6-12-18 1022PM PST
LFP6: Jen, I need a true crash course on what you're doing. I get lost tbh! [9:03 PM]
LFP6: Youre saying many cool things, but Im missing the core info [9:03 PM]
Jennifer Pearl: i can do that now if you want [9:04 PM]
LFP6: If you could give me at least a short description that would be cool [9:04 PM]
Astromon: jen is a genius all this stuf shes doing has proven test results!! i scored a 90 relieing on her programs adviuce and she has a few 100 percents using it! [9:04 PM]
Jennifer Pearl: what I am doing for NUPACK now or have done historically? [9:05 PM]
Astromon: just saying! [9:05 PM]
Astromon: both! [9:05 PM]
LFP6: Historically [9:05 PM]
LFP6: I have 0 context [9:05 PM]
Jennifer Pearl: ok [9:06 PM]
Astromon: having saying that lfp6 has a great idea making the site compatable with mobile devices way cooL stuf guys great work!! [9:06 PM]
Jennifer Pearl: so DPAT launches Vienna 2 using a sequecne from you teh file gereated from your script [9:08 PM]
Astromon: this is almost like a dev chat but more relaxed :) [9:07 PM]
LFP6: hehe [9:07 PM]
LFP6: Might want to post it, actually [9:07 PM]
Jennifer Pearl: the MFE secondary structure adn the pairing probabilities are generated in a file [9:08 PM]
LFP6: Help me out, whats MFE? [9:08 PM]
LFP6: Ive been under a rock [9:08 PM]
Astromon: thats can be a new think interesting chat .. and some one can post it if they deem it interesting at the wiki haha [9:08 PM]
Jennifer Pearl: Minimum Free Energy [9:09 PM]
LFP6: ooh [9:08 PM]
LFP6: So, free energy like we'd see? [9:09 PM]
Jennifer Pearl: the state teh RNA wnats to take most often out of the ensemble [9:10 PM]
LFP6: Is that what we see? [9:09 PM]
Jennifer Pearl: the secondary strucutre we use in game is the MFE secondary structure [9:10 PM]
LFP6: Kk [9:09 PM]
LFP6: I follow [9:09 PM]
LFP6: I knew this, but not the term [9:10 PM]
Jennifer Pearl: there are more possible secondary structures but the RNA will keep that shape most often since it is the lowest energy [9:11 PM]
LFP6: Yep [9:10 PM]
Astromon: yes [9:11 PM]
LFP6: Sounds like chemistry probably because it is [9:11 PM]
Astromon: hmm thats what ive noticed latley [9:11 PM]
Jennifer Pearl: this secondary strucutre is recorded by DPAT and it cross references the stacks with the dot plot [9:12 PM]
Jennifer Pearl: and records the minimum, avg or maximum pairing probabilites for each stack reproted in teh dot plot which is found in the files generated [9:13 PM]
Jennifer Pearl: molecular biology [9:13 PM]
Jennifer Pearl: yay!!! [9:13 PM]
Astromon: http://prntscr.com/bfp4ty [9:12 PM]
Jennifer Pearl: and chemistry [9:13 PM]
Jennifer Pearl: yay [9:13 PM]
Astromon: jen maybe you can help my question and I think it has to do what your talking about [9:13 PM]
Jennifer Pearl: it does this for every desing in a sublab [9:14 PM]
LFP6: How does it comeup with the plot? [9:14 PM]
LFP6: Generated on the fly? [9:14 PM]
Jennifer Pearl: creating a database of tsecondary structures and their pairing probabilites along with values from the partition function [9:16 PM]
Jennifer Pearl: the doit plot is generated on teh fly by Vienna 2 [9:16 PM]
Jennifer Pearl: in a file I just pull the pairing probabiites out of the file. It is not a real dot plot generated but a .ps file witht eh probabilities embeded in it [9:16 PM]
Jennifer Pearl: what do you need Astro [9:17 PM]
Astromon: sry i didnt mean to interupt i can ask another time [9:16 PM]
LFP6: Do you have sample data you could show me generated at this step? [9:17 PM]
LFP6: Where you generate probabilities [9:17 PM]
Astromon: id rather listen again tyour programs as you tel lfp6 << :) [9:17 PM]
Jennifer Pearl: just a sec [9:18 PM]
Astromon: i know a bit about all this working w/ you in the labs..im sure it would be harder for somone to understand not having that opertunity<> [9:19 PM]
LFP6: I havent been working with the labs much at all [9:19 PM]
Astromon: i need a link to the second part of your presentaionion though << :) [9:20 PM]
LFP6: If you go to the eternadev youtube it should be there [9:20 PM]
Astromon: well your script is what enabled jen to use her program thats pretty cooL [9:20 PM]
Astromon: oH okaY [9:20 PM]
LFP6: I just blindly provide the data lol [9:21 PM]
Astromon: ill ck it out later n:D [9:21 PM]
Astromon: writing a scipt that dumps all that data is way cool and wasnt done blindly i woldnt imagine [9:22 PM]
LFP6: Its kinda like "I pick stuff up and put zem down" [9:23 PM]
Jennifer Pearl: https://dl.dropboxusercontent.com/u/87351147/1stState_JRSenAMS2ONSub001.ps [9:24 PM]
Astromon: haha [9:23 PM]
Jennifer Pearl: that is the files for two states combined [9:25 PM]
Jennifer Pearl: scroll down to where it says ubox and lbox [9:25 PM]
Jennifer Pearl: ubox is the upper triangle and lbox is the MFE [9:25 PM]
Astromon: i cant open that file [9:25 PM]
LFP6: Geez thats a lot of data [9:25 PM]
LFP6: @Astro Open with notepad or whatnot [9:25 PM]
Astromon: jen is the cookie monster and Data is the cookies [9:26 PM]
LFP6: TextEdit if your a Mac guy [9:26 PM]
LFP6: @Jen I suppose I have a hard time making meaning out of this [9:26 PM]
Jennifer Pearl: just asecond [9:28 PM]
Astromon: oH that worked ty [9:27 PM]
LFP6: np [9:27 PM]
Astromon: oh yeah it was w/ the A/B puzzles [9:28 PM]
Astromon: that was so fun [9:28 PM]
Astromon: i can not wait till we have data for nupaK!! [9:28 PM]
LFP6: ie what do you mean by cross-referenced sec struct w/ dot plot, pairing probabilities per stack, patrition fn, etc [9:29 PM]
Astromon: yeah dpat looks at the data and sees paterns in bad and good designs [9:30 PM]
Astromon: but she can explain your question better than I<< :) [9:30 PM]
Astromon: i hope there are bubbles in the new gui [9:36 PM]
Astromon: maybe even stars like a planatarium [9:36 PM]
Astromon: when somone solves a puzzle a super nova happens! [9:37 PM]
Jennifer Pearl: im back [9:49 PM]
Astromon: wb jen [9:48 PM]
Jennifer Pearl: so i will explain the file a little bit more which will help with the cross referencing thing [9:49 PM]
Jennifer Pearl: basically ingore everythign but the ubox and lbox parts [9:49 PM]
Jennifer Pearl: http://prntscr.com/bfpd2j [9:50 PM]
Jennifer Pearl: http://prntscr.com/bfpdao [9:52 PM]
Astromon: 6:28 [9:53 PM]
Jennifer Pearl: i circled those because i forgot [9:54 PM]
Jennifer Pearl: it made snese when I did it [9:54 PM]
Astromon: oH [9:54 PM]
Jennifer Pearl: so it was more to be able ot point out something about hte file [9:55 PM]
Jennifer Pearl: 6and 28 are teh nucleotid numbers [9:56 PM]
Jennifer Pearl: 6 is hte lower number in th pair and 28 is the higher number [9:56 PM]
Jennifer Pearl: those make up the base pair and ht epairing probability is given to the right. [9:56 PM]
Astromon: made a0.95000 [9:56 PM]
Jennifer Pearl: it is given in percentages in decimal form [9:57 PM]
Jennifer Pearl: that is becasue it is the classic dot plot you seefor Vienna 2 and teh lower triangle or "lbox" is always a dark spot [9:57 PM]
Astromon: so you have all this data in a sublab [9:56 PM]
Astromon: do you study it and find paterns [9:57 PM]
Astromon: or does a program do that [9:57 PM]
Jennifer Pearl: i look at plots of stacks and have a program do alot of it now [9:58 PM]
Jennifer Pearl: based on my orignial analysis [9:58 PM]
Astromon: way cooL [9:58 PM]
Jennifer Pearl: so I take teh base pairs in th lbox which is part of the MFE secondary structure and get the pairing probabiity from the ubox [9:59 PM]
Jennifer Pearl: that is the cross referencing [9:59 PM]
Astromon: i saw your first pes it was great [9:58 PM]
Astromon: it explained alot to me [9:58 PM]
Astromon: im looking forward to seeing the second part<< [9:58 PM]
Jennifer Pearl: or recording the pairing probabilities for teh stacks in the MFE secondary structure [10:00 PM]
Astromon: what was it that saw we needed a long stack in the one sublabs [9:59 PM]
Astromon: the non switching one [10:00 PM]
Astromon: i got a 90 result from it [10:00 PM]
LFP6: Whats lbox and ubox? [10:01 PM]
Astromon: i mean the non switching area the puzzle did indeed switch [10:01 PM]
LFP6: Oh [10:01 PM]
Astromon: its like the xbox [10:01 PM]
Jennifer Pearl: lbox is the lower box of a clasic vienna 2 dot plot [10:02 PM]
LFP6: Its the pairing probability desired vs predicted? [10:02 PM]
Jennifer Pearl: lower triangle [10:03 PM]
Astromon: haha just joshing [10:02 PM]
Jennifer Pearl: pairing probailtiy is predicted [10:03 PM]
Astromon: maybe its L box [10:03 PM]
Jennifer Pearl: ubox is teh opper triangle [10:04 PM]
Astromon: and upper box [10:03 PM]
LFP6: So you have target structure and pairing probabilities. Now what? [10:03 PM]
Jennifer Pearl: i put those into a database for all the base pairs [10:04 PM]
Astromon: Upper and lower :) [10:03 PM]
Jennifer Pearl: there is a database in the program and a database genreated in afile [10:05 PM]
LFP6: Ok, then after that what do you do with them? [10:04 PM]
Jennifer Pearl: a file [10:05 PM]
Jennifer Pearl: I do analysis on it using my macro language [10:06 PM]
LFP6: An example of what you might do? [10:05 PM]
Jennifer Pearl: i have a family of commands that search through adn find stacks that only exist in a specific score range [10:06 PM]
Astromon: :) [10:06 PM]
LFP6: So where does target structure come into play? [10:06 PM]
Astromon: wOOt [10:06 PM]
Jennifer Pearl: target structure comes into play as the thing that is being searched for a pattern [10:08 PM]
Jennifer Pearl: in teh macro language [10:08 PM]
Jennifer Pearl: it aslo gives guidance for what pairing probabilites to look at [10:09 PM]
Jennifer Pearl: that is somehting I am moving from in NUPACK. I am doing everything from the upper triangle only [10:09 PM]
LFP6: Ok, I think I sorta follow [10:09 PM]
LFP6: What resouces have you put out? [10:09 PM]
Jennifer Pearl: say that you hade 4 desings [10:10 PM]
LFP6: I know the videos [10:09 PM]
Jennifer Pearl: i have forum posts that is it [10:10 PM]
Jennifer Pearl: adn the videos [10:10 PM]
LFP6: Wasnt there a paper or something? [10:10 PM]
Jennifer Pearl: yes [10:11 PM]
Astromon: thats next<> [10:10 PM]
Jennifer Pearl: on picking out ideal lengths of binding sites with DPAT [10:11 PM]
LFP6: Ah [10:10 PM]
LFP6: Cool [10:10 PM]
Astromon: oH yeah i meant a colplete one [10:11 PM]
LFP6: Ill need to dive into this more [10:11 PM]
Astromon: complete* [10:11 PM]
Jennifer Pearl: Sara takes those patterns and the stacks found and designs around that [10:12 PM]
Astromon: yeaH its way interesting and roven to work! [10:11 PM]
Jennifer Pearl: picking only desings that have those stacks [10:12 PM]
Astromon: proven** [10:11 PM]
Jennifer Pearl: that you specify [10:12 PM]
Astromon: it seems following this data even makes bad designs score well [10:12 PM]
Astromon: looking at my work w/ this it is true<< [10:13 PM]
Astromon: taking a good design and tweeking it in this manner could make it very strong [10:14 PM]
Jennifer Pearl: im glad to hear it has helped [10:15 PM]
Jennifer Pearl: does that give a general explanation LFP6 [10:15 PM]
Astromon: well you know the rtesults we got its great! [10:15 PM]
Astromon: ill look foword to pionters for nupak puzzles<>> :D [10:15 PM]
LFP6: Yes, I think so [10:16 PM]
LFP6: All right, I need to go to sleep. Thanks for the interesting info and discussion both of you! [10:16 PM]
Astromon: thanks you too lfp6 rest well! [10:16 PM]
Jennifer Pearl: where do you think we should post this [10:18 PM]
Jennifer Pearl: good night [10:18 PM]
LFP6: Hm... Astro, you have back to when we started talking too [10:17 PM]
Astromon: in a new area at the terna wike called interesting chat [10:17 PM]
LFP6: Id like that as well [10:17 PM]
LFP6: I like the idea :) [10:17 PM]
LFP6: Thanks, later all [10:18 PM]
