Closed since almost 14 years ago
Intermediate Overall PredictionMost of the server predictions for T0644 were very similar except for the first 32 residues (the region that was unaligned to the starting template). Getting this N-terminal region right (which is strongly predicted to be helical) will separate the best predictions from the worst ones for this CASP10 target. For this quick puzzle: focus on folding the first 32 residues with the rest of the protein. Please read the puzzle comments about fixing an error in the starting model!
Below is a screenshot of all the Baker-RosettaServer, Quark, and Zhang-Server predictions.
The first 33 residues are shown as lines so that you can see how conserved the rest of the protein is.
The key to this CASP10 target will be figuring out where the helix goes, so good luck!
And here is the link to two different secondary structure predictions for this target: http://fold.it/portal/node/992552#comment-17173
The rebuild tool takes structure assignment into account, so you can set the secondary structure to helix and then Rebuild. Foldit will then make a helix out of that region.
In the screenshot below, the starting model for this puzzle is shown in the first image and in the second screenshot I've made a register shift on the ENTIRE SHEET (spanning residues 127-134). This was done just like Intro Level 4-3: Flippin' Sheets (everyone's favorite :).
INCORRECT (before):
.
.
CORRECT (after):
.
.
You can now see in my corrected model that the residues that previously pointing down (127,129,131,133) are now pointing up.
We believe that this is the correct topology and that by doing this register shift you will be able to find higher scoring models (after some massaging of the protein).
To perform a register shift, you'll need to change residue 134 (Glu) from Sheet to Loop or else you'll get a Shift Failed Error, and changing 137 (Lys) to loop as well might be a good idea.
You'll want to use the Tweak tool to shift once towards the left (in the decreasing direction, towards the start of the protein).
Pink is your starting model, green is what we believe needs to be done:
.
Above is with the strand in front of it trimmed so you can see better, below is the full structure:
.
thanks beta
For those who wanted a bit more info like me, I have consulted the interweb and learnt a lot more about some of the validation errors in the PDB, however flipping the sheet is much more practical advice :)
'what are register errors?
For a segment of a model the assigned sequence is out of register with the actual(electron) density.' - Gerard J. Kleywegt
…'These are the sequence-structure mapping or sequence-register shift errors. These particular errors arise from the incorrect assignment of residue side chains for a protein sequence segment within the electron density map. As a result, the position of the backbone for this segment usually is not significantly affected, but side chains are shifted along the backbone and regions adjacent to the segment have insertions/deletions.'
Perhaps at the top of the webpage, next to WIKI & FAQ, we can add a VIDEOS button linking to here:
For now, we've post a general link to the youtube channel in the FAQ: http://fold.it/portal/info/faq
Hopefully even a video that is specific to 557 can be useful to a new player trying to learn about how to do a register shift (for example if they are trying to beat Level 4-3: Flippin' Sheets)
I wonder if you could tell us what the likely function of the long hydrophobic helix is? Is it to attach to a membrane or another protein? It was hard to decide whether to try to hide as many hydrophobes as possible or to leave them exposed for – well, whatever they're for.
It turns out the first 25 residues were not solved :-(
http://predictioncenter.org/casp10/target.cgi?id=90&view=all
