Task: You are to fold this protein (protein-T), given by its amino-acid sequence: 

ASN LEU TYR ILE GLN TRP LEU LYS ASP GLY GLY PRO SER SER GLY ARG PRO PRO PRO SER

(corresponding one-letter code for the same sequence: NLYIQWLKDGGPSSGRPPPS  ) 

Assume that nothing is known about this protein except its sequence above, 
and no similar ones are available in public databases, 
so standard tools based on machine learning won't work.
 
This is the so-called "de-novo" folding, the essence 
of the "Protein Folding" grand challenge of computational science. 


Steps: 

1. Familiarize yourself with the probelm. E.g. 

http://en.wikipedia.org/wiki/Protein_folding
http://folding.stanford.edu


2. We will approach the problem in the following manner, in three steps: A,B,  and C. 

(A) We will figure out what kind of numerical method we will need 
to attemp to solve the  problem. We will come up with an "energy landscape" 
that mimic real ones and see which methods work to find the global minimum. 

(B) We will get some ideas of which amino-acids might pair up in this protein using very simple arguments. The results will help us sanity check the computational solution we will obtain in teh last part. 


(C) 
once we are done with the above, we will use a commercial code AMBER (available 
on the class machine), and the ``best" numerical method for the problem 
you have identified, to actually try 
to fold a real (small) protein. 
We will visualize the results using free software VMD 
http://www.ks.uiuc.edu/Research/vmd/   or Pymol (see VT network software) 


DELIVERABLES:  

Brainstorming:  A+B (preliminary)
Mid-Term Report: A+B (final) 
Pre-final presentation: A+B+C (almost there)
Final report: A+B+C (final).