Feedback thread for "Cloud Lab"

Feedback Feature Suggestions
26

The ability to Input structure and raw reactivity data for an output of the EteRNA “blue-yellow” interface.

An idea occurred to me, while messing around with the reactivity data and interface in the RMDB, that it would be nice to have a similar interface in EteRNA itself, where a player could input the sequence, structure, reactivity data, and anything else necessary, and players would get a similar output to the data we get out of the game’s lab interface. The interface should be similar to the puzzle maker, and I believe this would allow players, and maybe even researchers, to look at raw reactivity data that may not be published in the game, in an easy to use interface.

27

I’d like to be able to preview the lab shape in the lab tool during the voting phase. This should at least include being able to see what the predefined/fixed bases are, but possibly even give it a “test-drive”, without necessarily saving the candidate sequence. I might change my vote for a lab if I could see better how it could play out.

Not sure if this has been said before. If so, I missed it and maybe it deserves to be said again.

29

There should indeed be a natural way to do it, but it’s possible right now actually: change ‘web/lab’ to ‘game/puzzle’ in the URL of the proposed lab.

31

Awesome! Work-arounds are great. However this also means it should be easy to add this feature directly to the proposed lab review page as a link!

33

Hmm, technically, they cannot link to the lab using this, because it actually links directly to the puzzle itself, allowing submission/voting. That would essentially bypass the entire idea of having a lab proposed in the first place, if players submitted and voted on all their designs before a lab became active.

35

I’m sure they can tweak it so that submissions are not allowed during the proposal phase – and should!

36

Multiple State RNA Lab Projects Constraint Request

Once we gain the ability to run projects with multiple states, I would like a few features built into it.

I personally want to run projects, that will not necessarily require the RNA molecule to switch its conformation in the presence of a ligand, but will instead be an RNA sequence that can potentially fold into 2, 3, or maybe even 4 or more possible states. However, the details of my project would entail setting a required free energy difference between the different states of the RNA.

As a written example, let’s say I have two structures. I will want one structure to be the minimum free energy fold, while the other, I may want its predicted free energy to be between 1 kilocalorie and 2 kilocalories greater than the MFE structure. Then, I may want to run 4 identical projects, each differing based on the measured number of kcal between the states.

A feature like this can also prove useful to FMN switch player projects as well, if they wish to set and test the effect of free energy on some designs.

However, as of right now, players are able to bypass constraints when submitting designs for voting. I propose that players should have the ability to lock specific constraints for design submission. For instance, if I wanted players to stabilize the second state of a RNA molecule, but had no preference for the first, I should be able to single out that structure constraint for fulfillment, while leaving the others as they currently are.

37

I just realized, it seems like it will be an awful lot of work to make comparison designs for synthesis when it comes to the player projects.

For Instance, I have been working on the creation of non functional, pseudo RNA switches, by taking the method the Das lab designed for the creation of an FMN mimic sequence, and applying this to internal loops of various sizes.

The project would ask players to design the switch, and allow them to vote on which ones will be synthesized. After this point, all designs chosen for synthesis will then be duplicated with the internal loop mimic. The only issue is, every sequence that will be synthesized with the mimic sequence will need to be manually input in at this time, which can take quite a bit of work.

I think the inclusion of the ability to transfer these solutions all at one time should be allowed at the very least.

A more complicated solution would be the ability to split the lab and its synthesis slots into multiple copies of the same structure, with all sequences for each structure to be chosen from one single “master” lab challenge. For example, I may want to test the stability of loop mismatches. The “master” lab would allow players to design and submit sequences as they see fit. Then, one of the copies may have a locked AAAA tetraloop, and all designs submitted to the master lab, are then submitted into the locked AAAA tetraloop copy as well. At the end of player voting, the number of requested synthesis slots will be split between the master lab and the copies, and the original sequences along with their constrained copies will be synthesized.

This will be especially useful to players who wish to test hypotheses that require minute changes to synthesized sequences, without needing to create an entirely new project.

As a final request in this post, if it hasn’t been mentioned yet, I believe it would be useful to project admins if they could split the slots in a project, with some going towards voter control, and the rest going towards admin control. This would be especially useful in the scenarios written above, where players vote for 40 slots, and the admin would be able to choose how the other 40 in a project are allocated, whether it is towards varying the solutions or a specific mutation in the sequence.

39

I agree that it may be sometimes useful for a lab to be part voting and part administrator control. Perhaps this could be done by having a non-binding voting phase followed by administrator selection of the final results.

I’m not as clear on the other part. Multiple shapes? Multiple constraints?

41

It would be one shape with copies.

We’ll call the original the master shape.

The master shape, is the challenge for players to design around, they can do whatever they wish.

The copies of the master shape on the other hand, will have sequence constraints. However. the sequences that will be synthesized for the copies are not original, but duplicates of those submitted by players for the master copy, with the sequence constraints intact.

This would mostly be for comparing results across a large number of designs, where it would not be in the player’s best interest to individually copy and paste the 100 designs from the original project into a new project with a constraint like all adenine loops, or C-G/G-C multibranch loop closing base pairs. Hence, the first option was a request for a method to copy and paste large sets of sequences across labs.

43

regarding ‘batch upload’ (brourd’s first point), jee is setting this up with me since its quite common for my lab to need several automatically generated sequences to test the experimental pipeline. will work on making this feature useable by all.

44

Official Lab and Results Discussion Within the Game Itself

With the immense influx of data that we are getting from the new synthesis protocols, we are starting to reach a point, where we are generating data that is not being analyzed by players. I have personally been thinking of trying to get myself to write some kind of weekly blog about the lab results and the game in general to then post in the game.

Elnando has been doing an amazing job, writing his own personal blog and theories about RNA in the wiki.

Some players, like Eli, have been putting down their thoughts in Google Docs.

I myself have been throwing around ideas and hypotheses in the forum.

However, none of this data is consolidated, and many players may not even be bothering to check this stuff out because they may not have a link, or they may not even be thinking of searching so deep for player theories.

So, my proposal? Official in-game discussion threads for the lab results.

Now, my original post ended up being unnecessarily long, so here are the highlights of the plan.

Ease of Access, and More Exposure of the Lab Results and Discussion

This is the most important part. Players should be able to open the discussion thread, and easily have access to links to the lab results. Right now, the results from Round 72 are located on page 2 of the previous labs, and some players may not realize this. And slowly, all the results for previous labs are going to be pushed farther and farther down, eventually becoming inaccessible.

So, in addition to including links to the lab results in these discussion threads, I believe that these threads will need more exposure. Sure, we can include a link in the recent news, but some players may not notice that. I believe something a bit more pronounced is necessary. How about a scrolling banner at the top of all menu related pages? I would also include something under “Looking for something to do?” in our feed.

Next, I would have all players automatically follow and receive a random comment each day from the discussion thread. This will also help to put a spotlight on the discussion of the lab results.

Finally, since there are several weeks between each lab round start and finish, I think the discussion should be staggered. For example, with EteRNA Round 72, we had 12 active labs. Since we are getting to a point of about 3-4 weeks between the posting of results, I say we split the discussion into 3-4 parts. Week 1 discussion would cover labs 1-4, week 2 discussion would cover labs 5-8, and the third week’s discussion would cover labs 9-12. This would focus players on a specific set of labs, instead of telling them to post whatever they thought was interesting about any one of the 12 labs in a 3 week period.

I definitely welcome player and developer input. Is there something we can do better? Is there some way we can integrate these discussions into the resources we already have? Should we even implement something like this, or would developer resources be better spent on something else?

46

Of course, I’m biased. But I would say, wiki would be the perfect platform in my opinion, specially the day that it would get “unified” with the main site, with an URL like http://eterna.cmu.edu/wiki/… I can easily imagine a bot creating the wiki lab pages on the basis of templates, and categorizing them automatically (proposed, current, closed, static, switches, etc), allowing for easy searches. A few applets for displaying thumbnails, 2D structures with or w/o associated SHAPE data. The wiki itself would provide for participation (anyone can edit), cross-linking (you and I could link our “divagations” to dedicated pages, explaining what the bloody hell is a “Hoogsteen edge” for instance), archiving, reorganizing and consolidating (wiki-peeps call that a “master-edit”), etc.

1 Like
47

The active labs that show up at the top of the page list get a lot more attention than the others, in my experience. My theory is that this is because they are at the top of the page list. So I propose the ordering of the list be shuffled periodically or randomized based on player ID or something. Just a thought. ElNando - approved