the constant region is replaced with the human form, the antibody is termed chimeric and the substem used is -xi-.
11:24 < yoleaux> 10 Jul 2015 12:20Z <kanzure> delinquentme: https://www.youtube.com/watch?v=JiyxPae4h5A&t=17m30s
11:24 < delinquentme> do we know anyone who knows this to be decidedly true?  ( its from wikipedia )
11:24 -!- gnusha [~gnusha@unaffiliated/kanzure/bot/gnusha] has quit [Ping timeout: 255 seconds]
11:24 -!- gnusha [~gnusha@unaffiliated/kanzure/bot/gnusha] has joined ##hplusroadmap
11:24 -!- Topic for ##hplusroadmap: biohacking, nootropics, transhumanism, open hardware | sponsored by lobsters everywhere, banned by the Federal Death Administration (5 times) | this channel is LOGGED: http://gnusha.org/logs | http://diyhpl.us/wiki | "ray kurzweil is a pessimist" - george church
11:24 -!- Topic set by kanzure [~kanzure@unaffiliated/kanzure] [Wed May 20 12:46:25 2015]
11:24 [Users ##hplusroadmap]
11:24 [ abetusk      ] [ catern      ] [ EnLilaSko      ] [ marchtemp      ] [ poohbear    ] [ superkuh ] 
11:24 [ Acty         ] [ chris_99    ] [ eudoxia        ] [ mf1008         ] [ Qfwfq       ] [ Taek     ] 
11:24 [ Adlai        ] [ crescendo   ] [ fenn           ] [ midnightmagic  ] [ redlegion   ] [ the8thbit] 
11:24 [ altersid     ] [ Daeken      ] [ gnusha         ] [ nickjohnson    ] [ rigel       ] [ ThomasEgi] 
11:24 [ andytoshi    ] [ delinquentme] [ Guest34971     ] [ night          ] [ ryankarason ] [ thundara ] 
11:24 [ archels      ] [ diginet     ] [ heath          ] [ nmz787         ] [ saurik      ] [ TMA      ] 
11:24 [ augur        ] [ dingo_      ] [ helleshin      ] [ nmz787_i1      ] [ sh          ] [ Urchin   ] 
11:24 [ Bakkot       ] [ Douhet      ] [ HEx1           ] [ nottimschmidt  ] [ sheena      ] [ vivi     ] 
11:24 [ balrog       ] [ dpk         ] [ JayDugger      ] [ nsh            ] [ sivoais     ] [ wrldpc1  ] 
11:24 [ berndj       ] [ drethelin   ] [ jrayhawk       ] [ p4nd4          ] [ smeaaagle   ] [ xrr      ] 
11:24 [ Betawolf     ] [ drewbot_    ] [ juri_          ] [ padz           ] [ strages     ] [ yoleaux  ] 
11:24 [ bkero        ] [ dustinm     ] [ justanotheruser] [ panax          ] [ strangewarp_] [ yorick   ] 
11:24 [ BobaMa       ] [ ebowden     ] [ juul           ] [ ParahSailin    ] [ streety     ] 
11:24 [ Burn_        ] [ eleitl      ] [ kanzure        ] [ pasky          ] [ Stskeeps    ] 
11:24 [ CaptHindsight] [ EnabrinTain ] [ kish           ] [ PatrickRobotham] [ super`      ] 
11:24 -!- Irssi: ##hplusroadmap: Total of 87 nicks [0 ops, 0 halfops, 0 voices, 87 normal]
11:24 -!- Channel ##hplusroadmap created Thu Feb 25 23:40:30 2010
11:24 -!- Irssi: Join to ##hplusroadmap was synced in 12 secs
11:24 -!- heath [~heath@unaffiliated/ybit] has quit [Ping timeout: 255 seconds]
11:26 -!- heath [~heath@unaffiliated/ybit] has joined ##hplusroadmap
11:26 < nmz787_i1> kanzure: it would be foreseeable to swap the electrochemical acid generation electrodes with some sensor-frontend and elucidate some data to increase synthesis fidelity
11:27 < kanzure> swap?
11:28 < nmz787_i1> from push to sense
11:28 < nmz787_i1> pin mux
11:28 < kanzure> ok pin mux
11:30 -!- nottimschmidt [~timschmid@13-195-195.client.wireless.msu.edu] has quit [Ping timeout: 264 seconds]
11:36 < kanzure> stalk: bernellev@gmail.com
11:38 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-egeeezwmfnjhahee] has joined ##hplusroadmap
11:47 -!- nottimschmidt [~timschmid@13-74-117.client.wireless.msu.edu] has joined ##hplusroadmap
11:47 -!- ant4t [~s3an@50.141.78.33] has joined ##hplusroadmap
11:55 -!- balrog [~balrog@unaffiliated/balrog] has quit [Ping timeout: 240 seconds]
12:00 < nmz787_i1> jrayhawk: are you or other paleo-types visually artistic at all? I am thinking it would be fun to have a shirt that reads 'I pass on grass' and has pictures of grain at the top of a stalk, popcorn, bread, corn syrup
12:01 < nmz787_i1> also rice
12:02 < nmz787_i1> maybe a picture of someone 'passing' it off to a cow
12:02 < nmz787_i1> 'pass the grass to the left-hand cow'
12:03 < CaptHindsight> from the Sniper paper "To the best of our knowledge, it is very difficult to supply a massive amount of synthetic oligonucleotides of extremely high standard, except through conventional cloning and random pick-and-place followed by individual identification"
12:04 < kanzure> random?
12:04 < nmz787_i1> i guess there are too many in their opinion to search each and every one
12:05 < nmz787_i1> and there's no better heuristic than random?
12:05 < CaptHindsight> heh, yeah random, just let the machine run wild
12:05 -!- nottimschmidt [~timschmid@13-74-117.client.wireless.msu.edu] has quit [Ping timeout: 240 seconds]
12:05 < kanzure> ok sounds like they mean "other than doing the full entire array, just use some random subset, because we don't want to wait for pick-and-place"
12:05 < kanzure> machines can pick-and-place very very quickly
12:06 -!- balrog [~balrog@unaffiliated/balrog] has joined ##hplusroadmap
12:07 < CaptHindsight> https://www.youtube.com/watch?v=u6KW8fIBjr8  Adept delta bot
12:08 -!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-qfnipffmztvesdfv] has quit [Quit: Connection closed for inactivity]
12:08 < CaptHindsight> https://www.youtube.com/watch?v=0-Kpv-ZOcKY  this is the one from a few years ago handling ball bearings
12:21 < kanzure> "It is stinky shit, but it's kinda working." - eleitl
12:22 -!- nottimschmidt [~timschmid@13-65-55.client.wireless.msu.edu] has joined ##hplusroadmap
12:27 -!- davm23 [~davm@adsl-107-203-26-177.dsl.tul2ok.sbcglobal.net] has joined ##hplusroadmap
12:32 < CaptHindsight> with laser interferometry, linear servos and temp control we can have few um repeatability at high speeds with a femtoliter inkjet
12:33 < CaptHindsight> lets just say 5um between centers for sample spots
12:33 < nmz787_i1> at that point we'd get further by just making a cheap photolithography mastering system
12:33 < nmz787_i1> using a bluray drive optical sled
12:33 < nmz787_i1> (they have near-diffraction limit optics)
12:34 < nmz787_i1> http://diyhpl.us/~bryan/papers2/optics/photolithography/High%20resolution,%20low%20cost%20laser%20lithography%20using%20a%20Blu-ray%20optical%20head%20assembly.pdf
12:34 < CaptHindsight> doesn't lithography mean flooding each layer at a time?
12:34 < CaptHindsight> and having to move masks around?
12:35 < nmz787_i1> not necessarily... it can mean making microfluidics in a closed-system
12:35 < CaptHindsight> oh, yeah different animal
12:35 < CaptHindsight> easy once i make the machines to make those machines
12:36 < nmz787_i1> lot more microfluidics research than just DNA (i.e. cell culturing)
12:37 < CaptHindsight> the blue ray sleds are nice for the laser diode but they rely on tracks for accuracy
12:37 < nmz787_i1> nah that's only for positioning
12:37 < nmz787_i1> the focus is sub-micron
12:37 < nmz787_i1> that's why i commented on that only after you mentioned interferometry
12:37 < CaptHindsight> well depends on how we define the sled
12:37 < nmz787_i1> the part with the optics, that slides
12:37 < CaptHindsight> sure, we use them all the time
12:38 < CaptHindsight> the 405nm laser diodes
12:39 < nmz787_i1> well the diodes themselves need quite a bit of work to get to sub-micron focus
12:39 < CaptHindsight> why hasn't anyone made a low cost diy blueray printer for sub-micron features?
12:39 < nmz787_i1> TEM00 and such
12:39 < nmz787_i1> CaptHindsight: because time on my part
12:39 < CaptHindsight> do I have to do everything?
12:40 < nmz787_i1> my idea without interferometry was to use a grid reticle under a video microscope, and machine vision... for feedback
12:40 < nmz787_i1> and something like a microscope stage for XY
12:42 < CaptHindsight> how many would use it or build it?
12:43 < CaptHindsight> my concern is that you give the plans out and people can't make it work since they have too little experience with fabrication anything
12:44 < kanzure> CaptHindsight: yes a future machine worth making is a way to easily and quickly make microfluidic devices. either laser diode laser cutter or maskless photolithography or something.
12:44 < CaptHindsight> then you have to spend lots of time holding their hands to make it work
12:44 < kanzure> CaptHindsight: we had a partial laser diode cutter thingy, but we couldn't do the optics
12:44 < kanzure> http://diyhpl.us/laser_etcher/laser_etcher/
12:45 < CaptHindsight> maybe give them the working mechanism and then they can wrap it and write software
12:46 -!- proofoflogic [uid65184@gateway/web/irccloud.com/x-wqcoozsjjgnawftq] has joined ##hplusroadmap
12:50 -!- MrHindsight [~not_sure@unaffiliated/capthindsight] has joined ##hplusroadmap
12:56 -!- genehacker [~chatzilla@128-193-154-1.ptpg.oregonstate.edu] has joined ##hplusroadmap
13:08 < kanzure> "Automatic error elimination by horizontal code transfer across multiple applications" http://people.csail.mit.edu/fanl/papers/codephage-pldi2015.pdf
13:16 -!- helleshin [~talinck@66-161-138-110.ubr1.dyn.lebanon-oh.fuse.net] has quit [Read error: Connection reset by peer]
13:19 -!- nmz787_i1 [~ntmccork@134.134.139.72] has quit [Quit: Leaving.]
13:21 -!- saurik [saurik@carrier.saurik.com] has quit [Ping timeout: 255 seconds]
13:21 -!- helleshin [~talinck@66-161-138-110.ubr1.dyn.lebanon-oh.fuse.net] has joined ##hplusroadmap
13:21 -!- saurik [saurik@carrier.saurik.com] has joined ##hplusroadmap
13:26 -!- nottimschmidt [~timschmid@13-65-55.client.wireless.msu.edu] has quit [Ping timeout: 256 seconds]
13:28 -!- Beatzebub [~beatzebub@d162-156-106-225.bchsia.telus.net] has joined ##hplusroadmap
13:30 -!- nmz787_i [ntmccork@nat/intel/x-fmxyjytcpyrvqtsx] has joined ##hplusroadmap
13:31 -!- Acty [uid89656@gateway/web/irccloud.com/x-lnomihmgkbtswyha] has quit [Quit: Connection closed for inactivity]
13:31 < nmz787_i> kanzure: we didn't stop the laser etcher because of the optics, did we? I thought fenn dropped offline and we didn't have anyone to help with CNC advice and modelling (and now fenn is less interested)
13:31 < nmz787_i> CaptHindsight: yes I think retrofitting a microscope is easy enough for the keen biologists
13:32 < nmz787_i> orders of magnitude moreso than a POSAM or derivative
13:32 < kanzure> no clue, i suppose that could be true? hm...
13:33 < nmz787_i> CaptHindsight: the microscope idea has been partially DIYed... http://openlabtools.eng.cam.ac.uk/Instruments/Microscope/   if sourcing microscopes to retrofit doesn't sound as appealing and repeatable as making your own
13:34 < nmz787_i> huh http://www.meetup.com/Cambridge-Synthetic-Biology-Meetup/events/221680133/
13:34 < nmz787_i> .title
13:34 < yoleaux> Hack the Lab: how to build a microscope - Cambridge Synthetic Biology Meetup (Cambridge, England) - Meetup
13:35 < nmz787_i> .title http://www.meetup.com/Makespace/events/221524812/
13:35 < yoleaux> Hack the Lab: how to build a microscope - Makespace (Cambridge, UK) (Cambridge, England) - Meetup
13:38 < kanzure> fenn: it just occurred to me that the pcr barcode method is not the same thing as dna hybridization overlap wings, the barcode is probably meant to be optimally-chosen for being different from all the other sequences, isn't it.... whereas overlap hybridization is just sharing relevant sequence between multiple strands.
13:40 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-egeeezwmfnjhahee] has quit [Quit: Connection closed for inactivity]
13:43 -!- nmz787_i1 [ntmccork@nat/intel/x-yjgbgkbahcqbepnl] has joined ##hplusroadmap
13:46 -!- nmz787_i [ntmccork@nat/intel/x-fmxyjytcpyrvqtsx] has quit [Ping timeout: 252 seconds]
14:11 -!- Acty [uid89656@gateway/web/irccloud.com/x-sozxtwlywxurstxe] has joined ##hplusroadmap
14:15 -!- Madplatypus [uid19957@gateway/web/irccloud.com/x-ktdebwyjroxkwhns] has joined ##hplusroadmap
14:15 -!- nmz787_i [~ntmccork@134.134.139.72] has joined ##hplusroadmap
14:16 -!- nmz787_i1 [ntmccork@nat/intel/x-yjgbgkbahcqbepnl] has quit [Ping timeout: 248 seconds]
14:20 -!- genehacker [~chatzilla@128-193-154-1.ptpg.oregonstate.edu] has quit [Ping timeout: 248 seconds]
14:24 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has quit [Ping timeout: 246 seconds]
14:25 -!- nottimschmidt [~timschmid@13-181-165.client.wireless.msu.edu] has joined ##hplusroadmap
14:37 -!- nmz787_i1 [ntmccork@nat/intel/x-qnyydaamforoisgn] has joined ##hplusroadmap
14:39 -!- nmz787_i [~ntmccork@134.134.139.72] has quit [Read error: Connection reset by peer]
14:41 -!- nmz787_i [~ntmccork@134.134.139.72] has joined ##hplusroadmap
14:43 -!- nmz787_i1 [ntmccork@nat/intel/x-qnyydaamforoisgn] has quit [Ping timeout: 256 seconds]
14:55 -!- genehacker [~chatzilla@128-193-152-178.ptpg.oregonstate.edu] has joined ##hplusroadmap
14:59 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Remote host closed the connection]
14:59 < nmz787_i> if there is an API call I'd like to block in a Windows executable, how would I skip it using a hexeditor? replace the call with a NOP or something?
15:03 -!- nottimschmidt [~timschmid@13-181-165.client.wireless.msu.edu] has quit [Ping timeout: 256 seconds]
15:07 -!- eudoxia [~eudoxia@r167-56-36-161.dialup.adsl.anteldata.net.uy] has quit [Quit: Leaving]
15:09 < juri_> nmz787_i: on the laser/CNC front, remember that i've been here to help with CNC and modeling for some time. just building up skills, waiting for this place to need me.
15:13 < juri_> I'm still no biologist, but when it comes to plastic printers, i know a thing or five hundred.
15:18 < nmz787_i> juri_: any noticeable improvement with the CAD repo? I remember the last time I tried it, it was giving pretty poor quality on the small features I was designing, and would have random 'holes' in the mesh
15:19 < juri_> nothing very notable. I'm still polishing the code, to make it easy enough for me to make radical changes.
15:20 < juri_> learning haskell, while fun, has been grueling.
15:21 -!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap
15:23 < kanzure> nmz787_i: yes, you can insert nops
15:24 < kanzure> also, use idapro
15:34 -!- MrHindsight [~not_sure@unaffiliated/capthindsight] has quit [Quit: Leaving]
15:36 -!- nottimschmidt [~timschmid@2601:405:4101:52db::822] has joined ##hplusroadmap
15:37 -!- Darius [~quassel@75.76.19.18] has joined ##hplusroadmap
15:45 -!- genehacker [~chatzilla@128-193-152-178.ptpg.oregonstate.edu] has quit [Ping timeout: 256 seconds]
15:59 -!- genehacker [~chatzilla@8-44.ptpg.oregonstate.edu] has joined ##hplusroadmap
16:12 -!- genehacker [~chatzilla@8-44.ptpg.oregonstate.edu] has quit [Ping timeout: 250 seconds]
16:15 -!- genehacker [~chatzilla@8-44.ptpg.oregonstate.edu] has joined ##hplusroadmap
16:15 -!- ant4t [~s3an@50.141.78.33] has quit [Ping timeout: 264 seconds]
16:24 < nmz787_i> kanzure: my nature-based calendar tells me a plover is a bird that hangs out in the NYC area
16:24 < nmz787_i> .wik plover
16:24 < yoleaux> "Plovers (/ˈplʌvər/ or /ˈploʊvər/) are a widely distributed group of wading birds belonging to the subfamily Charadriinae." — https://en.wikipedia.org/wiki/Plover
16:28 -!- ant4t [~s3an@50.141.77.140] has joined ##hplusroadmap
16:34 -!- genehacker [~chatzilla@8-44.ptpg.oregonstate.edu] has quit [Ping timeout: 250 seconds]
16:56 < kanzure> nmz787_i: yes, it's a bird. the plover chording system was named after the bird.
16:56 < kanzure> nmz787_i: why can't i find any evidence of single-molecule dilutions + pumping in the literature?
16:56 -!- ant4t [~s3an@50.141.77.140] has quit [Ping timeout: 256 seconds]
17:01 < kanzure> isn't this some sort of brownian motion thermodynamics limit
17:01 < kanzure> andytoshi: aren't you a physics person
17:04 -!- erasmus [~esb@unaffiliated/erasmus] has joined ##hplusroadmap
17:11 < nmz787_i> kanzure: the dilution thing is pretty common
17:11 < nmz787_i> that's how they dilute hybridomas for making monoclonal antibodies
17:12 < nmz787_i> s/dilute/isolate/
17:12 < CaptHindsight> we can start with an aluminum oxide filter with ~20nm pores
17:12 < nmz787_i> CaptHindsight: for?
17:12 < CaptHindsight> push the base through, and catch them a few at a time
17:13 < CaptHindsight> then we send that through a second pore and isolate the single molecules
17:14 < nmz787_i> wouldn't that just act to slow the flow down though? they aren't that much larger than a solvent molecule
17:15 < nmz787_i> hrmm, I've only ever really thought about it in terms of dilution and watching for it to pass by
17:15 < nmz787_i> aside from the molecular pump stuff, which I've stayed away from thinking much about
17:15 < CaptHindsight> well dilution helps to reduce the number of the molecules you want to isolate
17:17 < kanzure> isn't a single molecule pump supposed to be one of those impossible maxwell demon problems
17:17 < kanzure> .wik maxwell's demon
17:17 < CaptHindsight> if it's diluted you just watch for a single molecule to come through and then stop the flow
17:17 < yoleaux> "In the philosophy of thermal and statistical physics, Maxwell's demon is a thought experiment created by the physicist James Clerk Maxwell in which he suggested how the Second Law of Thermodynamics could hypothetically be violated. In the thought experiment, a demon controls a small door between two chambers of gas." — https://en.wikipedia.org/wiki/Maxwell%27s_demon
17:17 < nmz787_i> nah, those ion pumps do it all the time
17:17 < CaptHindsight> you can two filters act as a gate
17:17 < nmz787_i> even polymerase is a sort of nucleotide pump
17:18 < CaptHindsight> or build a polymer that loops
17:18 < CaptHindsight> aluminum oxide is easy to grow and have isolated pores
17:20 < kanzure> hmm so nevermind about it being impossible to isolate single (small) molecules in that way, perhaps it was single leptons or smaller particles
17:21 < nmz787_i> heh, answer... the demon generates entropy, so all is well
17:21 < nmz787_i> I wonder how many scientific articles mention words like demon... or more importantly, wizard
17:21 < kanzure> .gc site:nature.com wizard
17:21 < yoleaux> kanzure: Sorry, that command (.gc) crashed.
17:22 < CaptHindsight> or make a polymer protein hybrid that can act as a filter and gate
17:22 < kanzure> .gc site:sciencedirect.com wizard
17:22 < yoleaux> kanzure: Sorry, that command (.gc) crashed.
17:22 < nmz787_i> 7690
17:22 < CaptHindsight> electrically controlled
17:22 < kanzure> CaptHindsight: there are many membrane-bound proteins that act as filter and gate
17:22 < nmz787_i> CaptHindsight: yep, ion channel
17:22 < nmz787_i> .wik gated ion channel
17:22 < yoleaux> nmz787_i: Sorry, that command (.wik) crashed.
17:22 < nmz787_i> pfft
17:22 < CaptHindsight> so whats the problem then?
17:22 < kanzure> dpk: fixplz?
17:22 < kanzure> CaptHindsight: nobody has demonstrated it
17:22  * dpk kill kill kill Wikipedia
17:23 < nmz787_i> placing the ion channel in the right place/orientation
17:23 < kanzure> dpk: also gc is crashing
17:23 < CaptHindsight> oh you mean in nature there are ion channels
17:23 < CaptHindsight> sure
17:23 < kanzure> indeed, there are many many types of membrane-bound channels
17:23 < andytoshi> kanzure: i minored in physics
17:23 < andytoshi> what's up?
17:23 < nmz787_i>  nature pretty much has everything
17:23 < kanzure> andytoshi: single molecule pump. impossible?
17:23 < andytoshi> lol, hmm
17:23 < kanzure> small molecules, not giant dna molecules
17:23 < nmz787_i> proton pumps are like first-year topics
17:24 < andytoshi> hehe nmz787_i
17:24 < CaptHindsight> I'm saying we attach polymers to it to control them
17:24 < nmz787_i> .title http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2742554/
17:24 < yoleaux> Ion channels versus ion pumps: the principal difference, in principle
17:24 < andytoshi> i am definitely not more qualified than anyone else ..
17:24 < kanzure> qualifications are overrated anyway
17:24 -!- CyberEater [~Genesteal@c-69-255-209-151.hsd1.md.comcast.net] has joined ##hplusroadmap
17:25 < nmz787_i> https://en.wikipedia.org/wiki/Proton_pump
17:25 < kanzure> .wik proton pump
17:25 < yoleaux> "A proton pump is an integral membrane protein that is capable of moving protons across a biological membrane. Mechanisms are based on conformational changes of the protein structure or on the Q cycle." — https://en.wikipedia.org/wiki/Proton_pump
17:25 < nmz787_i> smallest molecule pump
17:25 -!- Darius [~quassel@75.76.19.18] has quit [Ping timeout: 255 seconds]
17:25 < CaptHindsight> graft polymerization onto an ion pump
17:25 < kanzure> protons are not molecules
17:25 < nmz787_i> and all the studies on RNA tugging kanzure
17:25 < nmz787_i> what!
17:25 < nmz787_i> well, I guess atoms aren't molecules
17:26 < nmz787_i> but molecules are atoms
17:26 < kanzure> molecules are atoms? are we sure about this :-)
17:26 < kanzure> they are made of atoms.
17:26 < nmz787_i> anyway, DNA tugging is a great google scholar search, I'm sure ;p
17:26 < kanzure> i'm not sure why i still can't find a single molecule pump in a machine
17:26 < CaptHindsight> quantum movement of base molecules
17:27 < CaptHindsight> kanzure: make one, somebody should
17:27 < nmz787_i> what's that cholesterol pump
17:28 < nmz787_i> brocolli induces it
17:28 < nmz787_i> brocolli
17:28 < kanzure> nanometer-scale molecular pump http://rafiki.tau.ac.il/~rabani/papers/paper44.pdf
17:28 < nmz787_i> broccoli
17:29 < nmz787_i> .wik spp1
17:29 < yoleaux> "SPP1 or SPP-1 may refer to:" — https://en.wikipedia.org/wiki/Spp1
17:29 < CaptHindsight> broccoli induced quantum teleporation of fats
17:29 < CaptHindsight> my next research project  :)
17:29 < nmz787_i> jrayhawk: you would probably remember that one... it was a bitter molecule that induced the pump
17:29 < kanzure> what's the gas transition temperature for nucleotides
17:29 < kanzure> or er.. what's the right question
17:30 < nmz787_i> jrayhawk: which I think might have been related to the whole removing old cholesterol before it could oxidize in-vivo
17:30 < jrayhawk> not offhand
17:31 < nmz787_i> ugh, damn gene acronyms
17:31 < CaptHindsight> so nobody has made an electrostatic base gun yet?
17:33 < kanzure> nmz787_i: one way to do a nucleotide pump would be the following.... get a protein that attaches to a single nucleotide, make the protein super massive and large, then manipulate that protein instead of the nucleotide. this can also work for other types of tags.
17:34 < kanzure> also possibly do laser conformational change to release the nucleotide
17:34 < CaptHindsight> a bead for single nucleotide
17:35 < kanzure> oh right, you wouldn't be able to guarantee the reaction
17:35 < kanzure> even after dissosciation
17:35 < kanzure> dissociation
17:35 < CaptHindsight> oh, small pore filters  https://www.bu.edu/meller/PDFs/AdvMat18_3149.pdf
17:36 < nmz787_i> kanzure: ion concentration could also work to release the molecule
17:36 < kanzure> CaptHindsight: yes that is the nanopore dna sequencing method
17:36 < CaptHindsight> kanzure: can also isolate single molecules
17:37 < nmz787_i> CaptHindsight: i've thought of using nano wires for programming hybridization areas...until I realized each wire would have to be the size of an atom (so they could hybridize)
17:37 < nmz787_i> possibly using a vacuum and electron gun could work
17:37 < nmz787_i> but then you'd need a reliable way to get the electron to stay in place and not leak off
17:38 < nmz787_i> and also a way to way the pattern with nucleotides
17:38 < nmz787_i> then you'd need to, umm, cryo freeze and add ligase
17:38 < nmz787_i> which also doesn't work in frozen non-aqueous conditions
17:38 < kanzure> "Mycobacterium smegmatis porin A (MspA) is the second biological nanopore currently being investigated for DNA sequencing. The MspA pore has been identified as a potential improvement over αHL due to a more favorable structure.[9] The pore is described as a goblet with a thick rim and a diameter of 1.2 nm at the bottom of the pore.[10] A natural MspA, while favorable for DNA sequencing because of shape and diameter, has a negative core ...
17:38 < nmz787_i> CaptHindsight: the big problem is positioning the nanopore.
17:39 < kanzure> ... that prohibited single stranded DNA(ssDNA) translocation. The natural nanopore was modified to improve translocation by replacing three negatively charged aspartic acids with neutral asparagines.[11]"
17:40 < CaptHindsight> nmz787_i: which problem are you trying to solve? Maybe I'm thinking of something else
17:40 < kanzure> "Another method is the use of nanoelectrodes on either side of a pore.[17][18] The electrodes are specifically created to enable a solid state nanopore's formation between the two electrodes. This technology could be used to not only sense the bases but to help control base translocation speed and orientation."
17:40 < nmz787_i> http://www.ncbi.nlm.nih.gov/pmc/articles/PMC413341/pdf/emboj00076-0052.pdf
17:41 < nmz787_i> The portal protein of bacteriophage SPP1: a DNA pump with 13-fold symmetry
17:41 < kanzure> nmz787_i: here you go, try this http://www.nist.gov/pml/div684/single_092209.cfm
17:42 < kanzure> "The water’s abrupt pressure drop—accompanied by a dash of detergent—breaks its surface tension, splitting it into small droplets.  ..... In the microfluidic channel, the water is laced with desired molecules of just the right concentration, so that resulting droplets each pick up on average just one molecule of interest."
17:42 < kanzure> oh... "Inside each droplet, the individual molecules of interest slosh around freely in the relatively roomy sphere"
17:42 < kanzure> so is it just one or not? >:(
17:42 < kanzure> the paper is http://proceedings.spiedigitallibrary.org/proceeding.aspx?articleid=784816
17:44 < kanzure> "Generation and mixing of subfemtoliter aqueous droplets on demand"
17:44 < kanzure> "Surface acoustic waves for on-demand production of picoliter droplets and particle encapsulation"
17:44 < kanzure> http://www.researchgate.net/profile/David_Collins15/publication/240306158_Surface_acoustic_waves_for_on-demand_production_of_picoliter_droplets_and_particle_encapsulation/links/0a85e539a8257e3a77000000.pdf
17:45 < kanzure> probably a piezo could do this too
17:45 < nmz787_i> well it's still just concentration based though
17:45 < nmz787_i> the droplet is not required
17:45 < kanzure> another description http://spie.org/x39621.xml?highlight=x2400
17:45 < nmz787_i> I can't remember if there are materials that only let water pass
17:45 < nmz787_i> as a way of concentrating the droplets
17:46 < nmz787_i> is tyvek water or oxygen?
17:46 < nmz787_i> .wik tyvek
17:46 < yoleaux> "Tyvek /taɪˈvɛk/ is a brand of flashspun high-density polyethylene fibers, a synthetic material; the name is a registered trademark of DuPont. It is often seen used as housewrap, a synthetic material used to protect buildings during construction." — https://en.wikipedia.org/wiki/Tyvek
17:46 < kanzure> "Using our devices, we have been able to produce single droplets with volumes under 1 femtoliter, suitable for performing single-molecule studies"
17:47 < kanzure> .title http://pubs.acs.org/doi/abs/10.1021/ac9014319
17:47 < yoleaux> An Error Occurred Setting Your User Cookie
17:47 < kanzure> "Generation and Mixing of Subfemtoliter Aqueous Droplets On Demand"
17:47 < kanzure> "by means of piezoelectric injection"
17:47 < nmz787_i> hah, they didn't use a global-shutter camera for that paper
17:47 < nmz787_i> the images have motion blur
17:48 < nmz787_i> err, interlacing artifacts
17:48 < kanzure> CaptHindsight: have you ever inkjet printed an emulsion?
17:49 < nmz787_i> kanzure: a flying droplet would dry!
17:49 < nmz787_i> (i'm saying that in a good way)
17:50 < kanzure> wouldn't dry in the right environment
17:51 < kanzure> anyway; i feel a little bit better now that at least one group has done a single molecule isolation system.
17:52 < CaptHindsight> kanzure: sure, that's how we get UV cured inks with thermal inkjet
17:52 < kanzure> ah good
17:52 < CaptHindsight> even easier with piezo
17:53 < kanzure> if your nozzle prints out too many emulsion droplets, you can maybe fix it by moving the head faster (so that they land in different spots anyway)
17:54 < CaptHindsight> aerosol jet
17:55 < CaptHindsight> can you use a low vapor pressure vehicle/solvent?
17:56 < CaptHindsight> then the droplet would not dry as quickly
17:56 < kanzure> CaptHindsight: how do you feel about money being used as a carrot for certain performance/testing targets? :P
17:57 < CaptHindsight> what target?
17:57 < CaptHindsight> and carrot for whom?
17:57 < CaptHindsight> whos money  :)
17:57 < kanzure> mine
17:58 < kanzure> well there's a blank section in the doc, so basically i'm figuring out money stuff- there's lots of options, like pay-at-completion, milestones, performance/testing targets, schedule, etc.
17:59 < CaptHindsight> oh for the inkjet?
17:59 < kanzure> yeah
17:59 < CaptHindsight> oh it's mostly for materials
17:59 < CaptHindsight> so it's up front
17:59 < CaptHindsight> for whatever materials
18:00 < CaptHindsight> cut it up in chinks if you want
18:00 < CaptHindsight> we work when there's funding
18:00 < CaptHindsight> whatever you are comfortable with
18:01 < kanzure> yeah i was thinking about that, and i'd be okay with upfront only as long as you grant me the option of having you build something else on the same granty if we can't figure out the chemistry
18:02 < kanzure> or, alternatively, if we can't order the chemistry for some reason (who knows- these people are fickle)
18:02 < nmz787_i> .wik nrec
18:02 < kanzure> s/fickle/absurd and they don't ship to people who are willing to pay
18:02 < yoleaux> "The National Robotics Engineering Center (NREC) is an operating unit within the Robotics Institute (RI) of Carnegie Mellon University." — https://en.wikipedia.org/wiki/National_Robotics_Engineering_Center
18:02 < nmz787_i> pfft
18:02 < nmz787_i> not what I meant
18:02 < CaptHindsight> kanzure: are you concerned about getting the chems?
18:02 < kanzure> 20% concerned?
18:02 < CaptHindsight> what are their hoops to jump through?
18:03 < nmz787_i> for a place like Sigma, you need to show a lease on a commercial property
18:03 < kanzure> nmz787_i: i'm thinking azco biotech
18:03 < kanzure> do we want liquid or solid reagents anyway?
18:04 < CaptHindsight> yeah, I jumped through their hoops years ago
18:04 < kanzure> whoops we can't buy liquid phosphoramidites from azco biotech
18:04 -!- ant4t [~s3an@50.141.79.168] has joined ##hplusroadmap
18:04 < nmz787_i> the POSAM paper said basically 'mix at your own risk (of getting the mix contaminated with O2 or water)
18:04 < CaptHindsight> kanzure: what else do you want to mount on the gantry?
18:04 < kanzure> https://azcobiotech.com/reagents-for-oligonucleotide-synthesis/phosphoramidites/dna-phosphoramidites/
18:05 < CaptHindsight> I have an account with Sigma
18:05 < kanzure> oh interesting
18:05 < nmz787_i> .title http://probesoftware.com/smf/index.php
18:05 < yoleaux> Probe Software Users Forum - Index
18:05 < kanzure> a sigma account is very helpful
18:05 < nmz787_i> (electron microprobe)
18:05 < CaptHindsight> don't use it since they charge 10-100x what I usually get things for
18:06 < kanzure> i think you might have to spend some time on the phone with these guys to figure out the bottle geometry and cap system, https://azcobiotech.com/cepa-bulk-dg-ibu-20-8110.html
18:07 < CaptHindsight> yeah saw those
18:07 < kanzure> huh, azco biotech claims to offer machine design services too. we should take advantage of that eventually.
18:08 < CaptHindsight> they probably have a 3rd party
18:08 < kanzure> "In 2014 the name of Azco Biotech was purchased by Global DNA Solutions, Inc. a California Corporation.  Today Global DNA Solutions, Inc. is doing business as AZCO Biotech"
18:09 < kanzure> nmz787_i: other fun things to use the gantry for if the phosphoramidite chemistr fails us?
18:10 < nmz787_i> lithography
18:10 < nmz787_i> put a bluray laser on the gantry and start whipping out microfluidics
18:11 < nmz787_i> (that version of the machine can come live near me)
18:11 < CaptHindsight> AZCO let me get to the checkout
18:11 < kanzure> so bluray laser microfluidics stuff, or maskless photolithography stuff?
18:11 < CaptHindsight> just some billing info a CC to pay
18:12 < nmz787_i> CaptHindsight: lots of sites will let you get there, then follow up with a phone call
18:12 < CaptHindsight> just a UPS verification of address
18:12 < kanzure> nmz787_i: he has a business at his location
18:12 < nmz787_i> yeah so it could be fine
18:12 < nmz787_i> there is no licensing AFAIK
18:13 < nmz787_i> they mostly just wanted to know that the biz was my biz
18:13 < nmz787_i> or i was part of the company, and that i was allowed to receive chem orders there
18:13 < CaptHindsight> 2 photon polymerization
18:14 < CaptHindsight> lithography and microfluidic fabrication
18:14 < CaptHindsight> I make all the photopolymers as well
18:15 < CaptHindsight> DNA anti counterfeiting marking
18:15 < CaptHindsight> I thought that the POSaM was proven to work years ago
18:16 < CaptHindsight> is there some secret sauce that they didn't mention or are they "pretending" that it works?
18:17 -!- proofoflogic [uid65184@gateway/web/irccloud.com/x-wqcoozsjjgnawftq] has quit [Quit: Connection closed for inactivity]
18:17 < kanzure> it works but there's lots of tweaking of everything
18:18 < nmz787_i> it works but sucks for fidelity and isn't easier than the previous tech
18:18 < CaptHindsight> isn't there a QC method using a CCD for testing each layer?
18:19 < nmz787_i> unless you're interested in arrays for microarray hybridization experiments
18:19 < nmz787_i> where lack of fidelity is usually more tolerable
18:19 < CaptHindsight> doesn't it fluoresce when the oxidation is occurring?
18:19 < nmz787_i> (because of signal to noise I guess)
18:19 < nmz787_i> the deblocking step is orange I think
18:19 < nmz787_i> the trityl group
18:20 < nmz787_i> but that isn't sensitive enough to show if one molecule skipped deblocking
18:20 < nmz787_i> and also the acid can eat away the side of the base, destroying the information-carrying capacity
18:21 < kanzure> there are also other problems like argon or nitrogen blowing away your samples, because they weren't linked to the surface well enough
18:21 < CaptHindsight> kanzure: if you're concerned about what to do with a flatbed inkjet take a look at SGIA  :)
18:21 < kanzure> or smearing with fingerprints causing nuclease contamination
18:21 < kanzure> SGIA?
18:21 < CaptHindsight> https://www.sgia.org/
18:22 < CaptHindsight> t-shirts, posters, banners etc
18:22 < CaptHindsight> printed electronics
18:22 < kanzure> hah, no i was thinking more along the lines of "what about printing yeast and ecoli cells"
18:22 < CaptHindsight> sell it to someone for that application
18:22 < kanzure> oh would they buy?
18:22 < nmz787_i> jrayhawk: I think it was Chris Masterjohn... but I cannot find it using 'chris masterjohn bitter induction' or 'bitter vegetables cholesterol "pump" masterjohn brocolli'
18:23 < CaptHindsight> the dimatix flatbed that size is like $40k
18:23 < kanzure> sigh... i should have guessed.
18:24 < nmz787_i> jrayhawk: I want to think it is close to irk2 or some acronym like that
18:24 < kanzure> oh right, so before i add in the upfront section, i need to figure out the future home for the bugger, i'll try to figure that out today i guess
18:24 < kanzure> nmz787_i: want this thing?
18:24 < kanzure> nmz787_i: would you operate it
18:25 -!- erasmus [~esb@unaffiliated/erasmus] has quit [Read error: Connection reset by peer]
18:27 < nmz787_i> kanzure: hrmm, all the chems having short shelf life seems stress-inducing
18:27 < kanzure> it's a lie
18:27 < kanzure> they are going to be opened in a neutral atmosphere
18:27 < kanzure> don't worry about it, the chemicals are like $9/bottle or something, i can pay that :P
18:28 < nmz787_i> didn't we already have a price list, and it was like $1200 per batch?
18:28 < kanzure> that might have been sigma prices
18:28 < kanzure> you're thinking of the price list from http://diyhpl.us/~bryan/nucleic/fbi-diybio-dna-v1.pdf
18:28 < kanzure> 3rd to last slide
18:29 < nmz787_i> I thought it was Glen Research
18:29 < kanzure> entirely possible
18:29 < nmz787_i> ah, yeah, that's the list
18:29 < nmz787_i> it's clearly a screenshot of google sheets
18:30 < nmz787_i> also the posam paper seemed to list an alternative two-component solvent to take the place of acetonitrile
18:30 < nmz787_i> bbl
18:30 < kanzure> wait you didn't answer
18:33 -!- nmz787_i [~ntmccork@134.134.139.72] has quit [Read error: Connection reset by peer]
18:33 < ParahSailin> is medical marijuana a real thing, or just a backdoor legalization?
18:34 < ParahSailin> second, if its useful for cancer patient relief, how is it generally recommended?
18:34 < ParahSailin> take it during a chemo session?
18:35 < ParahSailin> or only the rest of the time
18:35 < CaptHindsight> kanzure: http://www.myprintresource.com/product/10208199/fujifilm-north-america-corporation-fujifilm-dimatix-dmp-5000  no longer in production
18:36 < CaptHindsight> only 500 x 500 mm print area and that one was closer to $100K
18:37 -!- CyberEater [~Genesteal@c-69-255-209-151.hsd1.md.comcast.net] has quit [Ping timeout: 256 seconds]
18:37 -!- CyberEater [~Genesteal@c-69-255-209-151.hsd1.md.comcast.net] has joined ##hplusroadmap
18:38 < kanzure> CaptHindsight: don't the $200k machines come with support contracts etc.?
18:38 < kanzure> and on-site maintenance contracts
18:38 < CaptHindsight> they come with whatever they get you to pay for
18:38 < CaptHindsight> that is usually above the price of the printer in the fine print
18:39 < CaptHindsight> or they lease it to you with a service contract built in
18:39 -!- proofoflogic [uid65184@gateway/web/irccloud.com/x-qxwacvqkttuiwzxc] has joined ##hplusroadmap
18:40 < CaptHindsight> and they watch you like a hawk to see if have done anything like use a 3rd party ink
18:40 < CaptHindsight> so you void the service plan
18:40 -!- genehacker [~chatzilla@8-44.ptpg.oregonstate.edu] has joined ##hplusroadmap
18:43 < kanzure> ah interesting
18:43 < kanzure> it's about the same with dna synthesizers for that matter
18:43 < CaptHindsight> does it still need a home?
18:44 < kanzure> possibly- nathan ran off
18:47 < CaptHindsight> any private rooms at the nearby hackerspace?
18:48 < juri_> i imagine BUGS would enjoy it.
18:49 < juri_> and i can get there via the metro+bus, to operate.
18:51 < kanzure> juul would probably take it at counterculture labs
18:52 < CaptHindsight> https://counterculturelabs.org/ wow
18:52 < kanzure> superkuh: do you want it?
19:04 -!- genehacker [~chatzilla@8-44.ptpg.oregonstate.edu] has quit [Ping timeout: 264 seconds]
19:08  * superkuh reads the backlog.
19:10 < superkuh> No.
19:14 -!- knobuddy [~Razk@unaffiliated/knobuddy] has joined ##hplusroadmap
19:24 -!- erasmus [~esb@unaffiliated/erasmus] has joined ##hplusroadmap
19:27 -!- ant4t [~s3an@50.141.79.168] has quit [Ping timeout: 240 seconds]
19:33 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Quit: Leaving]
19:35 -!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap
19:44 -!- genehacker [~chatzilla@8-44.ptpg.oregonstate.edu] has joined ##hplusroadmap
19:54 < nmz787> bak
19:57 -!- CyberEater [~Genesteal@c-69-255-209-151.hsd1.md.comcast.net] has quit [Quit: Leaving]
19:57 -!- genehacker [~chatzilla@8-44.ptpg.oregonstate.edu] has quit [Ping timeout: 246 seconds]
19:57 -!- Manti5 [~Genesteal@c-69-255-209-151.hsd1.md.comcast.net] has joined ##hplusroadmap
19:58 -!- Manti5 [~Genesteal@c-69-255-209-151.hsd1.md.comcast.net] has left ##hplusroadmap []
20:07 < kanzure> nmz787: you haven't answered either
20:18 -!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-knzxqmqtwcqwsjcp] has joined ##hplusroadmap
20:20 < nmz787> kanzure CaptHindsight: why not aim for a sub-$500 CNC with 0.1-0.5 uM positioning repeatability, using interferometry (requiring, most likely, to either utilize the optics in the drives somehow (it /might/ be possible) or injection molding for plastic optics) or machine vision using a grid-recticle (but I imagine the feedback would be quite slow)
20:20 < kanzure> for the dnajet?
20:21 < nmz787> is that price-point not possible for that accuracy? in production, or rather once production can be achieved (through prototyping, making molds, etc)
20:21 < nmz787> it seems like if you could make a smaller one with more accuracy, if making them smaller would make the machine cheaper
20:22 < nmz787> then you could have a usable thing for litho, pipetbot, etc... and then go on to prototype dnajet chem
20:22 < nmz787> or straight to it
20:22 < nmz787> i don't think more than 25cm x 25cm machine is necessary (that's an 8 inch wafer)
20:23 < nmz787> (actually 8 inch is 20.32 cm)
20:24 < nmz787> the mesa cards seem quite expensive
20:24 < kanzure> if you don't want this thing, just say so
20:24 < nmz787> but that can be reduced later, and keep the fabric right?
20:24 < nmz787> or HDL?
20:24 < nmz787> verilog?
20:24 < kanzure> i can ship it to someone else
20:24 < kanzure> it's vhdl stuff
20:24 < kanzure> actually it's probably verilog but i hope it's systemc.
20:25 < nmz787> would this fit on an ice40 CaptHindsight or juri_ ? if not, how undersized is an ice40?
20:25 -!- Darius [~quassel@75.76.19.18] has joined ##hplusroadmap
20:25 < kanzure> (sleepmode)
20:26 < nmz787> kanzure: I don't think I have time to do the chem for the machine right now
20:26 < CaptHindsight> then you have to make boards, cables etc
20:26 < kanzure> nmz787: you wouldn't have to do the chemistry
20:26 < nmz787> kanzure: it is hard for me to make time, as the value proposition doesn't seem like a good ratio of effort to payoff
20:26 < kanzure> nmz787: although you would be asked to operate it (turn it on) and perform maintenance (like switching out chemicals)
20:26 < nmz787> or rather, the best ratio
20:26 < kanzure> and wifi config stuff
20:27 < nmz787> who will do the chem debug then?
20:27 < kanzure> 1.2 million spots/chip is not good ratio of effort to payoff?
20:27 < CaptHindsight> drivers, maybe a new CNC motion application....
20:27 < nmz787> yeah but the spots are not that great afterward... you still have tons of effort left
20:27 < kanzure> nmz787: i think captain is going to be doing the chemistry debugging
20:28 < nmz787> if it's "turn the thing on, and load chemicals" then it could sit here... but I think one of the biohacker spaces would have far more effort freely available that would be interested
20:29 < nmz787> I imagine
20:29 < nmz787> have you asked keoni?
20:29 < kanzure> keowho?
20:29 < nmz787> koeng
20:29 < kanzure> keoni gandall?
20:30 < kanzure> don'tk now this person
20:30 < nmz787> yeah
20:30 < nmz787> he has a bio-internship for a while now at UCLA I think, and hangs out with L.A. biohackers
20:30 < kanzure> oh the los angeles people, yeah i guess we could ask them
20:30 < kanzure> cory would be a good home for this
20:31 < nmz787> or maybe it's irvine
20:31 < nmz787> yeah
20:31 < nmz787> UCI
20:32 < kanzure> ok just sent an email to cory
20:33 < kanzure> jrayhawk: you want to run this?
20:35 < kanzure> fenn: also haven't asked you yet
20:36 -!- Madplatypus [uid19957@gateway/web/irccloud.com/x-ktdebwyjroxkwhns] has quit [Quit: Connection closed for inactivity]
20:39 < kanzure> 
20:39 < kanzure> jojack might be a good person to keep this, although i wouldn't expect him to do chemistry
20:44 < jrayhawk> Run what with the who now?
20:55 -!- Viper168_ [~Viper@unaffiliated/viper168] has joined ##hplusroadmap
20:57 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 252 seconds]
21:12 -!- Quashie [~boingredd@50.14.92.17] has joined ##hplusroadmap
21:18 -!- Quashie [~boingredd@50.14.92.17] has quit [Ping timeout: 264 seconds]
21:19 -!- knobuddy is now known as JimmyRussel
21:20 -!- JimmyRussel is now known as JamesRussel
21:20 -!- JamesRussel is now known as JimmySRussel
21:24 -!- Madplatypus [uid19957@gateway/web/irccloud.com/x-dvyjjsdompilywnq] has joined ##hplusroadmap
21:38 -!- Viper168_ is now known as Viper168
21:44 -!- Darius [~quassel@75.76.19.18] has quit [Remote host closed the connection]
21:49 -!- nottimschmidt [~timschmid@2601:405:4101:52db::822] has quit [Ping timeout: 248 seconds]
21:53 < kanzure> jrayhawk: dna synthesis equipment
22:04 < jrayhawk> i do not have a fume hood, no
22:05 < jrayhawk> if you want to buy me a fume hood, maybe
22:06 < jrayhawk> i have like zero lab experience; i suspect diybio folks would be a saner choice
22:25 < nsh> .t http://www.wired.com/2015/07/paradoxical-crystal-baffles-physicists/
22:25 < yoleaux> nsh: Sorry, I don't know a timezone by that name.
22:26 < nsh> .title
22:26 < yoleaux> Paradoxical Crystal Baffles Physicists | WIRED
22:26 < nsh> --
22:26 < nsh> The material, a much-studied compound called samarium hexaboride or SmB6, is an insulator at very low temperatures, meaning it resists the flow of electricity. Its resistance implies that electrons (the building blocks of electric currents) cannot move through the crystal more than an atom’s width in any direction. And yet, Sebastian and her collaborators observed electrons traversing orbits millions of atoms in diameter inside the crystal in response to
22:26 < nsh>  a magnetic field—a mobility that is only expected in materials that conduct electricity. Calling to mind the famous wave-particle duality of quantum mechanics, the new evidence suggests SmB6 might be neither a textbook metal nor an insulator, Sebastian said, but “something more complicated that we don’t know how to imagine.”
22:26 < nsh> --
23:06 < fenn> yep PCR barcode is meant to select a subset of the synthesis pool so it should have unique and maximally distant sequences
23:07 < fenn> i am personally not interested in running this dna synthesis machine and i don't have the facilities for it either
23:09 < fenn> i dont get why you're still talking about single molecule pumps
23:15 < fenn> eventually you will reinvent the ribosome
23:21 -!- Acty [uid89656@gateway/web/irccloud.com/x-sozxtwlywxurstxe] has quit [Quit: Connection closed for inactivity]
23:29 < fenn> if dna synthesis doesnt work out, the inkjet system could be useful for printing electronics; organic semiconductors and 3d circuit boards
23:37 -!- proofoflogic [uid65184@gateway/web/irccloud.com/x-qxwacvqkttuiwzxc] has quit [Quit: Connection closed for inactivity]
23:38 < fenn> fwiw counterculturelabs is right down the road from me and i sort of know kathy buehmann already (and juul i guess from this channel)
23:50 < fenn> if it's nearby i would be willing to get the machine set up in its new home and do some process debugging
--- Log closed Tue Jul 14 00:00:10 2015