The first 2 CASP 9 targets were released today: http://predictioncenter.org/casp9/targetlist.cgi
One target is for servers only and the other human target is 365 residues long so we will not be able to use it with Foldit unless we are able to break it down into smaller domains.
They announced that they are releasing 14 targets this week so we will have plenty to work with soon!
Until then, we are going to release some CASP training puzzles so you can train for CASP 9. These will be CASP 8 targets where we show you the native as a guide so you can see what moves needed to be done with the alignment tool in order to get closer to the native state.
We will also be posting an update soon that tells you which region of the alignment causes the "can't thread from alignment" message so that you can focus on fixing that part in the alignment tool.
Lastly, we will also be posting All Hands versions of the previous mini-CASP puzzles so that you can improve on each others' Foldit solutions. This could be a good way to select which models we should submit to CASP 9; we are allowed to submit 5 models per CASP target.
Hi,
Is there no way that you can adapt the contest design feature of the programme so that we can try out this first of the CASP targets? I looked at the design tool yesterday and it looks like it only does 40/80 residue extended chains. Can this be increased to….well…365 ?
It is not just the size of the protein that's a difficulty with these. If you look at the pages for the two human targets up now, you'll see in the "Additional Information" that they are part of larger complexes (dimers, trimers or hexamers). Basically this means that you need more than just that one protein itself to predict the structure accurately. We looked at T0517 for a while today, but concluded at this point that it's not appropriate for Foldit.
It's just the second day, and they will be releasing targets until July 17, so there should be plenty of good ones to work on.
Believe me, we are as anxious as you are!
We got very excited this morning when we saw that the new human target was only 159 residues and were really bummed when we realized that it really needs to be modeled as a trimer for the prediction to be successful.
We don't want to waste your time and post it as a useless puzzle that will just collapse on itself because one entire side is hydrophobic. Hopefully the CASP training puzzles will be more useful to all of you than that.
On the bright side it gives us more time to try out some of the other things that we need to test before submitting Foldit predictions to CASP 9 (such as using all hands to pick which solutions to submit).
If it makes you feel better, if CASP 9 has as many short monomeric proteins as CASP 8 did:
http://predictioncenter.org/casp8/targetlist.cgi?order=ASC&field=residues&view=all&view_targets=all
you will all have your hands full this summer!
