Protein Folding: The HP Model December 25, 2006
Posted by Geordie in Applications.trackback
The protein folding problem is this: given the primary structure of a protein (its linear sequence of amino acids), compute its three dimensional shape. Here is a good introduction.
One approximation that is often made to simplify the problem is to restrict the aminos to a lattice. A further restriction is to label all aminos as either H (hydrophobic) or P (hydrophilic).
Restricting to a lattice and H & P aminos gives the HP model, invented by Ken Dill. Finding the global energy minimum of this model of protein folding is NP-complete.
Scientists at Harvard University have developed a computer model that can fully map and predict protein folding for some 10 microseconds — about as long as some proteins take to assume their biologically stable configuration, and at least a thousand times longer than previous methods. It works for small proteins. This and further development of this method may limit the protein folding market for quantum computers.
Brian: Note that I didn’t claim QCs would actually be able to predict protein structure…my claim was that they could be used to find global energy minima of lattice models. Also I would take results (such as Harvard’s you reference) with a grain of salt. It’s easy to predict something, it’s hard to predict it correctly…
For example, the statement “the first computer model to fully replicate folding of a protein as happens in nature” is obviously incorrect and misleading. Unless all of modern physics is wrong, the way proteins actually fold is by evolution of the Schrodinger equation for all of its constituent electrons & nuclei (plus those of the solvent), and we all know how exponential that is.
You can compute protein flolding in many different ways, but in the real world, proteins fold only in certain (allowed) ways, so the problem is not the number of “fictive” folding calculations, but the constraints that allow only the “real” fold mechanisms to be profiled, eliminating the “false folds” from the computer models. And you usually need a supercomputer (f.eks. linux cluster with 1000 nodes) to do this kind of thing….
I attended an interesting bioinformatics seminar in 2004 sponsored by IBM and the Danish Technical university bioinformatics department, where IBM presended a special database application “Discoverylife” to calculate, among other things, protein folding.
Hi Geordie,
check this out, it looks interesting.
http://www.physorg.com/news109269396.html
also how many qbits are you up to now? when is your next demo?
congratulations on your latest financing