I just came across http://www.physorg.com/news79891654.html which claims:

"University of Chicago chemist David Mazziotti has developed a new method for determining the behavior of electrons in atoms and molecules, a key ingredient in predicting chemical properties and reactions. Results that formerly ranged from 71 to 96 percent accuracy have jumped to 95 to 100 percent. The contracted Schrödinger equation may soon become solvable with a package of computer software, according to Mazziotti."

The article refers to the paper here:

http://link.aps.org/abstract/PRL/v97/e143002

Is this research applicable to Rosetta@Home?

---

Rosetta uses a 'classical' approach in that the electronic structure and it's changes are incorporated into the model in approximate ways and not calculated in a direct fashion.

There have been a number of 'quantum' approaches (determining the electronic structure by solving Schroedinger's equations or approximations thereof) applied to the structure of peptides (the building blocks of proteins) as well as some small protein systems. But the general application (even using a method such as described here) of electronic structure for large molecular systems (thousands of atoms) are - typically - beyond the scope of generally available computational equipment.

-KEL

---

All 2-RDM methods scale to the fourth power in system size in storage and to the sixth power in processor time. The biggest thing we have done so far is benzene. This method is the last thing in the world anyone will ever use for protein folding. Quantum protein folding is likely to happen only, if ever, for linear scaling methods such as PW-LDA based CPMD.

Jeff