5 replies [Last post]
Joined: 11/10/2007
Groups: Window Group

CASP is coming up quickly, and the Alignment Tool will likely be important for most of the CASP puzzles. So we'd like it to be as effective and easy to use as possible.

Are there any suggestions for changes or new features for the alignment tool that would be useful?

Madde's picture
User offline. Last seen 2 years 33 weeks ago. Offline
Joined: 05/29/2008
Groups: Void Crushers
Apart from my and infjamc's

Apart from my and infjamc's suggestions in http://fold.it/portal/node/987340 I'd like a better ability of the tool to thread successful from an alignment. At least the tool could give us informations why it could not thread from the alignment.

spvincent's picture
User offline. Last seen 4 hours 25 min ago. Offline
Joined: 12/07/2007
Groups: Contenders
My 2 cents worth

I may be missing something obvious about the way it works but I find it difficult to use. It's very useful for getting some sort of a reasonable starting structure for the "straight line" puzzles, which otherwise involve so much tedious initial work that I'd stopped playing them. And using the initial threaded structures as a guide can be useful too.

But that's pretty much it for me, although I sometimes look at the other structures for inspiration. And judging by a non-scientific sampling of comments in chat I think others use it in that way only.

Some issues:

Threading invariably generates a structure that has a zero score and that also contains some residues with scores like -70,000 or whatever. By the time you've spent a rather frustrating half-hour dealing with these bits and finally getting a positive score, you've changed the protein quite a bit, probably negating any minor gains that might have been made from shuffling residues around in the top row.

And are you supposed to move the residues in the lower rows (templates) around? You're free to do so but it's not clear why you should.

Colours. There's no explanation for what they mean although it appears they refer to different groups of amino acids, partitioning the 20 amino acids into groupings other than hydrophobic/hydrophilic. A knowledge of which groups can be reasonably replaced with others is needed here.

Doesn't seem to work with proteins if any segments are locked. A pity this: if you could use threading on a small bit of an already folded protein: say copy just a loop from a template it could be very useful: a point already made I think by Madde and infjamc.

Documentation: videos are all very well but they're not really a substitute for having a written, detailed explanation that can be referred to when you need to look up a point you don't remember.

mimi's picture
User offline. Last seen 4 weeks 3 days ago. Offline
Joined: 11/17/2008
Groups: Contenders

Absolutely agree with spvincent.
There is an almost complete lack of useful information on which to base your adjustments.

What I find the most frustrating is to spend ages playing with the alignment to get a higher score only to find that it won't thread when the original "unimproved" alignment will.
Just tend to look at the results of the alignments and chose the one I feel is best as a starting point rather than waste time getting a better fit.

Brick's picture
User offline. Last seen 2 years 26 weeks ago. Offline
Joined: 07/15/2008
Groups: Beta Folders
Implementation is lacking

I'm sure someone put a lot of work into developing this, and as others have stated, it's useful for getting a starting structure on the straight line puzzles. Beyond that, it's like a lot of other features that have come along in the last year or so - the detailed and in-depth instructions on how to use the tool are missing. The video shows how the controls work, but not WHY any of the moves are being done. I think the developers assume we all know a lot more about this protein stuff than we really do. there's probably only a small handful that actually have the background to understand this. The rest of us just wiggle, shake and rebuild until it looks pretty and the score goes up. If anything, the skills are drying up, since now we just run sequences of recipes or scripts to get the best score we can and move on. I know I spend a LOT less time looking at the protein and thinking about how to improve it. I just wander by the PC, and if a script has stopped, I start another one.

Joined: 09/18/2009
Groups: SETI.Germany

That's right.
It shows how, but not why.
I don't understand the colors and letters.
Maybe you should give some graphics which show the single parts of the templates.
Anyway, it's a good thing that now the parts show up which made the aligment impossible.


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, Boehringer Ingelheim, RosettaCommons