FASTA amino acid sequence input->contests

Case number:845799-988998
Topic:Biochem
Opened by:Gray Thomas
Status:Closed
Type:Suggestion
Opened on:Monday, December 13, 2010 - 17:44
Last modified:Friday, February 24, 2012 - 22:55

For a class project recently I was asked to annotate a hypothetical protein in the Pedobacter heparinus genome. It seemed like the protein was related to beta barrel TonB receptive proteins, but I had no way of telling if the protein actually made a beta barrel or not. Remembering the foldit demonstration in the beginning of the semester I thought, "wouldn't it be great if I could put my amino acid sequence into foldit, and then have a competition to see who could find the lowest energy solution?"

I imagine that the implementation might begin with an enormous, long line of peptides, as they might spew from the ribosome, and then the user would go along the line disabling the lock and wiggling until a rough estimate of the structure is in place. All the while, if they noticed a better solution they could help the computer out. Slowly the two would produce a reasonable estimate of the structure, and scientific curiosity would be fulfilled.

I'm very fond of your program, it was the one shining beacon of hope for a robotics major in a required molecular biology class.

Cheers!

(Mon, 12/13/2010 - 17:44  |  12 comments)


beta_helix's picture
User offline. Last seen 12 hours 56 min ago. Offline
Joined: 05/09/2008
Groups: None

How long is your sequence? We can post your amino acid sequence as a freestyle puzzle (where it just starts as an extended chain) or as a contest if you don't want everyone playing it for some reason.

Send me a private message if you want me to set that up for you.

Joined: 09/13/2010
Groups: None

Are you telling me this already exits!!
Woot! No reason to keep it closed.
How easy is it for you to post freestyle puzzles?
Can you do 936aa? If so, here it is:

>644936442 YP_003093057 hypothetical protein [Pedobacter heparinus DSM 2366: NC_013061]
MKKCAFILLALAMFGTTQINAQNKPSKPVAPPLPLREISGIVKDSTDLEV
IGATVSLTSDKDTLKTATNNNGVFVFKNVKSATYTIVVQSIGYMKTPAML
FKQNDRIPRIVMDPIVLKEEKNTLNTVVINGEPSITYKTDTIEYKASDYV
VRENSTVDELLKKMEGMEVGTDGSLVHNGEAVLKAKLNGKEYLGGDVANA
IKNLPAEIVDKIQIVDDYGDEAARTGVKDGDPQKILNITTRTDKSVGNLL
RLNTAGGNNKRYDAGIFGTRINGNQQIGVRTGYKNTVNGVAGSGGGGGTG
GTTNDGNASFSIRDKIGKKIEVNMNYGFNGNDINALNSSFTHSFTNFRDP
VTQKDTLIATNTTRGNVSDNNSKNHNFKADIDFELDSNNFINFSPSINYS
SSLSLRSDSSYQTGFRHQRQISSNASTNTRPQLGAAISYTHRFGNKRRRL
SLHANLNSNNNDVEQQREANIVYFLNEDETERRDSLVNRIIARKNLQNTY
RGSITYVEPLSLNTQLELNAQLNYNGYSNDATTRNVTPTGYSEVIDSLSN
IFDYSFTQARISLNFRYGMVNTSKFRFSAGLTGIPSLLSGTKVSLGTSTE
RKSFNLIPIARMEYLWSRQHKISLNYSGSANEPSFDQIQPVKDVSNPQNP
RIGNPDLKASFTHTIRSEYNNYLADIKMNYSLNMNASFTQNSVVNNTIDV
LDPNVVVGPNQKAVYISETRYINVNGSYRLAGNYSVSRQFGNRKYNLSFS
GSANYNHRISMSNGDQNINTVKTFIERFGPRINPSDWFEVNPYVSHNYTK
SENTLLKNQNETSTLAFNVDGKIYFLKSWLFGYSASKNYVSGIKANLTSN
PFVVNAYLEKELLNRRGKISLHAFDILNQNNFVNRDYNDNGGYTDTKSNA
LSQYFMLRLSMRLQKWSGAKGRKGQPIRRRGDGSFM

spvincent's picture
User offline. Last seen 5 hours 46 min ago. Offline
Joined: 12/07/2007
Groups: Contenders

I fear a protein with 936 amino acids would be way too big for FoldIt to handle. Not necessarily FoldIt itself but the speed/memory of the machines it runs on. One of the current puzzles has 196 residues and that's pretty big by FoldIt standards: it runs very sluggishly on my C2D 2GHz setup (not that that's exactly state of the art these days).

Joined: 09/13/2010
Groups: None

Can I cut it into 200 residue chunks and look for beta-barrel pieces? I could use my friend's sick gaming computer if I needed to. How does the computational complexity scale for number of residues? Alternately, could I just deal with it being slow? I think it might be worth it. Also, how do you make a puzzle? Are you and Beta-helix developers who have a unique opportunity to create puzzles, or is there a part of the interface which I am overlooking?

infjamc's picture
User offline. Last seen 2 years 45 weeks ago. Offline
Joined: 02/20/2009
Groups: Contenders

Re: Gray Thomas

Regarding the computational complexity question: from the limited information I've heard, it's O(n^3) in the worst-case scenario. So, I would recommend cutting the puzzle into at least five pieces, especially considering that templates are not available (at least according to my BLAST search with the protein data bank option).

beta_helix's picture
User offline. Last seen 12 hours 56 min ago. Offline
Joined: 05/09/2008
Groups: None
Status: Open » Open

Sadly, most puzzles that we post are <150 residues.

If you create a contest using "CASP9 Target 611 residue protein", you'll understand why:
http://fold.it/portal/node/add/contest

Maybe you can run some domain predictions on it and parse it up into different domains.

Otherwise, if it's a 936-residue beta-barrel, I don't think Foldit is going to work for you.

Send me a Private Message if you want to discuss different options.

Joined: 05/20/2011
Groups: None

Good job guys for doing great with the project. Robotics major is not easy but I know you can do it guys. Looking forward to all your projects and outputs in the future. Update us.

Joined: 11/02/2011
Groups: None

I want sb to predict the structure of this short peptide:
SGTSEKERESGRLLGVVKRLIVCFRSPFP
and send it to my e-mail
cedar402@163.com
Greatly thanks to the players

hilary's picture
User offline. Last seen 7 years 25 weeks ago. Offline
Joined: 02/24/2012
Groups: None

would it be possible to post the sequence VSKKKKRKKDRSTTDSGGNNMKKCRARFGLDQQSQWCKPCRRKKKCIRYMEALNGNGPAEDGSCFDEHGS? If so, thank you very much.

beta_helix's picture
User offline. Last seen 12 hours 56 min ago. Offline
Joined: 05/09/2008
Groups: None

You can create a contest for this sequence:
http://fold.it/portal/node/add/contest

Just select "Freestyle Design: Variable Length" and then insert/mutate the residues to the ones you want!

hilary's picture
User offline. Last seen 7 years 25 weeks ago. Offline
Joined: 02/24/2012
Groups: None

Thanks for the link. will do.

beta_helix's picture
User offline. Last seen 12 hours 56 min ago. Offline
Joined: 05/09/2008
Groups: None
Status: Open » Closed

marking as resolved

Sitemap

Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, RosettaCommons