Associative memory Hamiltonians for structure prediction without homology: Alpha-helical proteins

Corey Hardin, Michael P. Eastwood, Zaida Luthey-Schulten, Peter G. Wolynes

Research output: Contribution to journalArticle

Abstract

Energy landscape theory is used to obtain optimized energy functions for predicting protein structure, without using homology information. At short sequence separation the energy functions are associative memory Hamiltonians constructed from a database of folding patterns in nonhomologous proteins and at large separations they have the form of simple pair potentials. The lowest energy minima provide reasonably accurate tertiary structures even though no homologous proteins are included in the construction of the Hamiltonian. We also quantify the funnel-like nature of these energy functions by using free energy profiles obtained by the multiple histogram method.

Original languageEnglish (US)
Pages (from-to)14235-14240
Number of pages6
JournalProceedings of the National Academy of Sciences of the United States of America
Volume97
Issue number26
DOIs
StatePublished - Dec 19 2000

ASJC Scopus subject areas

  • General

Fingerprint Dive into the research topics of 'Associative memory Hamiltonians for structure prediction without homology: Alpha-helical proteins'. Together they form a unique fingerprint.

  • Cite this