Sixth exercise: Structure prediction

Assignment due next Monday, April 25, 2005. Submit the sequence of a protein to 2 prediction servers. First try SWISS Model. If it works, then try one more. If it doesn't give you a model, then try 2 others including Rosetta. If you get a pdb file as output, make a Jmol page. Send me a short, < 1 page, report, and the Jmol page (if the pdb was available).

1. SWISS Model

If successful, you will receive a pdb file in an email attachment. If no suitable homology model exists, you will receive nothing! (Like this)

With better luck (>30% sequence identity to a known structure) this. Now you have a pdb file.

2. GenTHREADER

Results for nuoC

Results by email

3. Rosetta

This will return a pdb file by email (see email) (see Jmol)

4. Robetta

You must register for this prediction service.

Here is my result for secondary structure prediction.

It claims this molecule as the template for the 3D-structure (Jmol)

For 3-D prediction, select Fragment Libraries. You will get lots of small pdb files that predict the 3D sructure of segments of ~10 amino acids.

5. Prospector

Has not yet returned a prediction

6. 3D-PSSM

Here is my result. It gives a structure prediction in pieces. More manual work is required to build a model.

7. Fugue

Here is my result. It also gives a structure prediction in pieces.

PART 2

How to obtain the desired view in Jmol?

1. First open your pdb file in RasMol.

2. Using the mouse orient the protein for the desired view.

3. Type "save script 1xyz.spt"

4. Now open this script file in TextEdit or any word processor

5. Find the secton near the top that says something like this:

rotate z -75
rotate y -70
rotate x 72

6. To get the same view in Jmol, use these same angles except that you must do 2 things:

Add an initial line that says "rotate x 180"

and change the sign of the z rotation

7. So, in your Jmol modified-template file it should read:

rotate x 180
rotate z 75
rotate y -70
rotate x 72

8. Try it out.