Eterna3D - Testing and feedback

@LFP6 The chat window misplacement on the puzzle screen appears to be a dev site issue, not specific to the presence of the 3D window. I noticed it when testing another puzzle.

It was intentionally moved so that the 3D window would never start out below the chat - Iā€™ll need to look into how weā€™re supposed to address it.

Testing the 16S Slava made and running into odd behavior. The game is highlighting the base next to the actual base indicated by the cursor. My cursor is to the left of the black tag but not captured by the screenshot. The game is selecting the correct base when I click it to mutate, so itā€™s only a highlighting issue.
Screen Shot 2021-12-09 at 8.57.02 PM

Screen Shot 2021-12-09 at 9.11.17 PM

This is happening when I am in Pan mode. When in Rotate mode, Iā€™m unable to highlight or select any base in the 3D model.

The 2Dā€“>3D highlighting feature also is off by one base.

And the numbering isnā€™t displaying correctly. The numbering seems to be moved forward instead of being blocked by the front skittle.

Update: It looks like base 1 is missing from the 3D structure, so that could be why the highlighting is off. Might not be a code issue, might be a pdb/secstruct input issue.

How difficult would it be to connect the ā€œJump to Nucleotideā€ function to the 3D model? In large molecules, hunting through the 3D model to locate a base/helix will be challenging. Fixing the numbering issue to reduce visual clutter will help too.

I marked (cmd-click) a few bases in 2D and see the skittles flashing in 3D. That is really nice, that may be enough to navigate the 3D model. Donā€™t create a new Jump to Nucleotide tool for 3D, but thought Iā€™d mention it in case connecting the existing tool to 3D could be a quick addition.

1 Like

When making 3D puzzles, not only is it challenging to find a cif/pdb file that solely contains the desired molecule (they often contain several RNA chains and proteins), the RCSB usually does not provide the corresponding dot bracket notation. The solution I found is RNA central. Enter the four digit code or other designation and see if it pops up.

Update: Upon further examination, the dot-bracket notation from RNA Central often is not correct. Other tools to try are Nucleic Acid Database for a 2D diagram and entering the sequence at RNAfold for a a dot bracket.

March 21 Update
Marcin Magnus of the International Institute of Molecular Mechanisms and Machines (IMol) in Poland attended our town hall today when he saw he were discussing 3D models. He provided a link to his tool for getting secondary structure from a PDB file and extracting a chain. These should prove quite helpful when customizing a PDB file for an Eterna3D puzzle. The home page for all his tools is rna-tools.online

1 Like

Hello, Here is my presentation from the Eterna Town Hall - March 21, 2022

ā€œBuilding Your Own 3D Puzzleā€ by Craig Lollin [UserName: clollin]

Let me know if I can help anyone who is having trouble making a 3D puzzle.
Building Your Own 3D Puzzle - Eterna Presentation -3-21-2022.pdf (1.8 MB)

1 Like

Thanks for the presentation!

Iā€™m not yet familiar with Jmol so I havenā€™t tried the following in Jmol yet, but I did extract the RNA sequences for PDB 4WZJ using Chimera - including Google Drive links for them below in case anyone feels like making 3D puzzles with them. :slight_smile:

4WZJ - all RNAs in structure
4WZJ - single RNA from structure

Here are my thoughts after playing around with 3D. Excellent start. Butā€¦
If you really want to design 3D objects I suggest you look into a lego style interface. Jumping from 2D to 3D is a bit of hit and miss due to the natural coiling of each segment and the coil from loops kicking stacks in direction unknown. I would also test how robust RNAcomposer is if you are going to rely heavily on that application.

I really like how the tutorials focus the player on important areas to observe and examine more closely in the 3D model. I did just that and have a question/suggestionā€¦
Would it be possible to create a ā€œtoggleā€ on the background transparency of the 3D box to allow the user to more clearly visualize the structure without the distraction/clutter of the underlying puzzle bases and numbering? Perhaps a double-click hotspot on the 3 line bar at the top of the 3D box to go from transparent to solid background. Something that can be turned ON (solid background) or OFF (default transparent background) since both views are useful depending upon the situation. Just thought I would drop the idea here, but with full understanding that it is you coding wizards who know best!

1 Like