Foldit – A game that contributes to protein folding research | Hacker News Reader

Foldit – A game that contributes to protein folding research

fold.it

85 pointsjoubert20 comments

EteRNA is a newer game along the same lines, but for RNA folding. It's a 2D, browser-based game, so some might find it easier to play.

http://eterna.cmu.edu/web/

Love this one!

Just a note to everyone: "browser-based" means "flash" here.

Related: this game grew out of Seth Cooper's PhD dissertation [1] "A Framework for Scientific Discovery through Video Games", which won the ACM Dissertation Award in 2011. Although a lot of the content is Foldit-centric (protein algorithms, etc.), it still has a wealth of information on game design and gamification, well worth looking at for anyone interested in game development.

[1] http://grail.cs.washington.edu/theses/Cooper2011PhD.pdf

I was lucky enough to see him speak at the University of Michigan (http://www.cse.umich.edu/eecs/etc/events/showevent.cgi?2665).

Also, I just finished reading REAMDE, and Fold.it is extremely similar to the fictional APPIS discussed in the book. It's still a fairly untapped area, looking to games in order to gain real life advantages. Although besides Fold.it, ReCaptcha, DuoLingo and the ESP game seem to be using related strategies to yield results of some degree.

Related: You can also donate your spare CPU/GPU cycles to similar projects with http://folding.stanford.edu/

Someone's even trying to make the points awarded for time spent folding into a basis for a crypto currency http://foldingcoin.net/

False economy. "CPU/GPU cycles" are only ever "spare" if you have free power.

Last I heard the "folding" distributed computation projects were mostly only being used to publish papers on distributed computation; and not really to advance medical science... (In contrast to Foldit)

Truly a great idea. Take human beings who would otherwise be wasting time and put their brain power to good use, converting couch potatoes of the world into research scientist. Not only does the lab produce new scientific results, but it brings thousands of strangers into the fold of protein biochemistry.

Another such, for astronomy, is Galaxy Zoo: http://www.galaxyzoo.org/ .

Another one is Boinc http://boinc.berkeley.edu/

This has been around for about 6 years now. I always wondered what happened to this project. Good to know that it's still alive, though it's popularity would skyrocket if they ever released a mobile version on either iOS or Android.

This got popular on reddit about a month ago. This thread has a warning about community moderation policies: http://www.reddit.com/r/gaming/comments/2n97b1/to_anyone_who...

tl;dr Consider that unprofessional moderation may expose your IP address, email, or other personal information and that others may be encouraged to harass you before playing.

I remember an old news about Foldit from 2011, where Foldit players were able to decipher the structure of an enzyme that is the key to the way HIV multiplies.

Source : http://www.cnet.com/news/foldit-game-leads-to-aids-research-...

"Gamers were able to do what biochemists have been trying to do for a decade: decipher the structure of a protein called retroviral protease, an enzyme that is key to the way HIV multiplies...People have spatial reasoning skill - something computers are not yet good at."

Strange thing is, if you Google "foldit HIV" or "foldit AIDS" - there is no follow up news. There are some new articles on 2013, but they link straight back to 2011 articles.

http://scholar.google.com/scholar?cites=2961090759639019608&...

Cited by 192

EDIT: Although looking through that link, most of the articles are about the technique rather than the actual discovery. I did find a followup paper:

http://www.nature.com/nbt/journal/v30/n2/abs/nbt.2109.html

Thank you. The Nature's paper isn't accessible for free, but some bibliographies are open for all.

Would be cool if there were a mobile app.

Boinc has mobile app. The current version isn't working on Android 5.x, but the developers are working on it.

Source : http://boinc.berkeley.edu/

Why is machine learning not good for this?

if you read up on foldit they take a hybrid strategy, for proteins that are simpler, they are using automated techniques. For proteins that are more complicated and harder to solve where folding requires satisfying a variety of conflicting constraints, Foldit is an experiment in using computer assisted human solving.

Reply on news.ycombinator.com