Here are screenshots of the two puzzles. That is all I save.
Here is asteroid solution. No longer works, someone is moving the cheese.
AUUAUCACCCGAGAUAAAACACGCAGUGCAGCGUGAUAAUGCACAAAAUCGGGA
.(((((.{{{{{)))))…(((((.{{{{{)))))…}}}}}…}}}}}.
This solved in puzzlemaker half an hour ago and now it is not solving? Is that correct?
@JR1 The asteroid dot-bracket is .(((((.{{{{{)))))..(((((.{{{{{)))))....}}}}}....}}}}}.
You probably have the same dot-bracket, just the Forum formatting changing it on you.
Can it be fixed?
I know I was able to solve other puzzles with multiple pseudoknots.
Here’s a representative screenshot of the error I get when trying to load the puzzle:
And here are links to the two puzzles we’re all having issues loading (for future reference):
Asteroid headed for earth
Comet headed for earth
Both strings are the same.
Errors in puzzle maker: When there are pseudoknots:
1.) Some energy amounts are “0” when they should not be.
2.) If you delete a single or pair the design maps into a different configuration.
3.) Adding or subtracting pseudo brackets in structure string doesn’t work. I designed mine in a text file and then copy into puzzle maker. You get all sort of crazy results.
4.) When adding new brackets to structure you may need a error message of
unbalanced brackets.
5.) Very large numbers for delta difference. No longer means much.
Might want a pseudoknot icon that allows you to add/remove a pseudoknot in the puzzle itself (when the dust settles) if you are interested in designing pseudoknots in puzzle maker.
WRT the current puzzle crashes: I’ve identified an issue in the pair constraints that’s causing this. I’ll get the fix deployed probably tomorrow.
As far as the other puzzlemaker issues:
Some energy amounts are “0” when they should not be.
You may notice that when solving puzzles, we actually disable energies all together. There are a number of issues with energy calculations for pseudoknotted structures, plus we don’t currently have a good way to calculate how to display them even if they did work properly. It looks like for some reason in puzzlemaker they’re not getting completely disabled like they should be.
If you delete a single or pair the design maps into a different configuration.
I imagine this is due to the nature of how we represent things internally - what we work with ultimately isn’t the dot bracket, but a list of what bases are paired to what other bases. I’m guessing that it is “recalculating” the pseudoknotted structure, and choosing differently which pairs are treated “normally” vs as “pseudoknots” (as far as I know, there is no real difference, other than how we display them). If you’re seeing something else other than a re-configuration of what pairs are shown as pseudoknots, I’d need more details.
Adding or subtracting pseudo brackets in structure string doesn’t work. I designed mine in a text file and then copy into puzzle maker. You get all sort of crazy results.
Could you provide some examples? This may not be something we can easily resolve now, but might be able to look into with our upcoming rewrite
When adding new brackets to structure you may need a error message of unbalanced brackets.
Are you saying that it is showing incorrectly, or that it’s not showing when it should be?
Very large numbers for delta difference. No longer means much.
This is the nature of NUPACK - a high energy penalty means it encountered a structure that it determines is bogus, such as a U-U pair (which never occurs in its model) - applying this penalty is a way of indicating “this will never fold”.
On and for this, I think the ultimate goal will be to support magic glue in puzzlemaker - not sure that this can be done in the near term though
Here is a pseudoknot puzzle where I deleted NT45.
Resulting structure:
.(((((…)))))…(((…(…)))…)…
It looks like it changes the “{}” to “()” on a delete.
I added brackets to the original structure:
Structure:
.(((((…)))))…{…(((…}…)))…
It’s reversing single bracket with the 3NT stack. Is this going to work as far as seeing this in a RNA world?
Structure with single bracket:
.(((((…)))))…(((…{…)))…
There is no error now for single “()” so I guess no error for single bracket.,
Ah I think I see what’s going on - there’s some utility code that handles deleting bases via the UI, and it’s not doing its operations with pseudoknots enabled. Fix incoming!
Fixes released, changelog: Release v1.14.4 · eternagame/EternaJS · GitHub
Much better. Great fix. Thanks for your help.
Suggestion for upcoming events:
1.) Alt-Ctrl (Single NT 1)…Alt-Ctrl (Single NT 2) creates the pseudoknot. Right click… select “delete pseudoknot” from selection box to turn pseudoknot single NTs to regular NTs.
2’) Pseudoknot 2 state puzzles. Allow designer to create a pseudoknot in state 2 only that “switches” state 2 into a second configuration (like our 2 state switches now). Allows designer to create ratchets
and non-molecule switches. Take advantage of the thermodynamics of the RNA environment and makes for interesting labs.
I may have broken puzzle maker on stomping on existing structure with this string:
.(((((…}}}…{{{…)))))…(((((…}}}…{{{…)))))…(((((…}}}…{{{…)))))…(((((…}}}…{{{…)))))…
I hang when entering so can no longer use… Sorry about that.
I can enter through beam to puzzlemaker so I reset myself…