Can we get a change in dot plot (delta dot plot) graph added to the tools? It would make comparisons while making a puzzle a lot easier and comparisons between winners and others way easier.
Hello eternacac
Iām not sure if I understand the requested feature - are you thinking about an interface with which you can load 2 different sequences and compare dot plots?
Back to this old but good idea. Iād like to be able to more easily compare how two design solves differ by dot plot, ideally seeing the effect of changes made while editing the design.
One thought is to use different color channels for each of two designs, and because green color-blindness is most common, the default mode might be red for one design and green+blue for the second. (Alternatively yellow(=red+green) for one design and blue for the other. Since stronger bonds mean darker, the net effect would be reversed. Where the first design bonded more strongly the dot-plot would look slightly blue-greener (or bluer); where the second design bonded more strongly, the dot-plot might look slightly redder (or yellower). I donāt know it the difference would be strong enough to show visually - weād have to try it to see.
In editing designs, one idea might be to āfixā one design as a reference, then show the comparative dot plot with the delta-colors as editing changes were made. For 2-way switches, the ability to show both states plus the delta dot-plot at the same time would be a bonus! It would make it easier to see how modifications to the design were affecting the dot-plot.
Alternatively, one could simply do a split-screen with the dot-plot for one design (e.g. a reference) on the left and the dot-plot for another design (the one being edited or an alternative submission) on the right for visual inspection. Being able to run Up/Down the design lists on either side would help, possibly using the U/D keys to switch designs as in the design review mode right now.
Unfortunately the U/D keys donāt currently work after the dot-plot has been viewed, requiring going back to the design list, picking a new design, then back to the dot-plot mode. Even just fixing this would be a help.
It would be ideal to be able to compare a design currently being edited with other (submitted?) designs in order to check the effect of modifications as well as to compare different designs that have already been submitted.
To just throw this out as an addition to Jandersonleeās new comment, I believe that another useful addition to the dotplot would be the ability to interact with the dotplot and have that interaction highlighted on the sequence in real time. For example, if I click on a helix or base pair in the dotplot coordinates, it should highlight those corresponding bases on the sequence, and perhaps even modify the resulting GUI structure to form the structure reflected in the dotplot. In addition, there could be an alternative viewing mode for the RNA called āPairing Probabilities Modeā which changes all the bases to a color scale, and provides the pairing probability values, based on the native or target structure fold, alongside each basepair.
Iāve always wanted to design/evaluate via the dot-plot. I did do some work on a stand-alone web-based tool for a while that did some of this. I can probably fire it up if anyone wants to look at it. (PM me in game for access.) One challenge I found on larger designs was the dots become so small it is harder to correctly select a given pair.
Brourdās suggested pairing probabilities mode reminds me of the yellow/blue SHAPE lab results. I never was very good at reading that but I think it was very helpful for some folks.
While we are at it, it might also be good to be able to compare dot-plots from different energy models to see how a design has different predictions for folding in different energy models. If Iām not mistaken I think the dot-plot is currently always based on the Vienna energy model regardless of the model selected for folding.
Yeah, a similarity to SHAPE was the idea Iām aiming for, but with the ability to inter-convert between the different structures in the dot plot (and then changing the GUI structure to reflect that). It would be a cool addition to the UI, and would give a more tangible feel to the dotplot and being able to view alternative structures in the ensemble that the sequence may be folding into. Players can then ātargetā those bases that are in these alternative structures.
Iāve had the same issue with trying to use the dotplot coordinates and trying to pinpoint the base pair it corresponds to.