2018.12.18 Dev Chat: Difference between revisions

From Eterna Wiki

(Created page with "<p>Jieux: ‎Good evening friends! ‎[5:55 PM]<br />Astromon: ‎heya Jieux! ‎[5:56 PM]<br />Jieux: ‎Heya Astromon, how’s flips? ‎[5:56 PM]<br />LucaV...")
(No difference)

Revision as of 00:12, 9 December 2018

Jieux: ‎Good evening friends! ‎[5:55 PM]
Astromon: ‎heya Jieux! ‎[5:56 PM]
Jieux: ‎Heya Astromon, how’s flips? ‎[5:56 PM]
LucaV: ‎well, rank #12960 I’m gonna call it a day ‎[5:56 PM]
Jieux: ‎Good job LucaV ‎[5:57 PM]
whbob: ‎hi ‎[5:57 PM]
Astromon: ‎heya Whabob ‎[5:57 PM]
Gerry Smith: ‎evening all ‎[5:57 PM]
Astromon: ‎heya gerry ‎[5:57 PM]
Jieux: ‎Heya whbob, Gerry Smith. ‎[5:58 PM]
JR: ‎I,m here if this is roll call ‎[5:58 PM]
whbob: ‎hi gerry, jieux,zama,astro ‎[5:58 PM]
whbob: ‎hi JR ‎[5:58 PM]
Astromon: ‎hey JR ‎[5:58 PM]
Jieux: ‎Heya JR, nice to have you in chat… ‎[5:58 PM]
JR: ‎Who else is here? ‎[5:59 PM]
Gerry Smith: ‎we should create a matrix where each person can say hi to each other ‎[5:59 PM]
Jieux: ‎CHATONLINE (17) ‎[5:59 PM]
Astromon: ‎haha ‎[6:00 PM]
rhiju: ‎hi everyone! thanks for coming to the community chat ‎[6:00 PM]
Jieux: ‎Heya rhiju! ‎[6:01 PM]
Astromon: ‎Heya rhiju ‎[6:01 PM]
AndrewKae: ‎Hi Astro, Jieux, JR, Gerry, whbob, Zama, lfp6 ‎[6:01 PM]
Gerry Smith: ‎all at once…hi rhiju ‎[6:01 PM]
rhiju: ‎let me see if some of the other devs are logging in ‎[6:01 PM]
whbob: ‎hi rhiju, andrew ‎[6:01 PM]
mgotrik: ‎hi all ‎[6:01 PM]
Astromon: ‎hello’s ‎[6:02 PM]
LFP6: ‎Back from dinner :) ‎[6:02 PM]
rhiju: ‎in addition to @mgotrik, some of the other eterna devs are with me in my office ‎[6:02 PM]
Astromon: ‎\o ‎[6:02 PM]
Jieux: ‎Hello lurking labcoats! ‎[6:02 PM]
rhiju: ‎feriel melaine, who is doing a bunch of openTB stuff, and hannah wayment-steele, a newer student who is interested in switch design principles ‎[6:02 PM]
hwaymentsteele: ‎Hello! ‎[6:03 PM]
rhiju: ‎we do have some currently secret news about the latest lab round! ‎[6:03 PM]
rhiju: ‎@mgotrik do you want to share? ‎[6:03 PM]
Astromon: ‎sweet ‎[6:03 PM]
rhiju: ‎and then we can figure out how all this goodness is happening ‎[6:03 PM]
rhiju: ‎oh, @mgotrik has to get on his computer again ‎[6:04 PM]
rhiju: ‎while he’s setting up, i can leak a little of the news ‎[6:04 PM]
rhiju: ‎mgotrik and ann kladwang synthesized the top-voted solutions for the ATP-Malachite Green challenges from lightning round 8 ‎[6:04 PM]
rhiju: ‎which you all made, and omei helped collect ‎[6:04 PM]
mgotrik: ‎drum roll and Jieux came up with a successful design for the “ATP/MG Same State” puzzle ‎[6:05 PM]
rhiju: ‎in parallel, they also synthesized and tested 6 solutions that emerged from in vitro evolution efforts from @mgotrik’s prior work ‎[6:05 PM]
rhiju: ‎haha, i’ll let @mgotrik take over ‎[6:05 PM]
Astromon: ‎nice Jieux! ‎[6:05 PM]
mgotrik: ‎So Jieux’s puzzle (ID: 9080567 ‎[6:06 PM]
rhiju: ‎@mgotrik anyway players can see preview links of those results? ‎[6:06 PM]
mgotrik: ‎worked, and we got about a 4-fold change in the presence/absence of ATP ‎[6:06 PM]
rhiju: ‎and how did ‘evolution’ do @mgotrik? ‎[6:07 PM]
Jieux: ‎You should test the original design from which I made a mod… pretty sure I only did mods in that lab I believe. ‎[6:07 PM]
mgotrik: ‎you can see the results of that puzzle here: https://imgur.com/a/yK4193u ‎[6:07 PM]
rhiju: ‎@jieux the title is ‘FiskerMod 8 #all3’ ‎[6:08 PM]
Jieux: ‎The original design should be mentioned in the comments. ‎[6:08 PM]
mgotrik: ‎It did better than the ‘evolution’ attempt I had done on it ~2 years ago. That one also had about a 3-fold change, but was not as bright as Jieux’s at lower concentrations of the fluorescent dye ‎[6:08 PM]
rhiju: ‎can someone post a link? ‎[6:08 PM]
rhiju: ‎there are some other pretty amazing cases, in prior lightning rounds, that achieved fold ratios of >100 ‎[6:09 PM]
Astromon: ‎which sublaB ‎[6:09 PM]
Astromon: ‎lab-puzzle^ ‎[6:09 PM]
rhiju: ‎@astromon the sublab for 9080567 is “ATP/MG Same State” ‎[6:09 PM]
Astromon: ‎im not finding it ‎[6:10 PM]
Jieux: ‎https://eternagame.org/game/browse/9035753/?filter1=Id&filter1_arg1=9080567&filter1_arg2=9080567 ‎[6:10 PM]
Jieux: ‎Think that is it. ‎[6:10 PM]
mgotrik: ‎the results for all of these can be seen here: https://imgur.com/a/yK4193u. the third image is the characterization results of the evolution-derived sequences ‎[6:10 PM]
rhiju: ‎we’re trying to now understand what principles (if any) led to these particular designs ‎[6:10 PM]
Astromon: ‎oh that one ok ‎[6:10 PM]
Astromon: ‎it has 3 gu ‎[6:11 PM]
rhiju: ‎its a spiral ‎[6:11 PM]
mgotrik: ‎i was really surprised to see a positive result for that puzzle, so great work once again. There were no successes with the other ATP Same State (MGA) puzzle which we tested ‎[6:11 PM]
rhiju: ‎jieux’s comment is: "Mod of “#spiral #CCW #SB #aptamerselfturnoff” by Eli Fisker ‎[6:12 PM]
rhiju: ‎@jieux do you remember how you made the mod? ‎[6:12 PM]
rhiju: ‎did you introduce the three gu’s ‎[6:12 PM]
Astromon: ‎has a static mga ‎[6:12 PM]
Astromon: ‎lowish on percentages ‎[6:12 PM]
JR: ‎ Has a strong switching area between aptamers ‎[6:13 PM]
LFP6: ‎Original design by Eli I think: https://eternagame.org/game/browse/9035753/?filter1=Id&filter1_arg1=9075529&filter1_arg2=9075529 ‎[6:14 PM]
rhiju: ‎thanks @LFP6 ‎[6:14 PM]
DigitalEmbrace: ‎The hairpin on the aptamer is static. ‎[6:15 PM]
rhiju: ‎anyone know how to bring up in arcplot tool and post screenshots? ‎[6:15 PM]
Jieux: ‎it does look like I introduced the UG pairs. ‎[6:15 PM]
Astromon: ‎http://prntscr.com/ls6gtl ‎[6:16 PM]
rhiju: ‎looks like @jieux switched one C-G pair from Eli’s design (at 18-45) to a U-G ‎[6:16 PM]
Astromon: ‎arc-plot pic^ ‎[6:16 PM]
rhiju: ‎@astromon thnaks ‎[6:16 PM]
rhiju: ‎it is indeed spirally (but messy) ‎[6:17 PM]
Jieux: ‎but we don’t know how Eli’s original design performed do we? ‎[6:17 PM]
Astromon: ‎will be interesting to see what eli’s version scores ‎[6:17 PM]
lroppy: ‎y ‎[6:17 PM]
rhiju: ‎no we don’t know how well eli’s design performed; any predictions? ‎[6:17 PM]
Jieux: ‎Hey #1 :) ‎[6:17 PM]
JR: ‎I’m surprised the messy designs do so well. ‎[6:17 PM]
lroppy: ‎yo #2 ‎[6:17 PM]
Astromon: ‎I predict jx’s does better ‎[6:17 PM]
mgotrik: ‎we do not know how the original design performed, correct. there are 4-5 designs by eli that match the puzzle name you commented, so it’ll be up to you to figure out what puzzle you started with ‎[6:18 PM]
Jieux: ‎Well my feeling about such things is that weaker pairs near the area of interest make designs more switchy. ‎[6:18 PM]
LFP6: ‎@mgotrik I linked the one that matched that title specifically ‎[6:18 PM]
Jieux: ‎I think I started with the design that has the vote… that would be my guess… ‎[6:18 PM]
Astromon: ‎i think so too Jieux ‎[6:18 PM]
Astromon: ‎1-1 loops help also weakening the space between aptamers ‎[6:19 PM]
LFP6: ‎Also per @JR and @Jieux, it has been noted before that there’s a sweet spot where higher energy leads to better switches, so that it doesn’t get “stuck” in a particular state ‎[6:19 PM]
rhiju: ‎@JR yes, me too, its messy, but other folks have been saying to not pay attention to messiness. ‎[6:19 PM]
rhiju: ‎e.g., in https://getsatisfaction.com/eternagame/topics/player-dev-collaboration-needed-to-create-an-effective-rapid-in-round-select-test-publish-analyze-cycle?topic-reply-list[settings][page]=2#topic-reply-list ‎[6:19 PM]
mgotrik: ‎@LFP6 yeah, i see that there are 4 with that title, but only 1 is case-sensitive ‎[6:20 PM]
LFP6: ‎I’m imagining the “messy” plot is somewhat of an indication that it’ll readily switch since it doesn’t have a particularly high affinity for a given state? ‎[6:20 PM]
rhiju: ‎benbennet1 has been saying repeatedly: "i think this is mainly going into basic geometry. these spirals arent really important. its just the rna flattened and the bonds drawn as arcs. you need to focus on those faint lines in the center left, as ‎[6:20 PM]
rhiju: ‎@jieux did you pay attention to the x-ratios when doing your mods? ‎[6:20 PM]
LFP6: ‎@mgotrik The other ones have additional hashtags ‎[6:20 PM]
Jieux: ‎I don’t even know what an x-ratio is ‎[6:20 PM]
lroppy: ‎;-) ‎[6:20 PM]
rhiju: ‎ah, ok, that’s important . @jieux did you use arcplot tool at all? ‎[6:21 PM]
Jieux: ‎Nope :) ‎[6:21 PM]
JR: ‎X ratio is from arc plot last #. ‎[6:21 PM]
rhiju: ‎@astromon can you possibly run eli’s solution through arcplot? ‎[6:21 PM]
Jieux: ‎I don’t even know where the toolbox is. ‎[6:21 PM]
Astromon: ‎sure ‎[6:21 PM]
Astromon: ‎do we have the sequence ‎[6:22 PM]
mgotrik: ‎@LFP6 yes, that is a possibility. an important thing to look at re: messiness is where the messiness is. if the MG region is messy in absence of ligand, but cleans up in presence that would be a good predictor of switching in my opinion ‎[6:22 PM]
LFP6: ‎Oh also disregard my comment about there being additional tags on the other ones, that’s not the case ‎[6:22 PM]
rhiju: ‎@astromon @LFP6 found https://eternagame.org/game/browse/9035753/?filter1=Id&filter1_arg1=9075529&filter1_arg2=9075529 ‎ ‎[6:22 PM]
whbob: ‎the aptamer folded at it’s end and that pulled the 3’ part of the loop back in state 1. ‎[6:22 PM]
rhiju: ‎@mgotrik can you bring up some of the top solutions from the lightning round 5 tests? ‎[6:23 PM]
LFP6: ‎@mgotrik That certainly sounds logical ‎[6:23 PM]
rhiju: ‎the authors of the top 3 performing solutions (JR, jieux, and astromon) are all here in chat. ‎[6:23 PM]
whyyoubullyme: ‎sup my dudes ‎[6:23 PM]
mgotrik: ‎in the future, it may work better to reference other designs by their ID#'s. up to all of you of course, but would be less text and more instructive ‎[6:23 PM]
whyyoubullyme: ‎F O R T N I T E ‎[6:23 PM]
kingboo: ‎hello ‎[6:23 PM]
LFP6: ‎@mgotrik We do record ancestor IDs by the way ‎[6:24 PM]
whyyoubullyme: ‎sup my hippity hoppity homie ‎[6:24 PM]
Jieux: ‎Noted mgotrik, good idea. ‎[6:24 PM]
Astromon: ‎http://prntscr.com/ls6jgb ‎[6:24 PM]
LFP6: ‎As long as the “copy” tool within Eterna is used ‎[6:24 PM]
LFP6: ‎In the design list ‎[6:24 PM]
Astromon: ‎eli’s arcplot* ‎[6:24 PM]
LFP6: ‎(Not the copy/paste sequence) ‎[6:24 PM]
whyyoubullyme: ‎so on a scale of ten how epic is eterna ‎[6:24 PM]
mgotrik: ‎ah i was not aware of that. what is the best way for me to link several puzzle ID’s for diff puzzles. is it possible in a single link ‎[6:24 PM]
JR: ‎for my posts: first 3 numbers give the parent, secnd 2 numbers give the child ‎[6:25 PM]
LFP6: ‎@wybm We’re in the middle of a chat between scientists, devs, and players right now - please don’t chat right now unless you have something to contribute :) ‎[6:25 PM]
rhiju: ‎thanks @?astromon ‎[6:25 PM]
LFP6: ‎@mgotrik They won’t link between puzzles ‎[6:25 PM]
Astromon: ‎np< ‎[6:25 PM]
LFP6: ‎As I said, you need to use this thing: http://prntscr.com/ls6k5f ‎[6:26 PM]
rhiju: ‎the x-ratio for MGA went up from 10.63x (in Eli’s original) to 16.44x (in Jieux’s mod) ‎[6:26 PM]
Astromon: ‎when thier is time, I have made a picture to show some things I have noticed about the rapid round designs ‎[6:26 PM]
rhiju: ‎@mgotrik how about just post the IDs of those lightning round 5 designs with very high experimental performance ‎[6:27 PM]
rhiju: ‎@astromon this is a good time ‎[6:27 PM]
Jieux: ‎What does it mean for an X-Ratio to go up? ‎[6:27 PM]
rhiju: ‎also want to show everyone the very nice summary that JR put up of one of these designs: https://getsatisfaction.com/eternagame/topics/an-analysis-of-jr-theoassspin-sub00102 ‎[6:27 PM]
Astromon: ‎notice elis percentage is a bit higher than jx’s ‎[6:27 PM]
Astromon: ‎also ‎[6:27 PM]
Astromon: ‎ok tnx ‎[6:27 PM]
Astromon: ‎first here is my design from r5 ‎[6:28 PM]
rhiju: ‎that design’s ID is 88803108 @LFP6 can you find link? ‎[6:28 PM]
Astromon: ‎https://eternagame.org/game/browse/8789144/?filter1=Id&filter1_arg1=9019376&filter1_arg2=9019376 ‎[6:28 PM]
Gerry Smith: ‎X ratio up, most often just linear relationship with fold probablility going down. ‎[6:28 PM]
rhiju: ‎@astromon thanks ‎[6:28 PM]
mgotrik: ‎Two puzzles were particularly impressive from the Theophylline A SS (Spinach): https://eternagame.org/game/browse/8789144/?filter1=Id&filter1_arg1=8803108&filter1_arg2=8803108 and https://eternagame.org/game/browse/8789144/?filter1=Designer ‎[6:29 PM]
JR: ‎I think X ratio up means stronger bonds at the NT points ‎[6:29 PM]
rhiju: ‎on @astromon’s design there are some comments ‎[6:29 PM]
whbob: ‎I’ve heard it said that x50 might be a minimum requirement for a good design ‎[6:29 PM]
Astromon: ‎and here is the pic I made of its arc-plot ‎[6:29 PM]
rhiju: ‎(darn the applet disallows copy/pastof comments) ‎[6:29 PM]
Astromon: ‎http://prntscr.com/ls6l0u ‎[6:29 PM]
Gerry Smith: ‎stronger bonds even with very low fold probabilities? ‎[6:30 PM]
LFP6: ‎Ok so I’m digging into it, and it looks like Jieux’s synthesized solution is actually multiple mods deep ‎[6:30 PM]
LFP6: ‎Looks like you midded your own design some 8 times? :) ‎[6:30 PM]
LFP6: ‎*modded ‎[6:30 PM]
Jieux: ‎Me? ‎[6:30 PM]
DigitalEmbrace: ‎X-ratio is how often a design is predicted to fold in one state versus the other state ‎[6:30 PM]
DigitalEmbrace: ‎Or I should say how often a pair will bond ‎[6:31 PM]
Astromon: ‎Jiuex’s design is an exclusion ‎[6:31 PM]
Astromon: ‎so it has higher off percentage ‎[6:31 PM]
LFP6: ‎@Jieux Yes you! ‎[6:31 PM]
Astromon: ‎than same state winners ‎[6:31 PM]
JR: ‎ I like X ratios from arc plot balanced ‎[6:31 PM]
Astromon: ‎they all have low percentages ‎[6:31 PM]
Jieux: ‎I sometimes mod my mods… in recent labs… especially if the original design does not work in all 3 engines… first mod gets it into all 3 engines then subsequent mods are variations on that theme that also work in all three engines. ‎[6:31 PM]
rhiju: ‎@JR cool: what does balanced X ratios mean? ‎[6:32 PM]
LFP6: ‎http://prntscr.com/ls6m2p ‎[6:32 PM]
Jieux: ‎so say rhiju makes a design that only works in V2… I will try to make it work in #all3 then if I can, will then make variations of that mod moving forward… ‎[6:32 PM]
LFP6: ‎The original design doesn’t have an ancestor recorded, so you must’ve manually copy-pasted the solution ‎[6:32 PM]
JR: ‎A stronger NT pair at that nt point gives a higher X factor usually. ‎[6:32 PM]
JR: ‎there seems to be a sweet spot. ‎[6:33 PM]
Jieux: ‎Not sure what that means LFP6 ‎[6:33 PM]
rhiju: ‎@LFP6 before I forget, will you be able to record chatlog and post? I know you have to leave early ‎[6:33 PM]
LFP6: ‎@Jieux: You took a design by eli, copied the sequence, then submitted mods 8 times in succession ‎[6:34 PM]
rhiju: ‎@jieux – folding in all folding engines is also a constraint that JR used https://getsatisfaction.com/eternagame/topics/an-analysis-of-jr-theoassspin-sub00102 ‎[6:34 PM]
LFP6: ‎So it’s not a “direct” mod of the original ‎[6:34 PM]
LFP6: ‎@rhiju Yup ‎[6:34 PM]
Astromon: ‎most all the winners have a similar xratio between aptamers ‎[6:34 PM]
Astromon: ‎and the same state ones all have low on percentages ‎[6:35 PM]
rhiju: ‎@JR i want to understand more abuot te sweet spot. what number(s) in particular are you looking at, and what’s the sweet spot? ‎[6:35 PM]
Jieux: ‎I am still unsure whether having a design fold in the original Vienna engine is an asset or a liability… ‎[6:35 PM]
rhiju: ‎@astromon oh I see – “similar xratio between aptamers” is a kind of balancing – @jr is that also what you were referring to? ‎[6:36 PM]
Jieux: ‎I’m thinking if it’s outmoded it might be a liability. ‎[6:36 PM]
rhiju: ‎and @astromon thanks, I understood low on percentages, thanks to your graphic ‎[6:36 PM]
Astromon: ‎yes and its something i have noticed studying the winners ‎[6:36 PM]
rhiju: ‎that makes good sense ‎[6:36 PM]
JR: ‎IIf you get at weak design first and increase the nt strength at the arc post points fot get high and higher x factors. ‎[6:36 PM]
Astromon: ‎Eli brought it to my attention ‎[6:36 PM]
whbob: ‎the end of the loop doesn’t have the ag,uc like Omei mentioned. wonder what it would score with them? ‎[6:37 PM]
rhiju: ‎@astromon thanks, i can also bug @eli for more info off-line ‎[6:37 PM]
JR: ‎Too strong and the x factor starts decreasing ‎[6:37 PM]
Astromon: ‎2 static stems also seem to be a good thing as i showed in my picture i posted ‎[6:37 PM]
DigitalEmbrace: ‎The x-ratios of the aptamers should be approximately the same number, so both 40x for example. ‎[6:37 PM]
rhiju: ‎@digitalembrace thanks ‎[6:37 PM]
JR: ‎so you have to find a trade off. ‎[6:37 PM]
rhiju: ‎that is a feature that is totally new to me ‎[6:38 PM]
Gerry Smith: ‎sometimes you can get very low State 2 %'s and really high X’s…I’ve assumed that wasn’t good ‎[6:38 PM]
mgotrik: ‎@Jieux each of the engines have their strengths/weaknesses. One of the members tuning in has done a lot of analysis regarding the goodness of each engine. Perhaps she could enlighten everybody with a graphic sometime ‎[6:38 PM]
Astromon: ‎Digital Embrace made a great doc explaining some similarities between rapid winners ‎[6:38 PM]
rhiju: ‎i’m hearing that balanced x-factors for both aptamers is important, now from several folks ‎[6:38 PM]
rhiju: ‎here’s a proposal for a retrospective analysis of our data so far ‎[6:39 PM]
JR: ‎I think of the x factor as the probability to finding one another ‎[6:39 PM]
DigitalEmbrace: ‎But only for same state puzzles. ‎[6:39 PM]
rhiju: ‎we calculate x-factors for all designs, as well as separate percentages; and we also track balance in x-factors for the two aptamers ‎[6:39 PM]
Jieux: ‎When will we have “result results” from all the designs using the more accurate method? ‎[6:39 PM]
Gerry Smith: ‎how about a complete break in the aptamer? ‎[6:40 PM]
rhiju: ‎we need to do this in all engines, which is feasible ‎[6:40 PM]
mgotrik: ‎are the x-factors just the probability of specifically the closing pairs being paired? ‎[6:40 PM]
LFP6: ‎(Sorry for the lag, but the URL for JR’s design: https://eternagame.org/game/browse/8789144/?filter1=Id&filter1_arg1=8803108&filter1_arg2=8803108) ‎[6:40 PM]
JR: ‎If both are balanced it might make sense that the aptamer would form ‎[6:40 PM]
rhiju: ‎we can do this for all the lightning rounds as well as prior rounds like FMN/MS2, trp/MS2 ‎[6:40 PM]
rhiju: ‎then we look for enrichments/correlations of fold-ratios with features that you all listed above ‎[6:41 PM]
rhiju: ‎its very interesting that so far there has been little discussion of specific sequence features ‎[6:41 PM]
rhiju: ‎or structural features ‎[6:41 PM]
DigitalEmbrace: ‎@mgotrik Yes, and the arcplot tool focuses on the pairs at the closing ends of the aptamers. ‎[6:41 PM]
rhiju: ‎@JR yes, I got it now on balanced. finally! ‎[6:41 PM]
rhiju: ‎@JR you also pointed out a few other atrributes that you look for ‎[6:41 PM]
mgotrik: ‎@Jieux undetermined when we will have the “result results.” However, the recent data we are giving back is very accurate, just not comprehensive considering the few sequences tested ‎[6:42 PM]
JR: ‎distance is one ‎[6:42 PM]
rhiju: ‎yes i was about to paste that in “6.) The distance NTs have to travel to its associated pair in the next state is acceptably minimized.” ‎[6:42 PM]
Astromon: ‎structure feature i like are 1-1 loops between aptamers ‎[6:42 PM]
rhiju: ‎ah, yes, astromon 1-1 loops ‎[6:42 PM]
Jieux: ‎@ mgotrik: Understood. ‎[6:42 PM]
kingboo: ‎was there anything wrong with the last results? ‎[6:42 PM]
rhiju: ‎@mgotrik was also pointing out an intuition that 1-1 loops (‘mismatches’) are mportant featuers of good switches ‎[6:42 PM]
JR: ‎Nts that have long distances to travel to find it’s pair are suspect ‎[6:43 PM]
mgotrik: ‎@kingboo no nothing was wrong. the results are viewable here: https://imgur.com/a/yK4193u we will write up a summary and post it soon ‎[6:43 PM]
Jieux: ‎I think Eli’s original design had a 1-1 loop that I collapsed with a UG pair mod… ‎[6:43 PM]
rhiju: ‎@JR, just so I understand, the distance is between the pairing partner in one state and the pairing partner in the state 2? ‎[6:43 PM]
rhiju: ‎for each switching nucleotide? ‎[6:43 PM]
JR: ‎Yes ‎[6:43 PM]
whbob: ‎if there are 2 static stems, only the moving parts of the aptamer & loop will have % ‎[6:43 PM]
Jieux: ‎thus it would be good to see results from Eli’s design. ‎[6:43 PM]
kingboo: ‎was there anything that could have gone better? ‎[6:44 PM]
mgotrik: ‎@kingboo well of the 11 puzzles we tested (for two puzzles), only 1 was a ‘success’ in that exhibited switching. ‎[6:44 PM]
JR: ‎I like static end stems to decease moving parts and increase probabilities of other nts connecting if they do their job properly. ‎[6:45 PM]
kingboo: ‎just curious. sorry. ‎[6:45 PM]
mgotrik: ‎I would have loved to see a success for the second puzzle: ATP Same State (MG) ‎[6:45 PM]
Jieux: ‎make a 2nd lightning round just for that lab perhaps… ‎[6:46 PM]
Gerry Smith: ‎is there a quantative way to measure NT distance traveled? ‎[6:46 PM]
Astromon: ‎none of my desgns made it in (: ‎[6:46 PM]
Jieux: ‎Interesting quetion Gerry. ‎[6:46 PM]
JR: ‎I think Eli’s puzzles are on the right track. ‎[6:46 PM]
whbob: ‎it would be nice to see a levenshtein distance list to get design ID’s of similar designs ‎[6:46 PM]
whbob: ‎then later compare scores ‎[6:47 PM]
Astromon: ‎eli has a new hastag #+5 ‎[6:47 PM]
LFP6: ‎@whbob is it viable to make a network graph via levethian distances? ‎[6:47 PM]
Jieux: ‎Just googled levenshtein distance ‎[6:47 PM]
rhiju: ‎@JR on static stems, @astromon also pointed that out in his graphic and we’ve assumed for a long time that static stems are good ways to sequester nt’s. So that’s good so see its holding up ‎[6:47 PM]
Astromon: ‎means 5 pairs between aptamers ‎[6:47 PM]
mgotrik: ‎@whbob thats a good idea. I will look into something like that ‎[6:47 PM]
LFP6: ‎IE, so that you could trace further deviation from a node ‎[6:48 PM]
LFP6: ‎Not completely sure how that’d work, but would align with a project I had thrown together recently ‎[6:48 PM]
rhiju: ‎I wonder if there’s a way to just use jandersonlee’s arcplot tool ‘in batch’ to get all the numbers we need for percentages ‎[6:48 PM]
JR: ‎ In that lab though, the size os the spinich aptamer forces that constraint. ‎[6:48 PM]
rhiju: ‎i’ll ask omei about that offline ‎[6:48 PM]
Gerry Smith: ‎I had that leven program in R but it crashed. Trying to recreate using contract labor now ‎[6:48 PM]
Jieux: ‎One thing I noticed in my Comedy design is that there is a likelyhood of one of the static stems mismatching with the green glowy area… If I were to make another mod I would futz with this area and take this away. ‎[6:49 PM]
whbob: ‎@lfp6: yes, seeing how score relates to distance would be interesting. ‎[6:49 PM]
kingboo: ‎brb gotta eat dinner. ‎[6:49 PM]
whbob: ‎gerry’s good at that:) ‎[6:49 PM]
Jieux: ‎the static stem area… you can’t futz with the green glowy area :) ‎[6:49 PM]
Gerry Smith: ‎was… ‎[6:50 PM]
whbob: ‎:) ‎[6:50 PM]
rhiju: ‎@JR " In that lab though, the size os the spinich aptamer forces that constraint. ‎", which lab and constraint are you referring to? I want to understand more clearly ‎[6:50 PM]
Zama: ‎@mgotrik, I think Eli predicted only one of the ATP’s would be successful in this post. https://mail.yahoo.com/d/search/keyword=atp%20prediction?reason=invalid_cred ‎[6:50 PM]
LFP6: ‎@Gerry @whbob If you could help me figure out the algorithm, to generate the edges for the graph, there’s a high likelyhood that you could have a network graph in a relatively short amount of time :) ‎[6:50 PM]
mgotrik: ‎re: static stems. We think of them mostly as a way for you all to “reduce” the size of the puzzle region. e.g. if you can design a switch using 30/50 of the available bases, you can lock 20 of them into a static stem ‎[6:51 PM]
JR: ‎It would b e the puzzle that I posted the forum comment on. ‎[6:51 PM]
rhiju: ‎@Zama can you post the link into the game or getsat or some public spot (or copy/paste in the text of the prediction)? ‎[6:51 PM]
Astromon: ‎the one tested was one he predicted to do well ‎[6:51 PM]
rhiju: ‎@JR thanks ‎[6:51 PM]
Astromon: ‎i think we just didnt get the right designs in. ‎[6:51 PM]
rhiju: ‎@JR which constraint were you referring to? ‎[6:51 PM]
Astromon: ‎APTMG ss ‎[6:52 PM]
Gerry Smith: ‎working on a cluster analysis capability using leven dist of notational structures…programmer thinks next week. ‎[6:52 PM]
DigitalEmbrace: ‎I have a question: Were the lightning round switches intended to be reversible? And do you test for reversibility? ‎[6:52 PM]
Gerry Smith: ‎Leven module could be used for this too ‎[6:52 PM]
mgotrik: ‎@zama do i have to have a yahoo account to read that post? ‎[6:52 PM]
Astromon: ‎sry i meant Apt A ss mga ‎[6:52 PM]
Zama: ‎whoops- I thought I had done that sorry- https://getsatisfaction.com/eternagame/topics/player-dev-collaboration-needed-to-create-an-effective-rapid-in-round-select-test-publish-analyze-cycle?utm_source=notification&utm_medium=email&utm_campaig ‎[6:53 PM]
rhiju: ‎@astromon @Zama anyway you can paste the text of eli’s prediction here? ‎[6:53 PM]
JR: ‎The size of the spinach aptamer doesn’t leave much room for maneuvering ‎[6:53 PM]
rhiju: ‎oh thanks! ‎[6:53 PM]
JR: ‎Which could actually be a good thing. ‎[6:53 PM]
Gerry Smith: ‎if leven module works, I will forward it to you @lfp6 ‎[6:53 PM]
Astromon: ‎http://prntscr.com/ls6si5 ‎[6:54 PM]
Astromon: ‎eli’s vs predictions ‎[6:54 PM]
LFP6: ‎@gerry THanks ‎[6:54 PM]
JR: ‎The spinich aptamer doesn’t leave a lot of free NTs to play around with as compared with the other lab aptamers ‎[6:54 PM]
mgotrik: ‎@JR if you want more/less puzzle area on future puzzles let me know. I have just been sticking with what has been usual, but theres no reason we should have to ‎[6:54 PM]
Astromon: ‎all his predictions came true for the lightup round btw ‎[6:55 PM]
JR: ‎If the spinach aptamer does graet in this round, it brings up the question: ‎[6:55 PM]
Astromon: ‎although he wants us all to give all the puzzles our attention and care ‎[6:55 PM]
JR: ‎If we squeese another different aptamer in the same space, do we get the same results? ‎[6:56 PM]
rhiju: ‎@astromon yes i’m trying to read the posts on earlier predictions now, and on ATP more recently ‎[6:57 PM]
Brourd: ‎Hello @everyone ‎[6:58 PM]
tone: ‎Hi all. Thanks for the fireside chat ‎[6:58 PM]
whbob: ‎Jieux’s design almost had two static stems. 31-42 didn’t quite stick. ‎[6:58 PM]
Jieux: ‎Hello @Brourd ;) ‎[6:58 PM]
Astromon: ‎heya tone ‎[6:58 PM]
Astromon: ‎heya brourd ‎[6:58 PM]
mgotrik: ‎@JR that would be ideal, but it probably is not the case. most of the good puzzles have puzzle regions that pair with the aptamers, and you would need to maintain that same pairing (or equally strong pairing) with any new aptamer you swap ‎[6:58 PM]
whbob: ‎hi tone,brourd ‎[6:58 PM]
JR: ‎I agree ‎[6:59 PM]
AndrewKae: ‎@mgotrik I would have liked to see some extra NTs after the ATP on the ATP B Exclusion (Spinach), very limited possibilites for the spinach to come after the ATP ‎[6:59 PM]
Jieux: ‎FiskerMod 8 does have 2 static stems… ‎[6:59 PM]
rhiju: ‎@astromon @zama thanks for pointing us to eli’s predictions ‎[6:59 PM]
rhiju: ‎we need to understand those better! ‎[7:00 PM]
rhiju: ‎hey everyone, have to go! thanks for coming and contributing. things make more sense to me at least! ‎[7:00 PM]
AndrewKae: ‎bye rhiju ‎[7:00 PM]
JR: ‎see ya ‎[7:00 PM]
Astromon: ‎Ur welcome ‎[7:00 PM]
lroppy: ‎adios ‎[7:00 PM]
Brourd: ‎Take care ‎[7:00 PM]
whbob: ‎to all a good night:) ‎[7:00 PM]
kingboo: ‎goodbye ‎[7:00 PM]
Jieux: ‎Thank you for inviting me to participate, not sure I offered any help, but glad my designs performed pretty ok. ‎[7:01 PM]
Gerry Smith: ‎super helpful discussion - thank you ‎[7:01 PM]
Astromon: ‎read though’s getstats comments and you will get it! ‎[7:01 PM]
Astromon: ‎thanks for the chat lets do it agin somtime ‎[7:01 PM]
Jieux: ‎I thought we were going to be discussing something else entirely :) ‎[7:01 PM]
whbob: ‎afk ‎[7:01 PM]
mgotrik: ‎yes thanks everybody, you’ve all been very helpful. if you have any insights or thoughts that you think we might want to know, feel free to message me ‎[7:02 PM]