Gas station without pumps

2011 January 12

EteRNA, an Online Game

Filed under: Uncategorized — gasstationwithoutpumps @ 02:48
Tags: , , , ,

An article in today’s New York Times introduced a new “citizen science” game: EteRNA, an Online Game, Helps Build a New RNA Warehouse.

This game, following in the model of FoldIt for protein folding, has players designing RNA molecules.  Some of the designers of FoldIt helped design EteRNA.  The difference here is that the game is 2D (RNA secondary structure, rather than protein tertiary structure) and the task is to design RNA sequences rather than to fold given protein sequences.  (FoldIt has a protein design mode for researchers, but it isn’t part of the on-line game.)

EteRNA seems to be rather addictive.  There’s a bunch of crap on the screen (like a chat box that can’t be removed completely, and buttons that throw you out of the puzzle you are working on, which you can accidentally click when trying to color a base), but overall the interface is easy to learn and not too difficult to use.  There are times when I’ve wished for a more powerful operator that allows swapping the two bases of a pair, or replacing them with the other Watson-Crick pair, since it takes a lot of mousing back to the color palette to change two bases to different colors.

One cool thing about EteRNA is that the developers plan to synthesize some of the RNA designs that gamers come up with, in an attempt to check and improve the models used for folding RNA in the game.

It would be useful for gamers to know something about the base stacking model in EteRNA.  Luckily, one of the players has reverse-engineered the model so that the stacking contributions to the energy are available in a table.  It might be good to have some visual cues about stacking energy, though.

FoldIt and EteRNA seem to be fairly successful attempts to harness game players to do interesting optimization.  A different attempt, Phylo, which attempts to use humans to do sequence alignment, is a failure.  It is not fun to play, and the puzzles that people are given are small enough that would be faster to do the combinatorial optimization than to present the puzzles to the humans.  In short, Phylo is useful neither for gamers nor for science.

Additions: 12 Jan 2011

I played EteRNA for entirely too long last night (up until 4 a.m.).  I realized that there is a swap-bases tool, and that it isn’t the tool I want.

What I want is to have 5 tools selectable by typing a key: A, C, G, U, W for the 5 recognized base pairs (W for UG).  Clicking on a loop base would behave as now, putting in the A,C,G,U or U.  But on a paired base it would also fill in the other base appropriately with U,G,C,A, or G.  It shouldn’t matter whether the current display is the target structure or the current minimum-energy structure—the pairing is always part of the target design.

Two other things I’d like to see:

  1. a “randomize” option that can give me a random start (with all 5 tools applied randomly to a range of the RNA sequence)
  2. a bin of anti-design stems: pairs that formed in some incorrect minimum-energy structure but not in the desired structure.  I should be able to click on bases in the anti-design and have the tool applied to the base selected.  I would probably need to empty the bins occasionally as changes to the sequence make some anti-design stems no longer likely to appear in any minimum-energy structure.

8 Comments »

  1. This is the last thing I needed to see before my advancement.

    Comment by john — 2011 January 12 @ 11:16 | Reply

  2. I’m a champ at Phylo. I played Foldit when it was the new thing and never got into it. 3D+features was way too many variables for me.

    Now Eterna is basically about picking colors on a specific structure is how I see it, so that’s less daunting. For Eterna, you change colors, for Phylo, you change alignments. There still doesn’t seem to be a reason or rhyme with Eterna for me though.

    You can play Phylo with 8 sequences, and unless you’re hitting the high scores on those easily, I’m not sure how you can say it’s “too easy”.

    Jerome Waldispuhl recently spoke on using computers vs. humans and said “the problem of alignment with given phylogeny has been shown to be NP-complete…An exact method for solving the problem has been published in 2003 by B. Knudsen, but has been shown to be very limited in practice (3 and maybe 4 short sequences according to the paper).”

    The graphics aren’t as flashy, and the learning curve can be steep from the limited tutorial (I’m thinking about uploading some videos for beginners), but once you develop an intuition for Phylo, it’s amazing how much better 1 can do better than the computer. Besides, I like to listen to music and play this while I put it on mute.

    Comment by dude — 2011 January 12 @ 18:48 | Reply

    • The problem with Phylo is that the sequences are all short—short enough that the optimum dynamic programming is feasible, and that simple stochastic search does an almost perfect job of optimizing the cost function faster than problems can be presented to a human. The cost function is in many ways the wrong one and for the regions that haven’t beeen aligned automatically by the very fast heuristic aligner, the human solution is not any better biologically (even in the rare cases where it is better than just running a slightly more careful aligner on the same regions).

      Phylo is a hard game for humans, but the problem it is solving (in the sizes small enough to give to humans) is not hard for a computer. This makes it a fairly useless exercise in using human cognition to solve hard problems.

      Comment by gasstationwithoutpumps — 2011 January 12 @ 19:00 | Reply

  3. On the ap-bio mailing list, one teacher wrote
    > I don’t think I’ve seen anybody mention the Phlyo game on the
    > listserve yet. http://phylo.cs.mcgill.ca/eng/play.html Its a game
    > where one has to align genetic sequences over 3 or more species.
    > Kinda fun.
    >
    > There’s also one called “Fold It” where you have to predict how
    > polypeptide sequences will fold.
    >
    > Both work on the principle that our brains can detect patterns and
    > eliminate possibilities faster than computers.

    I’m fairly familiar with FoldIt, as it is in my research field (protein structure prediction) and one of my former PhD students is on the development team for it.

    FoldIt is doing some interesting things on a genuinely hard combinatorial optimization problem. The strategies used by good players look very little like the optimization techniques usually used on the problem, and some players are doing extremely well at finding good solutions. In short, it is a scientifically valuable project that makes good use of human skills. It is also fun.

    I’ve played Phylo a few times, and I have some familiarity with the multiple sequence alignment problem that they are having people work on. I don’t believe that the Phylo game provides a good use of human cognition or perception. Although multiple sequence alignment is hard (NP-hard in the number of sequences), the examples that they present to humans are tiny enough to be handled by routine computational methods—faster than the time it takes to deliver it to a human and get a result back. Furthermore, they do not give players enough time to do a really thorough job of optimizing. Like sudoku, the Phylo game is tough for humans and easy for computers. In short, it may have value to players as a game, but it provides no useful addition to computational biology.

    I think that EteRNA may be valuable to RNA computational biology, but it is still too soon to tell. I think it is a good teaching tool for teaching about RNA secondary structure, and the energy function they are using is typical of RNA secondary structure energy functions, including base-pair stacking terms as well as base-pair terms. I think they’ll need to add some tertiary structure information to the
    designs or provide more information about the distribution of energy in alternative stuctures if they want to make the game useful for actually designing RNA structures.

    Comment by gasstationwithoutpumps — 2011 January 14 @ 11:34 | Reply

  4. […] EteRNA, an Online Game […]

    Pingback by Blogoversary « Gas station without pumps — 2011 June 5 @ 10:50 | Reply

  5. […] EteRNA, an Online Game […]

    Pingback by Blog year in review « Gas station without pumps — 2012 January 1 @ 14:16 | Reply

  6. […] game so far to have made any real advances in molecular engineering. (The main other contender, EteRNA, though fun to play, does not capture enough in its scoring function for EteRNA play to do anything […]

    Pingback by Another success for FoldIt « Gas station without pumps — 2012 January 25 @ 20:20 | Reply

  7. […] EteRNA, an Online Game […]

    Pingback by Second Blogoversary « Gas station without pumps — 2012 June 2 @ 18:15 | Reply


RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: