Zama showed me an example yesterday. Here it comes.
zama: @eli, the counterclockwise design you have in your new Google Doc- looks clockwise to me?
zama: this looks counterclockwise to me
eli: The image you show above is clockwise. I count from the outside start of the circle.
Google doc with examples of different spiral types.
Ignore what I said about Vienna2, I can’t remember where I got that idea. And I just found the dictionary on the Wiki. Doesn’t cover newer terms but some good info for me to review. You guys have put so much time into creating tutorials and guides for new players, really impressive!
Player recommendations for the current lightning round (round 5)
@jandersonlee has added batch processing capabilities to his arcplot server, so it has now become realistic to process and evaluate large batches of designs. I’ve talked it over with some players, and we concur that it is time we start publishing some results to help collaboratively improve players’ ability to evaluate designs.
The process will continue to evolve, but for this round (voting closing Wednesday PM), I am inviting all players (whether or not you are using arcplot) to post your recommendations for the two labs here. There are no strict rules for this, but I’ll suggest the following guidelines:
Omei, These are the five i like for Trytophan. All designs have a xratio of 40-60x…Both r4 winners had an x ratio in this range. Thanks
https://eternagame.org/game/browse/8787266/?filter1=Id&filter1_arg1=8811363&filter1_arg2=881…
https://eternagame.org/game/browse/8787266/?filter1=Id&filter1_arg1=8967413&filter1_arg2=896…
https://eternagame.org/game/browse/8787266/?filter1=Id&filter1_arg1=8978825&filter1_arg2=897…
https://eternagame.org/game/browse/8787266/?filter1=Id&filter1_arg1=8981750&filter1_arg2=898…
https://eternagame.org/game/browse/8787266/?filter1=Id&filter1_arg1=8986304&filter1_arg2=898…
i think you are over complicating this. the main goal should be to have as little faint bonds as possible. none of this spiral stuff looks like it would help anything it seems. the goal is to get the arc plot to match the puzzle with as little interference as possible
i think you are focusing on the wrong issue. this has nothing to do with spirals. its got to do with matching probabilites rather than making a nice spiral. the arc plot is just a tool to prune potential interference to work in the wet lab
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 there is a small chance to mismatch there
those arc plots are messy with faint lines. not good.
Here are some Theophylline A Same State https://eternagame.org/game/browse/8787270/ designs that interest me this round:
seq,minAoff,minAon,minAx,maxAx,minRoff,minRon,minRx,maxRx,mfe1,mfe2,dFE,bonus,same,pp1,pp2,d1,d2
UAAGGUGAGGCGCCCAGCCCGACGGUUGAUACCAGUAGCGCAGCCUACCCUUGGCAGCAGUCCGGUAACGAAUGGCUGGGCGCCC,0.007,0.785,30.823,112.792,0.007,0.777,1.012,106.869,-37.500,-38.900,1.40,-4,1,0.992726,0.777362,0.488748,0.422383
UUAUUAUUAAGACAGUCUGAGCCGACGGAUACCAGAGACAAUAGUCUCCCUUGGCAGCUGGUAACGAAUGCUCAGGCUGACAAUU,0.017,0.903,1.815,52.704,0.017,0.876,1.000,52.539,-28.400,-29.800,1.40,-4,1,0.983335,0.875562,0.478785,0.436298
GUCAGUAGAAAGACUAACUGGCCGACAGAUACCAGCCAGUUCGCUGGCCCUUGGCAGCUGGUAACGAAUGCCAGGUCCUACUGAC,0.003,0.643,33.213,194.814,0.003,0.637,2.922,193.952,-33.000,-34.000,1.00,-4,1,0.996716,0.636938,0.470543,0.378125
CGCUGGUCGGACAGGUAACGAAUGACAGAUACCAGCUACGAAAGUAGCCCUUGGCAGCUGUCCGACUGUAAUAAUAGCGAAUUAU,0.005,0.716,1.012,155.224,0.004,0.588,154.439,154.504,-27.600,-28.700,1.10,-4,1,0.996193,0.587949,0.484057,0.356055
UUAAUUAUUAGUAGGACCUCACCGACAGAUACCAGCCAGUUGUCUGGCCCUUGGCAGCUGGUAACGAAUGUGAGGUCCUACAAUU,0.006,0.764,71.682,129.152,0.006,0.757,1.009,128.564,-37.100,-38.100,1.00,-4,1,0.994111,0.757111,0.472230,0.399191
CGCCAGUAACGGUUAAAGGCACCGACUGAUACCAGUGUCGAUAGACACCCUUGGCAGCAGGUAACGAAUGUGCCAAUUACUGGCG,0.010,0.847,1.062,81.533,0.010,0.597,52.088,81.203,-31.800,-32.200,0.40,-4,1,0.989887,0.597087,0.490786,0.358677
CUGAGCAUAAUCAGCAGCUCACCGACUGAUACCAGCCAGGCCGCUGGCCCUUGGCAGCGGGUAACGAAUGUGAGCUGCUGCUCAG,0.006,0.747,20.410,117.373,0.006,0.680,1.005,115.914,-39.000,-40.000,1.00,-4,1,0.994133,0.680065,0.491121,0.405857
CUGAGCAGCAUCAAUAGCUCACCGACUGAUACCAGCCAGGCCGCUGGCCCUUGGCAGCGGGUAACGAAUGUGAGCUGCUGCUCAG,0.006,0.747,19.682,117.471,0.006,0.680,1.006,116.006,-39.000,-40.000,1.00,-4,1,0.994139,0.679913,0.492882,0.407541
AAGUGGCUAACCGGUAACGAAUGUAACGAUACCAGUUCUGAAUUGGGCCCUUGGCAGCGUUACCGACGGUUGACUUAAGCCACUU,0.004,0.617,107.908,165.667,0.006,0.593,0.771,106.290,-29.000,-29.100,0.10,-4,1,0.994199,0.592568,0.490879,0.415939
All have static necks, and good metrics and the MFE in state 2 has both the reporter and aptamer sites. Some are spirals, some are not. Past labs (R101, R107) indicate that designs can perform well with these metrics:
like: maxAx>20,maxRx>50,minAon>50%,minRon>50%,d1<0.5,d2<0.5,dFE<1.5
I’ve also made some mods to some of these to fill in my last few slots.
I find there are more submissions in each puzzle than can reasonably be looked at so my approach tends to be a quick eyeballing.
I usually just look at all the submissions for a particular individual as this makes things more tractable: and will mostly use Vienna2 and the old browser continuously pressing d to flip through the structures. I keep meaning to look at the arc plot tool but have yet to get around to it.
What I look for is something I think of as cleanliness of design (for both states). It’s a little hard to define but it tends to mean a pleasing-looking geometric structure with an absence of too many kinks, small bulges and other oddities. Also the two structures must look sufficiently different that there’s a reasonable chance of good switching. Other considerations would be a reasonable balance in the nucleotide composition (i.e. more or less a quarter each of AUCG) and retaining the general structure when switching to Vienna or Nupack.
I try and only submit solution per designer and don’t look at the number of votes already cast (this might be a mistake as there’s a tendency to end up with a lot of solutions with only one vote)
Yes, to harness diversity of informed opinions, voting twice makes sense (a little more work). The first time to get a list of designs with at least one vote. And the second time (later), screening for designs with 1 or more votes and then changing your votes to those within that group. Your approach sounds thoughtful and different enough so it could add even more value in re-voting a second time…
For Tryptophan A Same State, these seem interesting to me:
seq, minAoff, minAon, minAx, maxAx, minRoff, minRon, minRx, maxRx, mfe1, mfe2, dFE, bonus, same, pp1, pp2, d1, d2
GCCUGAGAACGGUCCAAAUAACUCCUCCGACGAGGACCGGUACGGCCGCCACUCGGUAACGAAUGAGGAGUGGCGGAUCUCAGGC, 0.004, 0.703, 146.611, 182.506, 0.004, 0.702, 158.660, 182.411, -39.200, -39.200, 0.00, -4, 1, 0.996151, 0.701914, 0.008729, 0.034545
GCCUGAGAUCGGUCCAAGAGGUAACGAAUGCGAGGACCGGUACGGCCGCCACUCGCCGACUCUUGAUAAAGAUGGCGGCUCAGGC, 0.006, 0.774, 120.303, 139.307, 0.006, 0.707, 111.855, 139.204, -33.900, -35.100, 1.20, -4, 1, 0.994491, 0.706812, 0.234304, 0.209413
UCUGCAAACUAUGAGUGGCCGACGAACCGACGAGGACCGGUACGGCCGCCACUCGAAGGUUUGGUAACGAAUGCCACAAAGCAGA, 0.002, 0.594, 248.497, 249.017, 0.002, 0.556, 248.846, 248.855, -31.000, -31.700, 0.70, -4, 1, 0.997767, 0.555674, 0.431786, 0.360016
UCUGCUAACUAUGAGUGGCCGACGAACCGACGAGGACCGGUACGGCCGCCACUCGAAGGUUUGGUAACGAAUGCCACAAAGCAGA, 0.007, 0.803, 121.408, 121.482, 0.006, 0.750, 121.321, 121.329, -32.000, -33.300, 1.30, -4, 1, 0.993820, 0.749811, 0.412265, 0.353036
GUGAUCCACUAUGAGUGGCCGACGAAUCGACGAGGACCGGUACGGCCGCCACUCGAAGGUUUGGUAACGAAUGCCACAGGAUUAC, 0.006, 0.800, 114.672, 123.456, 0.006, 0.737, 123.360, 123.395, -30.100, -31.300, 1.20, -4, 1, 0.994026, 0.737162, 0.386218, 0.325203
GUGAUCCUACCUAUGAGUGACCGACACUGACGAGGACCGGUACGGCCGCCACUCGUAGGUGGUAACGAAUGUUAUUCAGGAUUAC, 0.013, 0.892, 23.948, 67.111, 0.012, 0.798, 66.968, 67.068, -30.700, -31.700, 1.00, -4, 1, 0.988102, 0.797973, 0.379268, 0.329051
GCUCAUUUUACGAGAGGCCGACGUGACUGACGAGGACCGGUACGGCCGCCACUCGUAUUUCAUGGUAACGAAUGCCAUUCUGAGC, 0.004, 0.702, 171.650, 182.943, 0.003, 0.522, 182.791, 182.796, -27.700, -28.700, 1.00, -4, 1, 0.997145, 0.521883, 0.438209, 0.367725
AAAGAAAAUACUCUUUUACGGGCGUCCGACCGAGGACCGGUACGGCCGCCACUCGGGUAACGAAUGACGGGAUCGUAAAAGAGAA, 0.012, 0.881, 73.944, 74.024, 0.010, 0.744, 73.597, 73.979, -32.100, -32.700, 0.60, -4, 1, 0.989889, 0.744144, 0.527156, 0.468019
ACGCUCCAAUUAUAACGGCCGACACGUUUACGAGGACCGGUACGGCCGCCACUCGAUCGUGGUAACGAAUGCCGAAAAGGAGCGA, 0.016, 0.910, 55.614, 56.246, 0.013, 0.756, 50.148, 56.154, -29.000, -30.400, 1.40, -4, 1, 0.986535, 0.756111, 0.340554, 0.284054
CCUCCUAUAGGAUGAUAUCGAAUCCGCCGACGAGGACCGGUACGGCCGCCACUCGGUAACGAAUGCGGCGACAUCCUGUAGGAGG, 0.011, 0.874, 72.217, 78.184, 0.010, 0.711, 68.434, 78.138, -41.100, -42.500, 1.40, -4, 1, 0.989604, 0.711440, 0.374165, 0.285274
Once again, all have relatively clean, symmetrical dot-plots, a strong, static neck, the aptamer and reporter both show in MFE for state 2, and the metrics are similar to those of prior lab winning designs (R101, R107) except for the seventh one whose d1 is just slightly high.
like: maxAx>20,maxRx>50,minAon>50%,minRon>50%,d1<0.5,d2<0.5,dFE<1.5
Here is a list of designs that I like for the two labs that are open for voting. It is my document with sets of different spirals. I have added a few extra designs for each the two labs and explained why I like them. It’s still a work in progress, but here is what I have so far:
Benbennet, you are correct that RNA do not make a literal spiral in lab, but in the arc plot. The reason I take interest in spirals in arc plots are that they have turned up in exceptional past lab winners.
Correct - a messy arc plot is not a goal in itself. Recently jandersonlee did some calculations over past lab designs that showed that the winning designs tend to have more clean arc plots. I start solving as I can, and as a consequence my plots are regularly messy. You are welcome to clean them up.
I have used jandersonlee’s arc plot tool to filter through all the Theophylline A Same State (MGA) designs submitted so far and used filtering to get rid of most of the designs. Then I got the tool spitting out arc plots and dumped them in a doc. Here are the resulting 144 arc plots. I have left the result of the full search at bottom of the doc.
If you average the arc plot pairing percentages for non-static pairs you can get a rough approximation of the experimental Fold Ratio. Perhaps this could be built into the arc plot generator as a way of averaging out the ratios figures given by the generator.
Obviously it’s not 100% accurate because the real world testing concentration is over a range, and my example doesn’t look at all bonding pairs, but I think it gives a better overall picture than the pairing ratios individually
Very interesting! In the meantime, now that the arcplot can also output data in .csv format, a column to do this could be added to a spreadsheet. Just tried it and I guess all the information you’re using isn’t in the .csv.
Multistate arcplot
I have an idea for multistate arc plot. (2+ states)
I will use a 3 state spiral puzzle I have made where all 3 states connectedly make a spiral - for illustration.
What I imagine is the arc plot as an opened book. Top page is the top arcplot for state 1. Bottom page is the bottom arc plot half for state 2.
!(https://d2r1vs3d9006ap.cloudfront.net/s3_images/1758555/RackMultipart20181024-91703-1qwt5ac-handbook-2435546_1920 2 inline.jpg?1540396640 “Image httpsd2r1vs3d9006apcloudfrontnets3_images1758555RackMultipart20181024-91703-1qwt5ac-handbook-2435546_1920__2__inlinejpg1540396640”)
Turn to next page and one could view an arc plot for state 2 and 3. It could also be 1 and 3. Or one could click an arrow to one of the sides for the page with arc plot for state 1 or 2, so one could swipe out one arc plot state for the next states arc plot.
Turn yet a page and one could view state 1 and 3 together.
Imaginary Arc plot tool with 3 states
I have been playing with your formula all morning and find that in some cases it comes amazingly close to results but in other cases, it is off considerably- like Jieux’s high-scoring Comedy 1 in Round 1. In any case, it made me look a lot closer at how the various numbers impact the results- thank you very much!
